PUB - Publikationen an der Universität Bielefeld

Diabetes mellitus is one of the most common diseases of our time and is one of the fastest rising global health risks. The life-threatening diabetes type 2 can be treated with the α glucosidase inhibitor acarbose, which is industrially produced by its natural producer, the soil bacterium *Actinoplanes* sp. SE50/110. While the acarbose biosynthesis (*acb*) gene cluster and the respective pathway have been extensively studied over decades, key regulatory elements for the understanding of acarbose biosynthesis and global transcriptional regulation are less investigated. Since acarbose production is coupled to growth, this study mainly focused on σ factors, which are transcriptional activators that are involved in major cellular processes such as growth phase-dependent and stress response regulation. It was demonstrated in less complex bacteria that the engineering of σ factors enabled the manipulation of natural product synthesis. The large genome, a complex life cycle, and expansion of regulation adds another layer of complexity to the investigation of filamentous actinomycetes. Comparison of *Actinoplanes* sp. SE50/110 to other filamentous actinomycetes shows an expansion of the σ factor families, which makes the manipulation of σ factors and its regulatory cascades challenging.

In this thesis the impact of σ factors and their antagonists on (i) acarbose yield (ii) cell morphology, (iii) regulation of the acb gene cluster, and (iv) stress response, were investigated in *Actinoplanes* sp. SE50/110 by the creation and characterization of gene overexpression and deletion mutants. Moreover, (v) binding motifs of their regulons were determined and the influence on (vi) production of other native natural compounds was investigated. Overexpression of the alternative σ factor coding *ACSP50_0507* (*sig0507*) or the anti-anti-σ factor coding *ACSP50_0284* achieved a two- or threefold increase in acarbose yield, respectively. Both strains showed an acarbose production phase extending into stationary growth, which is uncommon for *Actinoplanes* sp. SE50/110 where acarbose production ends at the beginning of the stationary growth phase. Expression of *sig0507* had a severe influence on mycelium morphology, showing hyperbranching and deformed cell shapes, while gene deletion had no influence on the strain’s physiology under standard growth conditions, but showed increased vulnerability to elevated osmolarities. The impact of σ factor engineering to enhance acarbose production was further investigated by transcriptome sequencing, which revealed an upregulation of acb genes during growth and at late stationary growth phase in the *sig0507* expression mutant. Genes coding for secreted and membrane associated proteins were transcriptionally upregulated. The cell morphology, gene upregulation and vulnerability of the *sig0507* deletion mutant to elevated osmolality further point an osmotic stress response-like function. The σ0507 DNA-binding motif was determined based on transcriptome sequencing data and shows high motif similarity to that of its Streptomyces coelicolor homolog σHSc. The motif was confirmed by in vitro binding of recombinantly expressed σ0507 His6 to the upstream sequence of a strongly upregulated gene. Autoregulation of σ0507 was observed and binding to its own gene promoter region was also confirmed. However, σ0507 does not directly regulate the acarbose biosynthesis gene cluster since it is not permanently upregulated under *sig0507* expression, and no σ0507 binding sequence was found within the gene cluster.

*ACSP50_0284* expression led to a similar phenotype like the *sig0507* expression, with a rendered cell morphology and extended acarbose production, most probably due to the temporary upregulation of the (*acb*) gene cluster. The gene is a homolog of the *S. coelicolor* anti-σHSc antagonist BldGAs. Due to an extended σ factor family, *Actinoplanes* sp. SE50/110 carries two homologs, of which ACSP50_0384 shows higher protein sequence similarity and phylogenetic relation to BldGAs. Since anti-σ factor antagonists have a regulatory function on protein level, transcriptomics allows for binding motif determination of an ACSP50_0284 regulated σ factor. This binding motif is similar to that of σ0507. A suspected regulatory connection of σ0507 and ACSP50_0284 is supported by further transcriptomic analysis which reveals that both share a large part of its regulons. However, its regulons are not identical, indicating that ACSP50_0284 it not the exclusive anti-σ0507 antagonist, but rather active on one of the σ factors of the same family. According to both their *S. coelicolor* homologs and regulatory similarity, *ACSP50_0507* is termed *sigHAs* and *ACSP50_0284* is termed *bldG2As*.

The transcriptional analysis led to the assumption that both proteins have a putative regulatory influence on the production of other native natural compounds and their associated gene clusters. Consequently, the natural product spectrum of *Actinoplanes* sp. SE50/110 was investigated, leading to the identification of five products of the 2 hydroxyphenylthiazoline family. Moreover, the structures of two novel molecules, thiazostatin C and thiazostatin D, were elucidated and their cytotoxic activity was demonstrated. It was further confirmed that manipulation of *sigHAs* and *bldG2As* also affected their production. Strikingly, a novel unidentified compound was discovered, which is exclusively produced under *bldG2As* expression, highlighting the potential of σ factor engineering to activate silent gene clusters and exploit the biosynthetic potential of established strains.

This work investigated the regulation of growth-related cell morphology and natural product synthesis, which appear to be linked in *Actinoplanes* sp. SE50/110. It successfully demonstrates that σ factor engineering can be successfully applied in regulatory complex bacteria to manipulate diverse aspects of growth and natural product synthesis. However, it reveals challenges for the investigation of regulatory networks, showing severe secondary effects on the whole transcriptome. It further shows that the expansion of regulatory network in complex species makes homology based functional predictions insufficient and limits function predictions to the respective regulatory family.