Refine
Year of publication
Document Type
- Doctoral Thesis (38) (remove)
Has Fulltext
- yes (38)
Is part of the Bibliography
- no (38)
Keywords
- Antigen carrier (1)
- Binding kinetic (1)
- Corynebakterium efficiens (1)
- Fatty Acid Synthase (1)
- Fatty Acids (1)
- Fatty acid synthases (1)
- Fettsäuresynthase (1)
- Fettsäuresynthase Typ I (1)
- Flavin homeostasis (1)
- Inthraszentin (1)
Institute
RNA research is very important since RNA molecules are involved in various gene regulatory mechanisms as well as pathways of cell physiology and disease development.1 RNAs have evolved from being considered as carriers of genetic information from DNA to proteins, with the three major types of RNA involved in protein synthesis, including messenger RNA (mRNA), transfer RNA (tRNA), and ribosomal RNA (rRNA).2 In addition to the RNAs involved in protein synthesis numerous regulatory non-coding RNAs (ncRNAs) have been discovered in the transcriptome. The regulatory ncRNAs are classified into small ncRNAs (sncRNAs) with transcripts less than 200 nucleotides (nt) and long non-coding RNAs (lncRNAs) with more than 200 nt.3
LncRNAs represent the most diverse and versatile class of ncRNAs that can regulate cellular functions of chromatin modification, transcription, and post-transcription through multiple mechanisms.4 They are involved in the formation of RNA:protein, RNA:RNA and RNA:DNA complexes as part of their gene regulatory mechanism.4,5 The RNA:DNA interactions can be divided into RNA:DNA heteroduplex formation, also called R-loops, and RNA:DNA:DNA triplex formation. In triplex formation, RNA binds to the major groove of double-stranded DNA through Hoogsteen or reverse Hoogsteen hydrogen bonding, resulting in parallel or anti-parallel triplexes, respectively. In vitro studies have confirmed the formation of RNA:DNA:DNA triplexes.6 However, the extent to which these interactions occur in cells and their effects on cellular function are still not understood, which is why these structures are so exciting to study (Chapter I RNA:DNA:DNA Triplexes).
This cumulative thesis investigates several functional and regulatory important RNAs. The first project involves the improved biochemical and biophysical characterization of RNA:DNA:DNA triplex formation between lncRNAs of interest and their target genes. Triplex formation was confirmed by a series of experiments including electromobility shift assays (EMSA), thermal melting assays, circular dichroism (CD), and liquid state nuclear magnetic resonance (NMR) spectroscopy. The following is a summary of the main findings of these publications.
In research article 5.1, the oxygen-sensitive HIF1α-AS1 was identified as a functionally important triplex-forming lncRNA in human endothelial cells using a combination of bioinformatics techniques, RNA/DNA pulldown, and biophysical experiments. Through RNA:DNA:DNA triplex formation, endogenous HIF1α-AS1 decreases the expression of several genes, including EPH receptor A2 (EPHA2) and adrenomedullin (ADM), by acting as an adaptor for the repressive human silencing hub (HUSH) complex, which has been studied by our collaborators in the groups of Leisegang and Brandes.
2) Triplex formation between HIF1α-AS1 and the target genes EPHA2 and ADM was investigated in biochemical and biophysical studies. The EMSA results indicated that HIF1α-AS1 forms a low mobility RNA:DNA:DNA triplex complex with the EPHA2 DNA target sequence. The CD spectrum of the triplex showed distinct features compared to the EPHA2 DNA duplex and the RNA:DNA heteroduplex. Melting curve analysis revealed a biphasic melting transition for triplexes, with a first melting point corresponding to the dissociation of the RNA strand with melting of the Hoogsteen hydrogen bonds. The second, higher melting temperature corresponds to the melting of stronger Watson-Crick base pairing. Stabilized triplexes were formed using an intramolecular EPHA2 DNA duplex hairpin construct in which both DNA strands were attached to a 5 nucleotide (nt) thymidine linker. This approach allowed improved triplex formation with lower RNA equivalents and higher melting temperatures. By NMR spectroscopy, the triplex characteristic signals were observed in the 1H NMR spectrum, the imino signals in a spectral region between 9 and 12 ppm resulting from the Hoogsteen base pairing. To elucidate the structural and sequence specific Hoogsteen base pairs 2D 1H,1H-NOESY measurements of the EPHA2 DNA duplex and the HIF1α-AS1:EPHA2 triplex were performed. The 1H,1H-NOESY spectrum of the HIF1α-AS1:EPHA2 triplex with a 10-fold excess of RNA was semi-quantitatively analyzed for changes in the DNA duplex spectrum. We discovered, strong and moderate attenuation of cross peak intensities in the imino region of the NOESY spectrum. This attenuation was proposed to result from weakening of Watson-Crick base pairing by Hoogsteen hydrogen bonding induced by RNA binding. The Hoogsteen interactions can be mapped based on the analysis of the cross peak attenuation in the NOESY spectra, which we used to generate a structural model of the RNA:DNA:DNA triplex. These biophysical results support the physiological function of HIF1α as a triplex-forming lncRNA that recruits the HUSH-epigenetic silencing complex to specific target genes such as EPHA2 and ADM, thereby silencing their gene expression through RNA:DNA:DNA triplex formation.
Identification of new natural products from nematode-associated bacteria using mass spectrometry
(2023)
This work aims to find unknown natural products produced by bacteria, that live in close association with nematodes and to elucidate their structure by using mass spectrometry.
The first chapter of this work is dedicated to the detection of hitherto unknown natural products by using a metabolomics approach and subsequent structure elucidation of said compounds. This chapter includes metabolomics analysis of Xenorhabdus szentirmaii wild type and knockout mutants, overproduction of the target compound, identification of derivatives from other strains and MS based structure elucidation.
The second and third chapters are about natural products that protect C. elegans from B. thuringiensis infections.
The second chapter deals with natural products that protect the nematode host without killing the pathogen. I deployed molecular biology methods to generate deletion and overproduction strains of a target compound, identified it via LC-MS/MS analysis and used LC-MS/MS and lipidomics to analyse the chemical properties of the active compound.
The third chapter aims at finding natural products, which are produced by Pseudomonas strains MYb11 and MYb12, respectively. These natural products display the ability to protect C. elegans by killing B. thuringiensis. I identified said compounds via fractionation and subsequent bioactivity testing. After identification, I generated production strains of the target compounds and elucidated the structure of the bioactive derivative.
The last chapter deals with the structure elucidation of peptides produced by an unusual GameXPeptide synthetase in Xenorhabdus miraniensis. I analysed producer strains of GameXPeptides using LC-MS and elucidated the structural differences between the known GameXPeptides, produced by P. luminescens TT01, and the unusual ones produced by X. miraniensis.
Mitochondria perform essential energetic, metabolic and signalling functions within the cell. To fulfil these, the integrity of the mitochondrial proteome has to be preserved. Therefore, each mitochondrial subcompartment harbours its own system for protein quality control. However, if the capacity of mitochondrial chaperones and proteases is overloaded, mitochondrial misfolding stress (MMS) occurs. Upon this stress condition, mitochondria communicate with the nucleus to increase the transcription of nuclear encoded mitochondrial chaperones and proteases. This proteotoxic stress pathway was termed the mitochondrial unfolded protein response (UPRmt) aiming at restoring protein homeostasis. Despite being discovered over 25 years ago, the signalling molecules released by stressed mitochondria as well as the corresponding receptor and transcription factor remain poorly understood. With this study, we aimed at characterising the underlying signalling events and mechanisms of how mitochondria react to misfolded proteins. First, we aimed to establish different methods to induce MMS that triggers the transcriptional induction of mitochondrial chaperones and proteases detected by quantitative polymerase chain reaction. We were able to induce UPRmt signalling by overexpression of an aggregation-prone protein and by knock-down or inhibition of mitochondrial protein quality control components. To study the signalling in a time-resolved manner, we focused on the usage of the mitochondrial HSP90 inhibitor GTPP and the mitochondrial LONP1 protease inhibitor CDDO.
Early time point RNA sequencing analysis of cells stressed with GTPP or CDDO revealed upregulated genes in response to oxidative stress. Indeed, measurements of mitochondrial superoxide with the fluorescent dye MitoSOX showed increased levels of reactive oxygen species (ROS) upon MMS induction. In contrast, there was no induction of mitochondrial chaperones and proteases when combining MMS with antioxidants. Compartment-specific targeting of the hydrogen peroxide sensor HyPer7 revealed increased ROS levels in the intermembrane space and matrix of mitochondria, followed by elevated ROS levels in the cytosol at later time points. The importance of cytosolic ROS for the signalling was supported by preventing UPRmt induction with an inhibitor blocking the outer mitochondrial membrane pore. Thus, ROS were identified as an essential UPRmt signal.
To understand which cytosolic factor is modified by ROS, redox proteomics was performed. Here, reversible changes on cysteine residues of the HSP40 co-chaperone DNAJA1 were observed upon MMS. Consequently, transcriptional induction of UPRmt genes was abolished by DNAJA1 knock-down. To understand the function of DNAJA1 during UPRmt signalling, quantitative interaction proteomics upon MMS revealed an increased binding to mitochondrial proteins and its interaction partner HSP70. Immunoprecipitation confirmed a ROS-dependent interaction between HSP40 and HSP70. Increased binding to mitochondrial proteins represented a cytosolic interaction of DNAJA1 with mitochondrial precursor proteins, whose accumulation was confirmed by western blot. Moreover, a fluorescent protein targeted to mitochondria accumulated in the cytosol during GTPP treatment, confirming a reduced import efficiency upon MMS. Preventing the accumulation of precursors by a translation inhibitor or depletion of a general mitochondrial transcription factor resulted in reduced UPRmt activation. Thus, DNAJA1 is essential for UPRmt signalling, since its oxidation by mitochondrial ROS and its enhanced recruitment to mitochondrial precursors allows the integration of both MMS-induced signals.
To link these findings to an increased transcription of mitochondrial chaperones and proteases, we screened for transcription factors accumulating in the nucleus upon MMS by cellular fractionation mass spectrometry. We demonstrated that specifically HSF1 accumulates in nuclei of cells stressed with GTPP or CDDO. Depletion of HSF1 by knock-down or knock-out resulted in the abrogation of the UPRmt-specific transcriptional response. HSF1 activation was visualised by nuclear accumulation on western blot, a process inhibited by ROS and precursor suppression. Moreover, DNAJA1 depletion prevented HSF1 activation. Ultimately, we proved by immunoprecipitation that the inhibitory interaction between HSF1 and HSP70 is reduced upon MMS.
Thus, we conclude that MMS increases mitochondrial ROS that are released into the cytosol. In addition, the import efficiency is reduced upon MMS, resulting in the accumulation of non-imported mitochondrial precursor proteins in the cytosol. Both signals are recognised via DNAJA1 oxidation and substrate binding. The concurrent recruitment of HSP70 to DNAJA1 results in the loss of the inhibitory HSP70-HSF1 interaction. Thus, active HSF1 can migrate to the nucleus to initiate transcription of mitochondrial chaperones and proteases. These findings are in accordance with observations in yeast, where mistargeted mitochondrial proteins activate cellular stress responses. Our results highlight a surprising interconnection and dependence of the mitochondrial and the cytosolic proteostasis network, in which the UPRmt is activated by a combination of two mitochondria-specific proteotoxic stress signals.
This work addresses the investigation of the biosynthesis mechanisms of type II polyketide synthase (PKS) and fatty acid synthase (FAS) derived specialized metabolites (SMs) from Photorhabdus laumondii.
The elucidation of the biosynthetic pathway of the bacterial 3,5-dihydroxy-4-isopropyl-trans-stilbene (IPS) was one of the major topics of this thesis. IPS exhibits several bioactive characteristics as it inhibits the phenoloxidase of insects, acts antibacterial, but also influences the soluble epoxide hydrolase which is involved in inflammatory reactions. It was recently approved as a treatment against psoriasis by the FDA and is the first Photorhabdus derived drug.
The stilbene generation in Photorhabdus requires the formation of the two acyl-carrier-protein (ACP) bound 5-phenyl-2,4-pentadienoyl- and isovaleryl-β-ketoacyl-moieties. The ketosynthase (KS)/cyclase StlD catalyzes a ring formation via a Michael-addition between the two intermediates which is then further processed by an aromatase. The formation of 5-phenyl-2,4-pentadienoyl-ACP was shown via in vitro assays with purified proteins by proving the influence of the KS FabH, ketoreductase FabG and dehydratase FabA or FabZ of the fatty acid metabolism. While E. coli was able to complement most of these enzymes in attempts to produce IPS in the heterologous host, the Photorhabdus derived FabH was not replaceable despite 73 % sequence identity with the E. coli based isoenzyme, acting as a gatekeeper enzyme for cinnamic acid (CA) moieties. Furthermore, the ability to incorporate meta-substituted halogenated CA-derivatives was shown in order to produce 3-chloro- and 3-bromo-IPS. While studying the stilbene biosynthesis, the ability of Photorhabdus and Xenorhabdus to produce hydrazines was also discovered.
The second investigated biosynthesis was the formation of benzylideneacetone (BZA). BZA is produced by Photorhabdus and Xenorhabdus strains acting as a suppressor for the immune cascade of insects and has also antibiotic activities towards Gram-negative bacteria. Due to its structural similarity towards CA and the intermediates during the stilbene formation, a shared mechanism for Photorhabdus and Xenorhabdus budapestensis was proposed due to their ability to produce CA. The production of BZA was also dependent on the stilbene related CoA-ligase, the ACP and FabH. It was verified in vitro and in vivo in E. coli yielding a 150-fold increase of the BZA production compared to the Photorhabdus and Xenorhabdus wildtype (WT) strains.
The second part of this work deals with the optimization of P. laumondii strains regarding the production titers of IPS. Therefore, several deletions of other SM related genes as well as promoter exchanges in front of stilbene related genes were carried out. These approaches were combined with the upregulation of the phenylalanine by heterologous plasmid expression, since it is the precursor of CA. Another approach applied in parallel was the optimization of the cultivation conditions with different media and supplementation with XAD-resins. It was proved that media rich on fatty acids or peptides led to higher optical densities of the cultures and thus to higher titers of stilbenes. Since IPS is inhibiting the phenoloxidase, an enzyme important for the insect immunity, it was hypothesized that cultivation in media containing insects might enhance the output of this SM. Starting from 23 mg/l of IPS in the P. laumondii WT strain, it was possible to increase the production levels to more than 860 mg/l by utilizing the mentioned approaches.
The last topic of this thesis focuses on the production of epoxidated IPS (EPS) and its derivatives. Under laboratory conditions, only a low titer of EPS was observed for the wildtype strain. However, the optimized IPS strains and IPS-production conditions could also be applied for EPS which led to higher productions and also to the detection of many new derivatives. Most of the EPS derivatives were amino acid or peptide derived acting as nucleophiles to open the epoxide ring and yielding β-amino-alcohols. However, purification and chemical synthesis attempts to obtain EPS failed due to its poor stability. Epoxides were utilized in in vitro assays with amino acids, peptides and proteins to get insights whether epoxidations might act as posttranslational modification in Photorhabdus. The reactions were performed with styrene oxide and stilbene oxide replacing EPS based on their structural similarity. The modifications were executed successfully although proteomics approaches with in vivo data are required to confirm these findings. During the purification attempts of EPS, further derivatives were detected. The structures of dimerized stilbenes, a cis-isomer of IPS and another derivative that might incorporate an amino-group in the resveratrol ring were proposed on the basis of the HPLC-MS data.
This cumulative dissertation examines learning in chemistry laboratories, focusing on the challenges and benefits of problem-based learning (PBL) for novices in the lab. It addresses the lack of consistent understanding about what should be learned in labs and why it's important. The research aims to understand what students learn, how they learn, and how lab learning can be improved.
A central concept in PBL labs is Information Literacy, defined as a sociocultural practice enabling learners to identify and use information sources within a specific context as legitimized by the practice community.
The first publication, Wellhöfer and Lühken (2022a), investigates the relationship between PBL and learner motivation. It identifies factors that can foster students' intrinsic motivation in a PBL lab. Autonomy is found to be a key factor, increasing student motivation and presenting a model of the autonomous scientific process. This model involves four steps: information acquisition, designing and applying experimental procedures, experimental feedback, and autonomous process optimization. The results suggest that intrinsic motivation in PBL labs can be enhanced by enabling students to independently execute these steps.
The second publication, Wellhöfer and Lühken (2022b), examines the information process students undergo during their first PBL lab. Using a sociocultural framework, it explores Information Literacy to understand students' handling of information and their perceptions of the information process. The findings reveal that in PBL labs, developing a practical, applicable experimental procedure is crucial for problem-solving and significantly shapes the information-acquisition process. This process is iterative, influenced by new information, leading to more precise information needs. Students assess information quality based on its usefulness for their problem, implementability (considering cognitive understanding, available equipment, and psychomotor skills), and safety.
Furthermore, the role of privileged knowledge forms in evaluating the quality of text sources is explored. Students viewed non-scientific sources as "poor" and scientific sources as "good," yet used both for information gathering. There were discrepancies between their assessment of source quality and actual use, indicating that perception of source quality doesn't always affect their practical decisions.
The third publication, Wellhöfer, Machleid, and Lühken (2023), investigates students' information practices in the lab, focusing on discourse between novice learners and experienced assistants. It shows that theoretical knowledge isn't sufficient for independent practical action, and students need actionable social information from experienced community members. The results highlight that information literacy in the lab for newcomers to a community of practice has distinctive features, and physical experience and tacit knowledge are crucial for learning the methods and group-specific knowledge of the practice community. The article demonstrates how learning information literacy in a practice community requires a social and physical experience and provides insights on how educators can support this process.
Locomotion, the way animals independently move through space by active muscle contractions, is one of the most apparent animal behaviors. However, in many situations it is more beneficial for animals to actively prevent locomotion, for instance to briefly stop before reorienting with the aim of avoiding predators, or to save energy and recuperate from stress during sleep. The molecular and cellular mechanisms underlying such locomotion inhibition still remain elusive. So, the aim of this study was to utilize the practical genetic model organism Caenorhabditis elegans to efficiently tackle relevant questions on how animals are capable of suppressing locomotion.
Nerve cells, mostly called neurons, are known to control locomotion patterns by activating some and inhibiting other muscle groups in a spatiotemporal manner via local secretion of molecules known as neurotransmitters. This study particularly focuses on whether neuropeptides modulate such neurotransmission to prevent locomotion. Neuropeptides are small protein-like molecules that are secreted by specific neurons and that act in the brain by activating G protein-coupled receptors (GPCRs) expressed in other target neurons. They can act as hormones, neuromodulators or neurotransmitters. DNA sequences coding for neuropeptides and their cognate receptors are similar across diverse species and thus indicate evolutionary conservation of their molecular signaling pathways. This could potentially also imply that regulatory functions of specific neuropeptides are also similar across species and are thus meaningful to unravel more general mechanisms for instance underlying locomotion inhibition.
Specifically, we find that the modulatory interneuron RIS constitutes a dedicated stop neuron of which the activity is sufficient to initiate rapid locomotion arrest in C. elegans while maintaining its body posture. Similar to its known function in larval sleep, RIS requires RFamide neuropeptides encoded by the flp 11 gene for this activity, in addition to GABA. Furthermore, we find that spontaneous calcium activity transients in RIS are compartmentalized and correlated with locomotion stop. These findings illustrate that a single neuron can regulate both stopping and sleeping phenotypes.
Secondly, we show that C. elegans RPamide neuropeptides encoded by nlp-22 and nlp-2 regulate sleep and wakefulness, respectively. We unexpectedly find that these peptides activate gonadotropin-releasing hormone (GnRH)-like receptors dose dependently and we highlight their sequence resemblance to other bilaterian GnRH-like neuropeptides. In addition, we show that these receptors are expressed in distinct subsets of neurons that are associated with motor behavior. Finally, we show that nlp 22 encoded peptides signal through GNNR 6 receptors to regulate larval sleep and that nlp 2 encoded peptides require both GNRR 3 and GNRR 6 receptors to promote wakefulness.
In sum, we find that locomotion inhibition in C. elegans is regulated by multiple, but evolutionary conserved RFamide and GnRH-like RPamide neuropeptidergic signaling pathways.
Mechanism of the MHC I chaperone TAPBPR and its role in promoting UGGT1-mediated quality control
(2022)
Information about the health status of most nucleated cells is provided through peptides presented on major histocompatibility complex I (pMHC I) on the cell surface. T cell receptors of CD8+ T cells constantly monitor these complexes and allow the immune system to detect and eliminate infected or cancerous cells. Antigenic peptides displayed on MHC I are typically derived from the cellular proteome and are translocated into the lumen of the endoplasmic reticulum (ER) by the ATP-binding cassette (ABC) transporter associated with antigen processing (TAP), which is part of the peptide-loading complex (PLC). In a process called peptide editing, the MHC I-dedicated chaperone tapasin (Tsn) selects peptides for their ability to form stable complexes with MHC I. While initial peptide loading is catalyzed in the confines of the PLC, the second quality control is mediated by TAPBPR, operating in the peptide-depleted cis-Golgi network. TAPBPR was shown to have a more fine-tuning effect on the presented peptide repertoire rather than initial peptide selection. The fundamental mechanism of peptide editing was illuminated by two crystal structures of TAPBPR in complex with peptide-receptive MHC I. Notably, one of these structures reported a structural element that inserted into the peptidebinding pocket. The so-called scoop loop was assumed to be involved in mediating peptide exchange but the underlying mechanism remained undefined. Additionally, latest results suggested that TAPBPR mediates the interaction of the glucosyltransferase UGGT1 with peptide-receptive MHC. To expand the current knowledge of quality control processes in the antigen presentation pathway, the contribution of the scoop loop in peptide editing and the role of TAPBPR in UGGT1-mediated quality control needs to be elucidated. In the first part of this study, TAPBPR proteins with various loop lengths were designed to scrutinize the contribution of the scoop loop in chaperoning peptidereceptive MHC I. In a light-driven approach, the ability of TAPBPR variants to form stable complexes with peptide-free MHC I was tested. These results demonstrated that in a peptide-depleted environment, the scoop loop is of critical importance for TAPBPR to chaperone intrinsically unstable, peptidereceptive MHC I clients. Moreover, fluorescence polarization-based assays allowed the pursuit of peptide exchange in different, native-like environments. Peptide displacement activities of TAPBPR variants illustrated that catalyzed peptide editing is primarily induced by structural elements outside the scoop loop. In a peptide-depleted environment, the scoop loop occupies the position of the peptide C-terminus and acts as an internal peptide surrogate. By combining complex formation and fluorescence polarization experiments, the scoop loop of TAPBPR was shown to be critically important in stabilizing empty MHC I and functions as an internal peptide selector. In the second part of this study, a novel in-vitro glucosylation assay was established to examine the role of TAPBPR in UGGT1-catalyzed re-glucosylation of TAPBPR-bound MHC I clients. Therefore, a peptide-free MHC I-TAPBPR complex with defined glycan species was designed which served as physiological substrate for UGGT1. By subjecting the recombinantly expressed HLA-A*68:02- TAPBPR complex and UGGT1 proteins to the new in-vitro system, UGGT1 was shown to catalyze the transfer of a glucose residue to the N-linked glycan of TAPBPR-bound Man9GlcNAc2-HLA-A*68:02. Moreover, a high-affinity, photocleavable peptide was applied to dissociate the MHC I-chaperone complex. However, in the absence of TAPBPR, no glucosyltransferase activity was observed. Generation of peptide-free MHC I through UV illumination also showed no activity, and only the addition of TAPBPR could restore UGGT1-mediated reglucosylation of the empty MHC I. Independent of the peptide status of HLAA*68:02, the combination of protein glycoengineering and LC-MS analysis implicated that UGGT1 exclusively acts on TAPBPR-chaperoned HLA-A*68:02. The newly established system provided insights into the function of TAPBPR during UGGT1-catalyzed re-glucosylation activity and quality control of MHC I. Taken together, the scoop loop allows TAPBPR to function as MHC I chaperone through stabilizing peptide-receptive MHC I. In a peptide-depleted environment, the loop structure serves as an internal peptide surrogate and can only be dislodged by a high-affinity peptide. Based on these findings, TAPBPR fulfills a dual function in the second level of quality control. On the one hand, TAPBPR functions as peptide editor, shaping the repertoire of presented peptides. On the other hand, TAPBPR mediates peptide-receptive MHC I clients to the folding sensor UGGT1. Here, TAPBPR is essential to promote UGGT1-catalyzed reglucosylation of the N-linked glycan, giving MHC I a second chance to be loaded with an optimal peptide cargo in the peptide loading complex.
Non-ribosomal peptide synthetases (NRPSs) are modular biosynthetic megaenzymes producing many important natural products and refer to a specific set of peptides in bacteria’s and fungi’s secondary metabolism. With the actual purpose of providing advantages within their respective ecological niche, the bioactivity of the structurally highly diverse products ranges from, e.g., antibiotic (e.g., vancomycin) to immunosuppressive (e.g., cyclosporin A) to cytostatic (e.g., echinomycin or thiocoralin) activity.
An NRPS module consists of at least three core domains that are essential for the incorporation of specific substrates with the 'multiple carrier thiotemplate mechanism' into a growing peptide chain: an adenylation (A) domain selects and activates a cognate amino acid; a thiolation (T) domain shuffles the activated amino acid and the growing peptide chain, which are attached at its post-translationally 4ʹ-phosphopantetheine (4'-PPant) group, between the active sites; a condensation (C) domain links the upstream and downstream substrates. NRPS synthesis is finished with the transfer of the assembled peptide to the C-terminal chain-terminating domain. Accordingly, the intermediate is either released by hydrolysis as a linear peptide chain or by an intramolecular nucleophilic attack as a cyclic peptide.
The NRPS’s modular character seems to imply straightforward engineering to take advantage of their features but appears to be more challenging. Since the pioneering NRPS engineering approaches focused on the reprogramming and replacement of A domains, several working groups developed advanced methods to perform a complete replacement of subdomains or single or multiple catalytic domains.
The first part of this work focusses parts of the publication with the title 'De novo design and engineering of non-ribosomal peptide synthetases', which follows up assembly line engineering with the development of a new guideline. Thereby, the pseudodimeric V-shaped structure of the C domain is exploited to separate the N-terminal (CDSub) and C-terminal (CASub) subdomains alongside a four-AA-long linker. This results in the creation of self-contained, catalytically active CASub-A-T-CDSub (XUC) building blocks. As an advantage over the previous XU concept, the characteristics (substrate- and stereoselectivity) assigned to the C domain subunits are likewise exchanged, and thus, no longer represent a barrier. Furthermore, with the XUC concept, no important interdomain interfaces are disrupted during the catalytic cycle of NRPS, allow to expect much higher production titers. Moreover, the XUC concept shows a more flexible application within its genus origin of building blocks to create peptide libraries. Additionally, with this concept only 80 different XUC building blocks are needed to cover the entire proteinogenic amino acid spectrum.
The second part of this work addresses the influence of the C domain on activity and specificity of A domains. In a comprehensive analysis, a clear influence of different C domains on the in vitro activation rate and the in vivo substrate spectrum could be observed. Further in situ and in silico characterizations indicate that these influences are neither the result of the respective A domains promiscuity nor the C domain’s proofreading, but due to an 'extended gatekeeping' function of the C domain. This novel term of an 'extended gatekeeping' function describes the very nature of interfaces that C domains can form with an A domain of interest. Therefore, the C-A interface is assumed to have a more significant contribution to a selectivity filter function.
The third part of this work combines the NRPS engineering with phylogenetic/evolutionary perspectives. At first, the C-A interface could be precisely defined and further identified to encode equivalent information corresponding to the complete C-A didomain. Moreover, the comparison of NRPSs topology reveals hints for a co-evolutionary relatedness of the C-A didomain and could be shown to reassemble even after separation. In this regard, based on a designed CAopt.py algorithm, the reassembling-compatibility of hybrid interfaces could be determined by scoring of the co-expressed NRPS hybrids. This algorithm also enables the randomization of the interface sequences, thus, leading to the identification of more functional interface variant, which cause significantly higher peptide production and could even be applied to other native and hybrid interfaces.
This work characterizes the post-PKS modifications of AQ-256. Additionally, the second part describes the establishment of an AQ production platform for electrolyte generation that can be utilized in redox-flow-batteries. Lastly, a silent BGC that encodes the genes for terpenoid biosynthesis was described and characterized with regards to product formation and putative ecological function.
In this thesis, we characterized megasynthases such as fatty acid synthases (FASs) and polyketide synthases. The obtained insights into structure and function were used to engineer such systems to produce new-to-nature compounds.
The in vitro characterization of megasynthases requires reproducible access to these enzymes in high quality. Therefore, we established purification strategies for the yeast FAS and the methylsalicylic acid synthase (MSAS) from Saccharopolyspora erythraea (SerMSAS) and applied the latter one on MSAS from Penicillium patulum (PenPaMSAS) and on 6-deoxyerythronolide B synthase (DEBS) module 6. With the purified samples, we were able to obtain initial structural data for SerMSAS and solve the complete structure of the yeast FAS (PDB: 6TA1). On the example of the yeast FAS, we could show that the sample can suffer from adsorption to the water-air interface during the grid preparation for electron microscopy and presented how the use of graphene-based grids can overcome this problem. The combined structural and functional analysis of the yeast FAS showed that the structural domains trimerization module and dimerization module 2 are not essential for the assembly of the whole system. Therefore, they can potentially be used for domain exchange approaches. The in-depth functional analysis of SerMSAS revealed that not SerMSAS itself releases the product, but a 3-oxoacyl-(acyl-carrier protein) synthase like enzyme within the gene cluster transfers 6-methyl salicylic acid from SerMSAS to another carrier protein for subsequent modifications. In contrast, we showed that PenPaMSAS can release its product by hydrolysis and that non-native substrates can be incorporated although at significantly slower turnover rates compared to the native starter substrate. Our further investigation demonstrated that the substrate specificity of the acyltransferase (AT) is a critical factor for the incorporation of non-native substrates.
With the insight from the functional and structural characterization, we engineered megasynthases for the biosynthesis of natural product derivatives. We targeted the AT of PenPaMSAS for active site mutagenesis and discovered a mutant which can transfer non-native substrates significantly faster (~200-300%). Additionally, the malonyl/acetyl transferase (MAT) of the mammalian FAS was used as a promising target for protein engineering because of its previously reported properties including polyspecificity, fast transfer kinetics, robustness, and plasticity. We showed that the MAT can transfer fluorinated substrates and accept the acyl carrier protein of DEBS module 6. By exchanging the substrate specific AT of DEBS with the polyspecific MAT of the mammalian FAS, we demonstrated an efficient DEBS/FAS hybrid and an optimal truncation site for the applied ATs. In contrast to the wild type system, the DEBS/FAS enzyme was able to synthesize demethylated and fluorinated derivatives. The production and purification of a fluoro-methyl-disubstituted polyketide was of particular interest, as it has a high potential for the generation of new drugs and shows the potential of protein engineering. Furthermore, the incorporation of the disubstituted substrate had important implication in the mechanistic details of the ketosynthase-mediated C-C bond formation.
Non-ribosomal peptide synthetase docking domains : structure, function and engineering strategies
(2021)
Non-ribosomal peptide synthetases (NRPSs) are known for their capability to produce a wide range of natural compounds and some of them possess interesting bioactivities relevant for clinical application like antibiotics, anticancer, and immunosuppressive drugs. The diverse bioactivity of non-ribosomal peptides (NRPs) originates from their structural diversity, which results not only from the incorporation of non-proteinogenic amino acids into the growing peptide chain, but also the formation of heterocycles or further peptide modifications like methylation, hydroxylation and acetylation.
The biosynthesis of NRPs is achieved via the orchestrated interplay of distinct catalytic domains, which are grouped to modules that are located on one or more polypeptide chains. Each cycle starts with the selection and activation of a specific amino acid by the adenylation (A) domain, which catalyzes the aminoacyl adenylate formation under ATP consumption. This activated amino acid is then bound via a thioester bond to the 4’-phosphopantetheine cofactor (PPant-arm) of the following thiolation (T) domain. Before substrate loading, the PPant-arm is post-translationally added to the T domain by a phosphopantetheinyl transferase (PPTase), which converts the inactive apo-T domain in its active holo-form. In the last step of the catalytic cycle, two T domain bound peptide building blocks are connected by the condensation (C) domain, resulting in peptide bond formation and transfer of the nascent peptide chain to the following module. Each catalytic cycle is performed by a C-A-T elongation module until the termination module with a C-terminal thioesterase (TE) domain is reached. Here, the peptide product is released by hydrolysis or intramolecular cyclisation.
In comparison to single-protein NRPSs, where all modules are encoded on a single polypeptide chain, multi-protein NRPS systems must also maintain a specific module order during the peptide biosynthesis. Therefore, small C-terminal and N-terminal communication-mediating (COM) domains/docking domains (DD) were identified in the C- and N-terminal regions of multi-protein NRPSs. It was shown that these domains mediate specific and selective non-covalent protein-protein interaction, even though DD interactions are generally characterized by low affinities.
The first publication of this work focuses on the Peptide-Antimicrobial-Xenorhabdus peptide-producing NRPS called PaxS, which consists of the three proteins PaxA, PaxB and PaxC. Here, in particular the trans DD interface between the C-terminal attached DD of PaxB and N-terminal attached DD of PaxC was structurally investigated and thermodynamically characterized by isothermal titration calorimetry (ITC), yielding a dissociation constant (KD) of ~25 µM, which is a DD typical affinity known from further characterized DD pairs. The artificial linking of the PaxB/C C/NDD pair via a glycine-serine (GS) linker facilitated the structure determination of the DD complex by solution nuclear magnetic resonance (NMR) spectroscopy. In comparison to known docking domain structures, this DD complex assembles in a completely new fold which is characterized by a central α-helix of PaxC NDD wrapped in two V-shaped α-helices of PaxB CDD.
The first manuscript of this work focuses on the application of synthetic zippers (SZ) to mimic natural docking domains, enabling the easy assembly of NRPS building blocks encoded on different plasmids in a functional way. Here, the high-affinity interaction of SZs unambiguously defines the order of the synthetases derived from single-protein NRPSs in the engineered NRPS system and allows the recombination in a plug-and-play manner. Notably, the SZ engineering strategy even facilitates the functional assembly of NRPSs derived from Gram-positive and Gram-negative bacteria. Furthermore, the functional incorporation of SZs into NRPS modules is not limited to a specific linker region, so we could introduce them within all native NRPS linker regions (A-T, T-C, C-A).
The second publication and the second manuscript of this thesis again focus on the multi-protein PaxS, in particular on the trans interface between the proteins PaxA and PaxB on a molecular level by solution NMR. Therefore, the PaxA CDD adjacent T domain was included into the structural investigation besides the native interaction partner PaxB NDD. Before a three-dimensional structure could be obtained from NMR data, the NH groups located in the peptide bonds had to be assigned to the respective amino acids of the proteins (backbone assignment). Based on these backbone assignments, the secondary structure of PaxA T1-CDD and PaxB NDD in the absence and presence of the respective interaction partner were predicted.
The structural and functional characterization of the PaxA T1-CDD:PaxB NDD complex is summarized in manuscript two. The thermodynamic analysis of this complex by ITC determined a KD value of ~250 nM, whereas the discrete DDs did not interact at all. The high-affinity interaction allowed to determine the solution NMR structure of the PaxA T1-CDD:PaxB NDD complex without the covalent linkage of the interaction partners and an extended docking domain interface could be determined. This interface comprises on the one hand α-helix 4 of the PaxA T1 domain together with the α-helical CDD, and on the other hand the PaxB NDD, which is composed of two α-helices separated by a sharp bend.
...
The deubiquitinase USP32 regulates non-proteolytic ubiquitination in the endosomal-lysosomal system
(2021)
The regulation of essential cellular processes requires tightly controlled and directed transport of proteins and membranes. The highly dynamic endosomal and lysosomal system forms the key network for exchange and trafficking of molecules with its early endosomes, recycling endosomes, late endosomes, lysosomes, and additionally autophagosomes.
In this system, the small GTPase Rab7 has an essential role at the late endosomal stage regulating vesicle transport, tethering, and fusion, and retromer mediated receptor recycling back to the trans-Golgi network (TGN). Thus, Rab7 is also important for autophagosomes and lysosomes.
Lysosomes do not only represent the end point of the degradation pathway with several feeder pathways. But these organelles are also a dynamic signaling hub for a variety of metabolic processes. The ever-important regulator of cellular biosynthetic pathways mTORC1 dynamically associates with lysosomes where it is activated. mTORC1 activation is a complex multi-step process where a series of signaling events converge in dependence of amino acid levels thereby enabling interactions between the lysosomal v-ATPase, Ragulator complex (consisting of LAMTOR1-5), and Rag GTPases.
Ubiquitin signals are involved in almost all cellular processes. With this, their regulatory mechanism is also described for the endosomal-lysosomal system as well as mTORC1 signaling. Deubiquitinases (DUBs) release conjugated ubiquitin from proteins and thereby maintain the dynamic state of the cellular ubiquitinome.
The ubiquitin-specific protease 32 (USP32) is a poorly characterized DUB with only emerging cellular function. However, its predicted domain structure includes two unique domains within the entire DUB family. It has been linked to the development of breast cancer and small cell lung cancer. Furthermore, overexpressed GFP-USP32 was localized at the TGN, and a global mass spectrometry-based DUB interactome study suggested an interaction with the retromer complex. Based on these data, USP32 was a very interesting candidate to study its cellular function in this PhD project.
To investigate the function without disease background, a polyclonal USP32 knockout (USP32KO) RPE1 cell line was generated using the CRISPR/Cas9 technology. First experiments revealed different protein expression levels in various cell lines, and a subcellular localization of USP32 at membranes of the Golgi and lysosomal compartments. In a subsequent SILAC-based ubiquitinome analysis potential substrates of USP32 were identified. Interestingly, various proteins of the endosomal-lysosomal system were detected with enriched non-proteolytic ubiquitination upon USP32 depletion.
The further characterization of Rab7 as USP32 substrate confirmed the USP32-sensitive ubiquitination of Rab7 at lysine (K) residues 191 and 194. The ubiquitination in USP32KO cells did not change the subcellular localization of Rab7, but enhanced the interaction with the effector protein RILP. This implied that Rab7 was either more active or RILP had higher affinity to ubiquitinated Rab7. The subsequent results verified this theory. The retromer mediated recycling of CI-M6PR back to the TGN was faster or more efficient in USP32-depleted cells.
Accompanying this, levels of hydrolases were enriched in lysosomes isolated from USP32KO cells. Notably, USP32 had no direct effect on expression level or assembly of the retromer complex itself.
The observed lysosomal phenotypes connected another identified substrate to the function of USP32 in the endosomal-lysosomal system: LAMTOR1. LAMTOR1 is a component of the Ragulator complex and thus involved in the activation of mTORC1 at the lysosomal surface. Similar as for Rab7, the first experiments to characterize LAMTOR1 as USP32 substrate confirmed the USP32-sensitive ubiquitination at K20 independent of amino acid availability. However, ubiquitination of LAMTOR1 decreased its lysosomal localization in untreated and amino acid starved USP32KO cells. The following label-free interactome study detected a reduced interaction of LAMTOR1 and subunits of the lysosomal v-ATPase upon loss of USP32. This resulted in a shifted subcellular localization of mTOR (subunit of mTORC1) away from lysosomes. Furthermore, direct substrates of mTORC1 were less or slower re-phosphorylated after long amino acid starvation and re-activation of mTORC1 in USP32KO cells indicating a reduced mTORC1 activity.
Both USP32-dependent regulations of Rab7 and LAMTOR1/Ragulator converged in enhanced autophagic processes analyzed by increased LC3 levels upon amino acid starvation and USP32 depletion.
In summary, the presented thesis described the diverse role of USP32 in the endosomal and lysosomal system, and contributes to the understanding of novel ubiquitin signals in this context.
Ziel dieser Doktorarbeit war es, die Bedeutung der Kristallstrukturbestimmung aus Pulverdaten (SDPD) herauszuarbeiten und etwaige Grenzen durch neue Methodenentwicklungen zu erweitern, insbesondere bei Analyse der Paarverteilungsfunktion (PDF).
Die Effizienz der SDPD konnte anhand der erfolgreich gelösten Kristallstruktur von Carmustin (1,3 Bis-2-chlorethyl-1-nitrosoharnstoff, C5H9Cl2N3O2) aufgezeigt werden. [CS01]
Die Grenzen der SDPD wurden ausgelotet und erfolgreich erweitert. Nach weit verbreiteter kristallographischer Meinung ist die Strukturlösung mittels des simulierten Temperns (simulated annealing, SA) bei mehr als 25 zu bestimmenden Parametern problematisch oder unmöglich. Die pharmazeutischen Salze Lamivudin-Camphersulfonat (LC) und Aminogluthethimid-Camphersulfonat (AC) konnten, trotz ihrer hohen Anzahl an Freiheitsgraden von 31 für LC bzw. 37 für AC erfolgreich bestimmt werden. Die Strukturlösung von AC war herausfordernd und nicht direkt bei Anwendung der SA-Methode möglich. Nach einer intensiven Fehleranalyse stellte sich heraus, dass nicht die Grenzen der SA-Methode ausschlaggebend für das anfängliche Scheitern der Strukturlösung waren, sondern falsch extrahierte Intensitäten des vorangegangenen Pawley-Fits. Nach Behebung dieser Fehlerquelle war die Strukturlösung von AC problemlos. [CS02]
Mittels SDPD kann die absolute Konfiguration chiraler Verbindungen nicht direkt bestimmt werden. Durch Kristallisation der zu bestimmenden chiralen Verbindung mit einem chiralen Gegenion bekannter Konformation in einer simplen Säure-Base-Reaktion zu einem diastereomeren Salz und nachfolgender SDPD konnte eine neue Methode entwickelt werden, um die Konfigurationsbestimmung aus Pulverdaten zu ermöglichen. Diese Methode wurde anhand der drei pharmazeutischen Salze (R)-Flurbiprofen-(R)-Chinin (FQ), (2R5S)-Lamivudin-(R)-Camphersulfonat (LC) und (R)-Aminogluthethimid-(R)-Camphersulfonat (AC) aufgezeigt: In allen drei Fällen konnte die korrekte Konfiguration des pharmazeutischen Wirkstoffes mit den hierfür entwickelten Kriterien erfolgreich bestimmt werden. [CS03, CS04]
Durch Kombination der klassischen SDPD mit neuen methodischen Ansätzen konnten die Kristallstrukturen der schlecht kristallinen organischen Pigmente 2-Monomethylchinacridon (MMC, C21H14N2O2) und 4,11-Difluorchinacridon (DFC, C20H10N2O2F2) bestimmt werden, obwohl aufgrund ihrer geringen Kristallqualität keine sinnvolle Indizierung möglich war.
Für die Kristallstrukturbestimmung von DFC lieferte der neu entwickelte Global-Fit des Programms FIDEL mögliche Strukturmodelle mit ähnlich guter Übereinstimmung an das experimentelle Pulverdiagramm. Die Rietveld-Verfeinerung der Strukturmodelle in Kombination mit der Anpassung der Kristallstruktur an die PDF-Daten und kraftfeldbasierter Gitterenergieminimierung konnte einen geeigneten Strukturrepräsentanten von DFC liefern. [CS05, CS06]
Im Fall von MMC war eine Kombination der Methoden von Rietveld-Verfeinerung, Verfeinerung an die PDF-Daten und Gitterenergieminimierung zielführend zur Bestimmung der Orientierungs-Fehlordnung von MMC im Kristall. MMC ist hierbei die erste organische Verbindung, deren Fehlordnung durch Anpassung an die PDF bestimmt werden konnte. [CS07]
Große Erfolge konnten bei der Methodenentwicklung der PDF-Analyse erzielt werden. Die Bestimmung von Kristallstruktur organischer Verbindungen durch Anpassung an die PDF ohne vorherige Kenntnis der Gitterparameter oder Raumgruppe wurde durch die Entwicklung des PDF-Global-Fits erreicht. Lediglich die PDF-Kurve und eine Molekülstruktur werden als Input benötigt. Die Strukturlösung beruht auf einem globalen Optimierungs-Ansatz, bei welchem in ausgewählten Raumgruppen Zufallsstrukturen erzeugt werden. Die Zufallsstrukturen werden mit den experi¬mentellen Daten verglichen und entsprechend ihres Ähnlichkeitsindexes, basierend auf der Kreuz-Korrelation, sortiert. [CS08, CS09] Die vielversprechendsten Kandidaten werden in einem einge¬schränkten simulierten annealing-Ansatz an die experimentelle PDF angepasst. Eine nachfolgende Strukturverfeinerung der besten Strukturmodelle liefert die korrekte Kristallstruktur. Der Erfolg des PDF-Global-Fits wurde am Beispiel der Barbitursäure aufgezeigt: Ausgehend von 300 000 Zufallsstrukturen konnte die korrekte Kristallstruktur von Barbitursäure bestimmt werden. Barbitursäure ist hierdurch die erste organische Verbindung, deren Lokalstruktur durch Anpassung an die PDF bestimmt wurde, ohne Input oder Vorgabe von Gitterparametern oder Raumgruppe.[CS10]
This work comprises the investigation of four different biosynthesis gene clusters from Xenorhabdus. Xenorhabdus is an entomopathogenic bacterium that lives in mutualistic symbiosis with its Steinernema nematode host and together they infect and kill insect larvae. Xenorhabdus is well known for the production of so-called specialised metabolites and many of these compounds are synthesised by non-ribosomal peptide synthetases (NRPSs) or NRPS-polyketide synthase (PKS)-hybrids. These enzymes are organised in a modular manner and produce structurally very diverse molecules, often with the help of modifying domains and tailoring enzymes. In general, the genes involved in the biosynthesis are organised in so-called biosynthetic gene clusters (BGCs) in the genome of the producing strain. Exchanging the native promoter with an inducible promoter, e.g. PBAD, allows the targeted activation of the BGC and in turn the analysis of the biosynthesis product via LC-MS analysis.
The first BGC investigated in this work is responsible for the biosynthesis of xenofuranones. Based on gene deletions, this work shows that the NRPS-like enzyme XfsA produces a carboxylated furanone intermediate which is subsequently decarboxylated by XfsB to yield xenofuranone B. The next step in xenofuranone biosynthesis is the O-methylation of xenofuranone B to yield xenofuranone A. A comparative proteomics approach allowed the identification of four methyltransferase candidates and subsequent gene deletions confirmed one of the candidates to be responsible for methylation of xenofuranone B. The proteome analysis was based on the comparison of X. szentirmaii WT and X. szentirmaii Δhfq because distinct levels of the methylated xenofuranone A were observed when the xfs BGC was activated in either WT or Δhfq strain. Hfq is a global transcriptional regulator whose deletion is associated with the down regulation of natural product biosynthesis in Xenorhabdus. The strong PBAD activation of the xfs BGC also allowed the detection of two novel xenofuranone derivatives which arise from incorporation of one 4-hydroxyphenylpyruvic acid as first or second building block, respectively.
PBAD based activation of the second BGC addressed in this work lead to the detection of a novel metabolite and compound purification allowed NMR-based structure elucidation. The molecule exhibits two pyrrolizidine moieties and was named pyrrolizwilline (pyrrolizidine + twin (German: “Zwilling”)). The BGC comprises seven genes and single gene deletions as well as heterologous expression in E. coli and NRPS engineering were conducted to investigate the biosynthesis. The first two genes xhpA and xhpB encode a bimodular NRPS and a monooxygenase which synthesise a pyrrolizixenamide-like structure, similar to PxaA and PxaB in pyrrolizixenamide biosynthesis. It is suggested that the acyl side chain incorporated by XhpA is removed by the α,β-hydrolase XhpG. The keto function is then reduced by two subsequent two electron reductions catalysed by XhpC and XhpD. One of these two reduced pyrrolizidine units most likely is extended with glyoxalate prior to non-enzymatic dimerisation with the second pyrrolizidine moiety. To finally yield pyrrolizwilline, L-valine is incorporated, probably by the free-standing condensation domain XhpF.
The third BGC investigated is responsible for the production of a tripeptide composed of β-D-homoserine, α-hydroxyglycine and L-valine and is referred to as glyoxpeptide. This work demonstrates that the previously observed glyoxpeptide derivative is derived from glycerol present in the culture medium. Furthermore, this work shows that the monooxygenase domain, which is found in an unusual position between motifs A8 and A9 within the adenylation domain, is responsible for the α-hydroxylation of glycine. It is suggested that the α-hydroxylation of glycine renders the tripeptide prone to hydrolysis via hemiacetal formation. Hence, the XgsC_MonoOx domain might be an interesting candidate for further NRPS engineering.
The fourth BGC addressed is responsible for the production of xildivalines and this work describes two additional derivatives which are detected only when the promoter is exchanged and activated in the X. hominickii WT strain but not in X. hominickii Δhfq. Deletion of the methyltransferase encoding gene xisE results in the production of non-methylated xildivalines. It remains to be determined when the N-methylation of L-valine takes place. It is discussed that the methyltransferase could act on the NRPS released product but also during the assembly. The peptide deformylase is not involved in the proposed biosynthesis as xildivaline production is detected in a ΔxisD strain. The PKS XisB features two adjacent, so-called tandem T domains. The inactivation of the first or the second T domain by point mutation causes decreased production titres of detected xildivalines in the respective mutant strain when compared to the wild type.
Polyketides are highly valuable natural products, which are widely used as pharmaceuticals due to their beneficial characteristics, comprising antibacterial, antifungal, immunosuppressive, and antitumor properties, among others. Their biosynthesis is performed by large and complex multiproteins, the polyketide synthases (PKSs). This study solely focuses on the class of type I PKSs, which arrange all their enzymatic domains on one or more polypeptides. Despite their high medical value, little is known about mechanistic details in PKSs.
One central domain is the acyl transferase (AT), which is present in all PKSs and channels small acyl substrates into the enzyme. More precisely, the AT loads the substrates onto the essential acyl carrier protein (ACP), which subsequently shuttles the substrates and all intermediates for condensation and modification to additional domains to build the final polyketide.
Some PKSs use their domains several times during biosynthesis and work iteratively – these are called iterative PKSs. Others feature several sets of domains, each being used only once during biosynthesis – these PKSs are called modular PKSs. All PKSs or PKS modules consist of minimum three essential domains to connect the acyl substrates. Three modifying domains are optional and can enlarge the minimal set. According to the domain composition, the acyl substrate is fully reduced, partly reduced, or not reduced at all. This variation of modifying domains accounts for the huge structural and therefore functional variety of polyketides.
Even though the structure of fatty acids is not exactly reminiscent of polyketides, their biosynthetic pathways are closely related. Fatty acid biosynthesis is carried out by fatty acid synthases (FASs), which share many similarities with PKSs. Both megasynthases feature the same domains, performing the same reactions to connect and modify small acyl substrates. In contrast to PKSs, FASs always contain one full set of modifying domains which is used iteratively, leading to fully reduced fatty acids.
The present thesis extensively analyzes the AT of different PKSs in its substrate selectivity, AT-ACP domain-domain interaction, and enzymatic kinetic properties. The following key findings are revealed through comparison: 1.) ATs of PKSs appear slower than the ones of FASs, which may reflect the different scopes of biosynthetic pathways. Fatty acids as essential compounds in all organisms are needed in high amounts for physiological functions, whereas polyketides as secondary metabolites only require basal concentrations to take effect. 2.) The slower ATs from modular PKSs do not load non-native substrates even in absence of the native substrates. This is different to the faster ATs from iterative PKSs and FASs, which indicates high substrate specificity solely for the ATs from modular PKSs and emphasizes their role as gatekeepers in polyketide synthesis. 3.) The substrate selectivity can emerge in either the first or the second step of the AT-mediated ACP loading and is not assured by a hydrolytic proofreading function.
Moreover, a mutational study on the AT-ACP interaction in the modular PKS 6-deoxyerythronolide B synthase (DEBS) shows that single surface point mutations can influence AT-mediated reactions in a complex manner. Data reveals high enzyme kinetic plasticity of the AT-ACP interaction, which was also recently demonstrated for the interaction in a type II FAS.
Based on these findings, the mammalian FAS is engineered towards a modular PKS-like as- sembly line with the long-term goal to rationally synthesize new products. Basically, three important aspects need to be considered: 1.) AT’s loading needs to be splitted in specific loading of a priming substrate by a priming AT and in specific loading of an elongation substrate by an elongation AT. 2.) FAS-based elongation modules need to be designed with varying domain compositions for introducing functional groups in the product. 3.) Covalent and non-covalent linkers need to be designed for connection of priming and elongation modules.
This study focuses on the first aspect, splitting loading of priming and elongation substrates. An elongation substrate-specific AT is installed in the mammalian FAS via domain swapping. Since ATs from modular PKSs were proven to be substrate specific, these are used to exchange the mammalian FAS AT. This work demonstrates that it is extremely challenging to create stable and functional chimeras, but first essential steps are taken. Proper domain boundaries for AT swapping are established and a stable chimera with 70 % wild type AT activity is created. However, this chimera is only of limited value for application in an elongation module due to the intrinsic slow turnover rate of the wild type AT. Using another PKS AT, a stable elongation module is designed and analyzed in its activity in combination with a priming module. These experiments demonstrate that the loading of priming substrates are successfully suppressed in the elongation module, but nonetheless only minor turnover rates are detected in the assembly line.
...
The application of natural products (NPs) as drugs and lead compounds has greatly improved human health over the past few decades. Despite their success, we still need to find new NPs that can be used as drugs to combat increasing drug resistance via new modes of action and to develop safer treatments with less side effects.
Entomopathogenic bacteria of Xenorhabdus and Photorhabdus that live in mutualistic symbiosis with nematodes are considered as promising producers of NPs, since more than 6.5% of their genomes are assigned to biosynthetic gene clusters (BGCs) responsible for production of secondary metabolites. The investigation on NPs from Xenorhabdus and Photorhabdus can not only provide new compounds for drug discovery but also help to understand the biochemical basis involved in mutualistic and pathogenic symbiosis of bacteria, nematode host and insect prey.
Nonribosomal peptides (NRPs) are a large class of NPs that are mainly found in bacteria and fungi. They are biosynthesized by nonribosomal peptide synthetases (NRPSs) and display diverse functions, representing more than 20 clinically used drugs. Although a large number of NRPs have been identified in Xenorhabdus and Photorhabdus, the advanced genome sequencing and bioinformatic analysis indicate that these bacteria still have many unknown NRPS-encoding gene clusters for NRP production that are worth to explore. Therefore, this thesis focuses on the discovery, biosynthesis, structure identification, and biological functions of new NRPs from Xenorhabdus and Photorhabdus.
The first publication describes the isolation and structure elucidation of seven new rhabdopeptide/xenortide-like peptides (RXPs) from X. innexi, incorporating putrescine or ammonia as the C-terminal amines. Bioactivity testing of these RXPs revealed potent antiprotozoal activity against the causative agents of sleeping sickness (Trypanosoma brucei rhodesiense) and malaria (Plasmodium falciparum), making them the most active RXP derivatives known to date. Biosynthetically, the initial NRPS module InxA might act iteratively with a flexible methyltransferase activity to catalyze the incorporation of the first five or six N-methylvaline/valine to these peptides.
The second publication focuses on the structure elucidation of seven unusual methionine-containing RXPs that were found as minor products in E. coli carrying the BGC kj12ABC from Xenorhabdus KJ12.1. To confirm the proposed structures from detailed HPLC-MS analysis, a solid-phase peptide synthesis (SPPS) method was developed for the synthesis of these partially methylated RXPs. These RXPs also exhibited good effects against T. brucei rhodesiense and P. falciparum, suggesting RXPs might play a role in protecting insect cadaver from soil-living protozoa to support the symbiosis with nematodes.
The third publication presents the identification of a new peptide library, named photohexapeptide library, which occurred after the biosynthetic gene phpS was activated in P. asymbiotica PB68.1 via promoter exchange. The chemical diversity of the photohexapeptides results from unusual promiscuous specificity of five out of six adenylation (A) domains being an excellent example of how to create compound libraries in nature. Furthermore, photohexapeptides enrich the family of the rare linear D-/L-peptide NPs.
The fourth publication concentrates on the structure elucidation of a new cyclohexapeptide, termed photoditritide, which was produced by P. temperata Meg1 after the biosynthetic gene pdtS was activated via promoter exchange. Photoditritide so far is the only example of a peptide from entomopathogenic bacteria that contains the uncommon amino acid homoarginine. The potent antimicrobial activity of photoditritide against Micrococcus luteus implies that photoditritide can protect the insect cadaver from food competitor bacteria in the complex life cycle of nematode and bacteria.
The last publication reports a new family of cyclic lipopeptides (CLPs), named phototemtides, which were obtained after the BGC pttABC from P. temperata Meg1 was heterologously expressed in E. coli. The gene pttA encodes an MbtH protein that was required for the biosynthesis of phototemtides in E. coli. To determine the absolute configurations of the hydroxy fatty acids, a total synthesis of the major compound phototemtide A was performed. Although the antimalarial activity of phototemtide A is only weak, it might be a starting point towards a selective P. falciparum compound, as it shows no activity against any other tested organisms.
The dodecin of Mycobacterium tuberculosis : biological function and biotechnical applications
(2020)
Biological Function of Bacterial Dodecins
In this thesis, the dodecins of Mycobacterium tuberculosis (MtDod), Streptomyces coelicolor (ScDod) and Streptomyces davaonensis (SdDod) were studied. Kinetic measurements of the flavin binding of MtDod revealed that the dodecin binding pocket is filled in two distinct steps, for which a kinetic model then was established and verified by experimental data. The analysis with the two-step model showed that the unique binding pocket of dodecins allows them to bind excessive amounts of flavins, while at low flavin concentrations, flavin is released and only weakly bound. This function of flavin buffering prevents accumulation of free oxidised flavins and therefore helps to keep the redox balance of the cell and prevents potential cell damage caused by excessive free flavins. To further gain insights into the role of bacterial dodecins, the effect of knocking out the dodecin encoding gene in S. davaonensis was analysed. The knockout strain showed increased concentrations of various stress related metabolites, indicating that without dodecin the cellular balance is disrupted, which supports the role of dodecins as a flavin homeostasis factor.
With a self-designed affinity measurement method based on the temperature dependent dissociation of the dodecin:flavin complex, which allowed parallel screening of multiple conditions, it was shown that MtDod, ScDod and SdDod have much higher affinities towards FMN and FAD under acidic conditions. Under these conditions, the three dodecins might function as a FMN storage. M. tuberculosis encounters multiple acidic environments during its infection cycle of humans and can adopt a state of dormancy. During recovery from the dormant state, a flavin storage might be beneficial. For some Streptomyces species it was reported that the formed spores are slightly acidic and therefore ScDod and SdDod could function as flavin storages for the spores. Further details on the flavin binding mechanism of MtDod were revealed by a mutagenesis study, identifying the importance of a histidine residue at the fourth position of the protein sequence for flavin binding, but contrary to expectations, this residue seems only to be partly involved in the pH related affinity shift.
The data, reported in this thesis, demonstrates that bacterial dodecins likely function as flavin homeostasis factors, which allow overall higher flavin pools in the cell without disrupting the cellular balance. Further, the reported acid-dependent increase in binding affinity suggests that under certain conditions bacterial dodecins can also function as a flavin storage system.
Application of the Dodecin of M. tuberculosis
In this thesis, the stability of MtDod, ScDod SdDod and HsDod was analysed to find a suitable dodecin for the use as a carrier/scaffold. Therefore, a method to easily measure the stability of dodecins was designed, which measures the ability of the dodecamer to rebind flavins after a heating phase with stepwise increasing temperatures. Using this assay and testing the stability against detergents by SDS PAGE, showed that the dodecamer of MtDod possesses an excellent stability against a vast array of conditions, like temperatures above 95 °C, low pH and about 2% SDS. By solving the crystal structure of ScDod and SdDod, the latter forming a less stable dodecamer, combined with a mutagenesis study, the importance of a specific salt bridge for dodecamer stability was revealed and might be helpful to find further highly stable dodecins.
In addition to the intrinsic high stability of the MtDod dodecamer, also the robustness of the fold was tested by creating diverse MtDod fusion constructs and producing them in Escherichia coli. Here it was shown that MtDod easily tolerates the attachment of proteins up to 4-times of its own size and that both termini can be modified without affecting the dodecamer noticeably. Further, it was shown that MtDod and many MtDod fusion constructs could be purified in high yields via a protocol based on the removal of E. coli proteins through heat denaturation and subsequent centrifugation. In a case study, by fusing diverse antigens from mostly human proteins to MtDod and using these constructs to produce antibodies in rabbits, it was demonstrated that MtDod is immunogenic and presents the attached antigens to the immune system.
The here reported properties of MtDod and to a lesser degree of other bacterial dodecins, show that bacterial dodecins are a valuable addition to the pool of scaffold and carrier proteins and have great potential as antigen carriers.
Polyketide synthases (PKSs) are large megaenzymes that occur in bacteria, fungi, and plants and produce polyketides, a class of secondary metabolites. Many polyketide natural products exhibit high biological activities e.g. as antibiotics or anti-fungal compounds. The modular architecture of assembly line PKSs makes them exciting targets for engineering approaches via the exchange of whole modules or single domains. Although many engineering attempts have been pursued over the last three decades, the resulting chimeric PKSs often exhibit decreased turnover rates or diminished product yields.
In this thesis, new approaches to engineer chimeric PKSs were explored, each targeting a different aspect of the chimeric system: First the relative contribution of protein-protein and protein-substrate recognition on the turnover of chimeric PKS was assessed, revealing the importance of protein-protein interactions between the acyl carrier protein (ACP) and the ketosynthase (KS) domain in the chain translocation step. Directed evolution experiments followed to optimize the protein-protein interaction across a chimeric interface. Additionally, different junction sites for the generation of chimeric PKSs were compared, showing the ability for recombination without interfering with the chain translocation reaction, and highlighting the use of SYNZIP domains to bridge PKS modules. To optimize chimeric PKSs even further, multipoint mutagenesis of KS domains was established, with positive effects on the activity of chimeric systems.
To support engineering attempts, several structure elucidation techniques were combined with in silico modeling to characterize the architecture of a PKS module and the domain-domain interactions within it. Preliminary results show a strong conformational flexibility of the PKS module and the great potential of these techniques to define the multitude of transient interactions in PKS modules.
This work deals with the characterization of three different type II polyketide synthase systems (PKS II) from the Gram-negative bacteria Xenorhabdus and Photorhabdus.
Particular attention was paid to a biochemically underexplored class of aryl polyene (APE) pigments. Bioinformatic analysis of enzymes involved in the biosynthesis and the in vitro reconstruction proved that the synthesis of APEs involves an unusual fatty acid-like elongation mechanism. Furthermore, the discovery of unexpected protein-protein interactions provided new insights into the multienzyme complex formation of this unusual PKS II system. Through collaboration with the groups from Prof. Michael Groll and junior Prof. Nina Morgner, two protein complexes were structurally solved and several native protein multimerization events were identified and allowed us to suggest a possible protein-interaction network. The results are summarized in publication ‘An Uncommon Type II PKS Catalyzes Biosynthesis of Aryl Polyene Pigments’ (first author; J. Am. Chem. Soc.).
In addition to in vitro-analysis, in vivo-studies were used to investigate the APE compound produced by X. doucetiae in more detail. The activation of the silent biosynthetic gene cluster (BGC) led to the detection of the APE compound in the homologous host. Further combination of homologous expression and targeted deletions of the APE BGC revealed an APE-lipid-like structure. MS-based analyses and purification of intermediates allowed us to deduce structural building blocks of the APE-lipid, which is composed of an APE structural core, a glucosamine residue and an unusual long-chain fatty acid with unusual conjugated double bonds and a phosphoethanolamine head group. In combination with the above stated in vitro-data, we assumed a plausible biosynthetic mechanism of the APE-lipid. The results are summarized in the section ‘Additional Results: Tracing the Full-length APE’.
The biosynthesis of isopropylstilbene (IPS) has already been well-studied by the Bode laboratory and the group of Prof. Ikuro Abe. Studies with Photorhabdus laumondii TT01 by the Bode group revealed the distributed locations and functions of the genes involved in biosynthesis, which originate from two pathways. Particularly, the Bode group first demonstrated that an unusual ketosynthase/cyclase (StlD) catalyzes the condensation of 5-phenyl-2,4-pentadienoyl-ACP and isovaleryl-beta-ketoacyl-ACP via a Michael addition. Such a pathway for stilbene formation is distinct from those widespread in plants. The Abe group solved the structure and biochemical mechanism of StlD and further investigated the aromatization reaction of the aromatase StlC. However, the generation of the required cinnamoyl-precursor 5-phenyl-2,4-pentadienoyl-ACP as a Michael acceptor for this cyclization reaction remained elusive. In this work, we were able to reconstitute the synthesis of the Michael acceptor in vitro, by the action of enzymes from the fatty acid biosynthesis. With the knowledge about the crucial cross-talk from primary and specialized metabolism, we further determined the minimal endowment for stilbene production in a heterologous host. Here, the discovered AasS enzyme StlB is responsible for the generation of cinnamoyl-ACP and among others, plFabH plays a key role as gatekeeper enzyme for further processing. With this information in hand, we were able to obtain IPS production in E. coli. These results are presented in the manuscript ‘Biosynthesis of the Multifunctional Isopropylstilbene in Photorhabdus laumondii Involves Cross-talk Between Specialized and Primary Metabolism’ (co-first author, manuscript).
The biosynthesis of the orange-to-red-pigmented anthraquinones (AQs) is the best-studied type II PKS system according to preliminary results. While several investigations by Brachmann et al. discovered the BGC and the overall product spectrum of the main AQ-256 and its methylated derivatives, data of Quiqin Zhou (Bode group) performed biochemical in vitro analysis paired with in vivo heterologous expression of the ant-genes antA-I. This led to the identification of shunt products that indicated an AQ-scaffold derived from an octaketide intermediate that gets shortened to a heptaketide by the hydrolase AntI, resulting in the main anthraquinone AQ-256. This PKS-shortening mechanism was further confirmed by the protein crystal structure of AntI by the Groll group (publication, minor contributions, co-author, Chem Sci. ‘Molecular Mechanism of Polyketide Shortening in Anthraquinone Biosynthesis of Photorhabdus luminescens’). Further substrate analysis of the P. luminescens AQ-producer and mutants revealed an inhibitory effect of cinnamic acid against the hydrolase AntI. Cinnamic acid might therefore be involved in regulation of AQ biosynthesis (‘Anthraquinone Production is Influenced by Cinnamic Acid’, first author, manuscript).
Biochemical analysis from Quiqin Zhou with the minimal PKS of the AQ-synthase further revealed the exclusive activation of the AQ-ACP by the PPTase AntB. The PPTase is insoluble alone but gets stabilized by the CoA-ligase, most likely inactive, working as a chaperone. Thus, the minimal PKS endowment to produce the octaketide scaffold compromises, besides the ACP, the KS:CLF heterodimer and the MCAT, the co-occurrence of the PPTase AntB and the CoA-ligase AntG. For the first time, X-ray crystallography depicted a minimal PKS in action, by obtaining the structural data of native complexes from an ACP:KS:CLF, the KS:CLF alone and an ACP:MCAT in their non-active and active forms. It was possible to confirm a KS-bound hexaketide, which was built upon heterologous expression of the KS:CLF. Mutagenesis with amino-acids proposed to be involved in protein-protein interactions in the ACP:KS:CLF complex revealed some interesting protein-interaction sites. Additionally, an induced-fit mechanism of the MCAT with the ACP during the malonylation reaction confirmed a monodirectional transfer reaction (‘Structural Snapshots of the Minimal PKS System Responsible for Octaketide Biosynthesis’ co-author, manuscript under review).