Curriculum Vitae

Neural networks are widely used in bioinformatics to extract features from morphological, structural, and sequence data of different taxa. A key question is whether such features are compatible with a known phylogenetic tree describing the evolutionary relationships among the taxa. We address this question with a machine learning approach that takes taxon-specific data and a reference tree as input, and trains a neural network to produce a latent feature space whose pairwise distances are consistent with the tree topology. Our approach builds on the established role of quartets in distance-based phylogeny, leading to a quartet-based loss function for neural network training. In a proof-of-concept study using bacterial ribosomal RNA sequences, we show that the learned feature distances closely match the reference phylogeny. This framework can be applied to diverse biological data types, providing a principled way to incorporate phylogenetic constraints into neural network-based feature extraction.

Bacterial pathogens, such as Yersinia pseudotuberculosis, encounter temperature fluctuations during host infection and upon return to the environment. These temperature shifts impact RNA structures globally. While previous transcriptome-wide studies have focused on RNA thermometers in the 5’-untranslated region of virulence-related messenger RNAs, our investigation revealed temperature-driven structural rearrangements in the small RNA CyaR (cyclic AMP-activated RNA). At 25°C, CyaR primarily adopts a conformation that occludes its seed region, but transitions to a liberated state at 37°C. By RNA sequencing and in-line probing experiments, we identified the Shine-Dalgarno sequence of ompX as a direct target of CyaR. Interestingly, the ompX transcript itself exhibits RNA thermometer-like properties, facilitating CyaR base pairing at elevated temperatures. This interaction impedes ribosome binding to ompX and accelerates degradation of the ompX transcript. Furthermore, we observed induced proteolytic turnover of the OmpX protein at higher temperatures. Collectively, our study uncovered multilayered post-transcriptional mechanisms governing ompX expression, resulting in lower OmpX levels at 37°C compared with 25°C.

Pathogenic bacteria often encounter fluctuating reactive oxygen species (ROS) levels, particularly during host infection, necessitating robust redox-sensing mechanisms for survival. The LysR-type transcriptional regulator (LTTR) OxyR is a widely conserved bacterial thiol-based redox sensor. However, members of the Rhizobiales also encode LsrB, a second LTTR with potential redox-sensing function. This study explores the roles of OxyR and LsrB in the plant-pathogen Agrobacterium tumefaciens. Through single and combined deletions, we observed increased H2O2 sensitivity, underscoring their function in oxidative defense. Genome-wide transcriptome profiling under H2O2 exposure revealed that OxyR and LsrB co-regulate key antioxidant genes, including katG, encoding a bifunctional catalase/peroxidase. Agrobacterium tumefaciens LsrB possesses four cysteine residues potentially involved in redox sensing. To elucidate the structural basis for redox-sensing, we applied single-particle cryo-EM (cryogenic electron microscopy) to experimentally confirm an AlphaFold model of LsrB, identifying two proximal cysteine pairs. In vitro thiol-trapping coupled with mass spectrometry confirmed reversible thiol modifications of all four residues, suggesting a functional role in redox regulation. Collectively, these findings reveal that A. tumefaciens employs two cysteine-based redox sensing transcription factors, OxyR and LsrB, to withstand oxidative stress encountered in host and soil environments.

Agrobacterium tumefaciens is a plant pathogen, broadly known as the causal agent of the crown gall disease. The soil bacterium is naturally resistant to beta-lactam antibiotics by utilizing the inducible beta-lactamase AmpC. Our picture on the condition-dependent regulation of ampC expression is incomplete. A known regulator is AmpR controlling the transcription of ampC in response to unrecycled muropeptides as a signal for cell wall stress. In our study, we uncovered the global transcriptional regulator LsrB as a critical player acting upstream of AmpR. Deletion of lsrB led to severe ampicillin and penicillin sensitivity, which could be restored to wild-type levels by lsrB complementation. By transcriptome profiling via RNA-Seq and qRT-PCR and by electrophoretic mobility shift assays, we show that ampD coding for an anhydroamidase involved in peptidoglycan recycling is under direct negative control by LsrB. Controlling AmpD levels by the LysR-type regulator in turn impacts the cytoplasmic concentration of cell wall degradation products and thereby the AmpR-mediated regulation of ampC. Our results substantially expand the existing model of inducible beta-lactam resistance in A. tumefaciens by establishing LsrB as higher-level transcriptional regulator.

RNA secondary structures fulfill diverse regulatory functions in the cell. Their investigation is therefore key to understanding biological processes. The Lead-Seq method established here makes an important contribution to this by enabling transcriptome-wide mapping of RNA secondary structures in vivo. Another method presented here complements the in silico prediction of RNA structures. In an inverse folding based pre-training approach, deep learning methods are enhanced with prior knowledge from the thermodynamic model. This enables the inference of features from only a few data points, as shown in the example of transcription terminators. Last, a novel quartet loss is introduced that continues the integration of existing knowledge in deep learning methods. The trained models reproduce a known phylogeny and can be used to identify which features evolve along the given tree.

It is well-established that neural networks can predict or identify structural motifs of non-coding RNAs (ncRNAs). Yet, the neural network based identification of RNA structural motifs is limited by the availability of training data that are often insufficient for learning features of specific ncRNA families or structural motifs. Aiming to reliably identify intrinsic transcription terminators in bacteria, we introduce a novel pre-training approach that uses inverse folding to generate training data for predicting or identifying a specific family or structural motif of ncRNA. We assess the ability of neural networks to identify secondary structure by systematic in silico mutagenesis experiments. In a study to identify intrinsic transcription terminators as functionally well-understood RNA structural motifs, our inverse folding based pre-training approach significantly boosts the performance of neural network topologies, which outperform previous approaches to identify intrinsic transcription terminators. Inverse-folding based pre-training provides a simple, yet highly effective way to integrate the well-established thermodynamic energy model into deep neural networks for identifying ncRNA families or motifs. The pre-training technique is broadly applicable to a range of network topologies as well as different types of ncRNA families and motifs.

Small RNAs (sRNAs) are universal posttranscriptional regulators of gene expression and hundreds of sRNAs are frequently found in each and every bacterium. In order to coordinate cellular processes in response to ambient conditions, many sRNAs are differentially expressed. Here, we asked how these small regulators are regulated using Agrobacterium tumefaciens as a model system. Among the best-studied sRNAs in this plant pathogen are AbcR1 regulating numerous ABC transporters and PmaR, a regulator of peptidoglycan biosynthesis, motility, and ampicillin resistance. We report that the LysR-type regulator VtlR (also known as LsrB) controls expression of AbcR1 and PmaR. A vtlR/lsrB deletion strain showed growth defects, was sensitive to antibiotics and severely compromised in plant tumor formation. Transcriptome profiling by RNA-sequencing revealed more than 1,200 genes with altered expression in the mutant. Consistent with the function of VtlR/LsrB as regulator of AbcR1, many ABC transporter genes were affected. Interestingly, the transcription factor did not only control the expression of AbcR1 and PmaR. In the mutant, 102 sRNA genes were significantly up- or downregulated. Thus, our study uncovered VtlR/LsrB as the master regulator of numerous sRNAs. Thereby, the transcriptional regulator harnesses the regulatory power of sRNAs to orchestrate the expression of distinct sub-regulons.

The dynamic conformation of RNA molecules within living cells is key to their function. Recent advances in probing the RNA structurome in vivo, including the use of SHAPE (Selective 2′-Hydroxyl Acylation analyzed by Primer Extension) or kethoxal reagents or DMS (dimethyl sulfate), provided unprecedented insights into the architecture of RNA molecules in the living cell. Here, we report the establishment of lead probing in a global RNA structuromics approach. In order to elucidate the transcriptome-wide RNA landscape in the enteric pathogen Yersinia pseudotuberculosis, we combined lead(II) acetate-mediated cleavage of single-stranded RNA regions with high-throughput sequencing. This new approach, termed ‘Lead-seq’, provides structural information independent of base identity. We show that the method recapitulates secondary structures of tRNAs, RNase P RNA, tmRNA, 16S rRNA and the rpsT 5′-untranslated region, and that it reveals global structural features of mRNAs. The application of Lead-seq to Y. pseudotuberculosis cells grown at two different temperatures unveiled the first temperature-responsive in vivo RNA structurome of a bacterial pathogen. The translation of candidate genes derived from this approach was confirmed to be temperature regulated. Overall, this study establishes Lead-seq as complementary approach to interrogate intracellular RNA structures on a global scale.