Refine
Year of publication
- 2014 (147) (remove)
Document Type
- Doctoral Thesis (147) (remove)
Language
- English (147) (remove)
Keywords
- Maus (8)
- Thrombozyt (5)
- Genexpression (4)
- Regulation (4)
- T-Lymphozyt (4)
- Biodiversität (3)
- Klimaänderung (3)
- Leistungsbewertung (3)
- Molekularbiologie (3)
- Netzwerk (3)
Institute
- Graduate School of Life Sciences (47)
- Theodor-Boveri-Institut für Biowissenschaften (30)
- Physikalisches Institut (14)
- Institut für Informatik (10)
- Fakultät für Physik und Astronomie (6)
- Fakultät für Chemie und Pharmazie (5)
- Institut für Psychologie (5)
- Institut für Theoretische Physik und Astrophysik (5)
- Institut für Geographie und Geologie (4)
- Institut für Mathematik (3)
Sonstige beteiligte Institutionen
- DLR (1)
- Department of Biochemistry (1)
- IBMP - Institut für Biomedizinische und Pharmazeutische Forschung in Nürnberg-Heroldsberg (1)
- Institut für Tierökologie und Tropenbiologie (1)
- Institut für klinische Neurobiologie (1)
- Instituto de Hygiene Montevideo, Uruguay (1)
- Klinische Mikrobiologie am Universitätsklinikum Erlangen (1)
- MRB Research Center for Magnetic-Resonance-Bavaria (1)
- Universitätsklinikum Münster (1)
ResearcherID
- I-5818-2014 (1)
Organisms have evolved endogenous clocks which allow them to organize their behavior, metabolism and physiology according to the periodically changing environmental conditions on earth. Biological rhythms that are synchronized to daily changes in environment are governed by the so-called circadian clock. Since decades, chronobiologists have been investigating circadian clocks in various model organisms including the fruitfly Drosophila melanogaster, which was used in the present thesis.
Anatomically, the circadian clock of the fruitfly consists of about 150 neurons in the lateral and dorsal protocerebrum, which are characterized by their position, morphology and neurochemistry. Some of these neurons had been previously shown to contain either one or several neuropeptides, which are thought to be the main signaling molecules used by the clock. The best investigated of these neuropeptides is the Pigment Dispersing Factor (PDF), which had been shown to constitute a synchronizing signal between clock neurons as well as an output factor of the clock.
In collaboration with various coworkers, I investigated the roles of three other clock expressed neuropeptides for the generation of behavioral rhythms and the partly published, partly unpublished data are presented in this thesis. Thereby, I focused on the Neuropeptide F (NPF), short Neuropeptide F (sNPF) and the Ion Transport Peptide (ITP). We show that part of the neuropeptide composition within the clock network seems to be conserved among different Drosophila species. However, the PDF expression pattern in certain neurons varied in species deriving from lower latitudes compared to higher latitudes. Together with findings on the behavioral level provided by other people, these data suggest that different species may have altered certain properties of their clocks - like the neuropeptide expression in certain neurons - in order to adapt their behavior to different habitats.
We then investigated locomotor rhythms in Drosophila melanogaster flies, in which neuropeptide circuits were genetically manipulated either by cell ablation or RNA interference (RNAi). We found that none of the investigated neuropeptides seems to be of equal importance for circadian locomotor rhythms as PDF. PDF had been previously shown to be necessary for rhythm maintenance in constant darkness (DD) as well as for the generation of morning (M) activity and for the right phasing of the evening (E) activity in entrained conditions. We now demonstrate that NPF and ITP seem to promote E activity in entrained conditions, but are clearly not the only factors doing so. In addition, ITP seems to reduce nighttime activity. Further, ITP and possibly also sNPF constitute weak period shortening components in DD, thereby opposing the effect of PDF. However, neither NPF or ITP, nor sNPF seem to be necessary in the clock neurons for maintaining rhythmicity in DD.
It had been previously suggested that PDF is released rhythmically from the dorsal projection terminals. Now we discovered a rhythm in ITP immunostaining in the dorsal projection terminals of the ITP+ clock neurons in LD, suggesting a rhythm in peptide release also in the case of ITP. Rhythmic release of both ITP and PDF seems to be important to maintain rhythmic behavior in DD, since constantly high levels of PDF and ITP in the dorsal protocerebrum lead to behavioral arrhythmicity.
Applying live-imaging techniques we further demonstrate that sNPF acts in an inhibitory way on few clock neurons, including some that are also activated by PDF, suggesting that it acts as signaling molecule within the clock network and has opposing effects to PDF. NPF did only evoke very little inhibitory responses in very few clock neurons, suggesting that it might rather be used as a clock output factor. We were not able to apply the same live-imaging approach for the investigation of the clock neuron responsiveness to ITP, but overexpression of ITP with various driver lines showed that the peptide most likely acts mainly in clock output pathways rather than inter-clock neuron communication.
Taking together, I conclude that all investigated peptides contribute to the control of locomotor rhythms in the fruitfly Drosophila melanogaster. However, this control is in most aspects dominated by the actions of PDF and rather only fine-tuned or complemented by the other peptides. I assume that there is a high complexity in spatial and temporal action of the different neuropeptides in order to ensure correct signal processing within the clock network as well as clock output.
Tumor angiogenesis is essential for the growth of solid tumors as their proliferation and survival is dependent on consistent oxygen and nutrient supply. Anti-angiogenic treatments represent a therapeutic strategy to inhibit tumor growth by preventing the formation of new blood vessels leading to starvation of the tumor. One of the best characterized anti angiogenic therapeutics is the monoclonal antibody bevacizumab (Avastin), which targets and neutralizes VEGF leading to disruption of the VEGF signaling pathway. Until today, bevacizumab has found its way into clinical practice and has gained approval for treatment of different types of cancer including colorectal cancer, non-small cell lung cancer, breast cancer and renal cell carcinoma. Signaling of VEGF is mediated through VEGF receptors, mainly VEGFR2, which are primarily located on the cell surface of endothelial cells. However, there has been evidence that expression of VEGF receptors can also be found on tumor cells themselves raising the possibility of autocrine and/or paracrine signaling loops. Thus, tumor cells could also benefit from VEGF signaling, which would promote tumor growth. The aim of this study was to investigate if bevacizumab has a direct effect on tumor cells in vitro. To this end, tumor cell lines from the NCI-60 panel derived from four different tumor types were treated with bevacizumab and angiogenic gene and protein expression as well as biological outputs including proliferation, migration and apoptosis were investigated. Most of the experiments were performed under hypoxia to mimic the in vivo state of tumors. Overall, there was a limited measurable effect of bevacizumab on treated tumor cell lines according to gene and protein expression changes as well as biological functions when compared to endothelial controls. Minor changes in terms of proliferation or gene regulation were evident in a single tumor cell line after VEGF-A blockade by bevacizumab, which partially demonstrated a direct effect on tumor cells. However, the overall analysis revealed that tumor cell lines are not intrinsically affected in an adverse manner by bevacizumab treatment.
Besides the functional analysis of tumor cells, embryonic stem cell derived endothelial cells were characterized to delineate vascular Hey gene functions. Hey and Hes proteins are the best characterized downstream effectors of the evolutionary conserved Notch signaling pathway, which mainly act as transcriptional repressors regulating downstream target genes. Hey proteins play a crucial role in embryonic development as loss of Hey1 and Hey2 in mice in vivo leads to a severe vascular phenotype resulting in early embryonic lethality. The major aim of this part of the thesis was to identify vascular Hey target genes using embryonic stem cell derived endothelial cells utilizing a directed endothelial differentiation approach, as ES cells and their differentiation ability provide a powerful in vitro system to study developmental processes. To this end, Hey deficient and Hey wildtype embryonic stem cells were stably transfected with an antibiotic selection marker driven by an endothelial specific promoter, which allows selection for endothelial cells. ESC-derived endothelial cells exhibited typical endothelial characteristics as shown by marker gene expression, immunofluorescent staining and tube formation ability. In a second step, Hey deficient ES cells were stably transfected with doxycycline inducible Flag-tagged Hey1 and Hey2 transgenes to re-express Hey proteins in the respective cell line. RNA-Sequencing of Hey deficient and Hey overexpressing ES cells as well as ESC-derived endothelial cells revealed many Hey downstream target genes in ES cells and fewer target genes in endothelial cells. Hey1 and Hey2 more or less redundantly regulate target genes in ES cells, but some genes were regulated by Hey2 alone. According to Gene Ontology term analysis, Hey target genes are mainly involved in embryonic development and transcriptional regulation. However, the response of ESC-derived endothelial cells in regulating Hey downstream target genes was rather limited when compared to ES cells, which could be due to lower transgene expression in endothelial cells. The limited response also raises the possibility that target gene regulation in endothelial cells is not only dependent on Hey gene functions alone and thus loss or overexpression of Hey genes in this in vitro setting does not influence target gene regulation.
The discontinuous mountain permafrost zone is characterized by its heterogeneous distribution of frozen ground and a small-scale variability of the ground thermal regime. Large parts of these areas are covered by glacial till and sediments that were exposed after the recession of the glaciers since the 19th century. As response to changed climatic conditions permafrost-affected areas will lose their ability as sediment storage and on the contrary, they will act as source areas for unconsolidated debris. Along with modified precipitation patterns the degradation of the discontinuous mountain permafrost zone will (temporarily)
increase its predisposition for mass movement processes and thus has to be monitored in a differentiated way.
Therefore, the spatio-temporal dynamics of frozen ground are assessed in this study based on results obtained in three glacier forefields in the Engadin (Swiss Alps) and at the Zugspitze (German Alps). Sophisticated techniques are required to uncover structural differences in the subsurface. Thus, the applicability of advanced geophysical methods is tested for alpine environments and proved by the good 3D-delineation of a permafrost body and by the detection of detailed processes in the active layer during snow melt. Electrical resistivity tomography (ERT) approaches (quasi-3D, daily monitoring) reveal
their capabilities to detect subsurface resistivity changes both, in space and time. Processes and changes in regard to liquid water content and ice content are observed to exist at short distances even though the active layer is not subject to a considerable thickening
over the past 7 years. The stability of the active layer is verified by borehole temperature data. No synchronous
trend is recognized in permafrost temperatures and together with multi-annual electrical resistivity data they indicate degradation and aggradation processes to occur at the same time. Different heat transfer mechanisms, especially during winter, are recognized by means of temperature sensors above, at, and beneath the surface. Based on surface and borehole temperature data the snow cover is assessed as the major controlling factor for the thermal regime on a local scale. Beyond that, the debris size of the substrate, which modifies the snow cover and regulates air exchange processes above the ground, plays a crucial role as an additional buffer layer. A fundamental control over the stability of local permafrost patches is attributed to the ice-rich transient layer at the base of the active layer. The refreezing of melt water in spring is illustrated with diurnal ERT monitoring data from glacier forefield Murtèl.
Based on these ERT and borehole temperature data a conceptual model of active layer processes between autumn and spring is developed. The latent heat that is inherent in the transient layer protects the permafrost beneath from additional energy input from the surface as long as the refreezing of melt water in spring prevails and sufficient ice is build up each spring. Permafrost sites without a transient layer show considerably higher
temperatures at their table and are more prone to degradation in the years and decades ahead. As main investigation area a glacier forefield beneath the summits of Piz Murtèl and Piz Corvatsch in the Swiss Engadin was chosen. It is located west of the well-known
rock glacier Murtèl. Here, a permafrost body inside and adjacent to the lateral moraine was investigated and could be delineated very well. In the surrounding glacier forefield no further indications of permafrost occurrence could be made. Geophysical data and temperature values from the surface and from a permafrost borehole were compared with long-term data from proximate glacier forefield Muragl (Engadin). Results from both
sites show a considerable stability of the active layer depth in summer while at the same time geophysical data demonstrate annual changes in the amount of liquid water content and ice content in the course of years.
A third investigation area is located in the German Alps. The Zugspitzplatt is a high mountain valley with considerably more precipitation and thicker snow cover compared to both Swiss sites. In close proximity to the present glacier and at a large talus slope beneath the summit crest ground ice could be observed. The high subsurface resistivity values and comparable data from existing studies at the Zugspitze may indicate the presence of sedimentary ice in the subsurface of the karstified Zugspitzplatt. Based on these complementary data from geophysical and temperature measurements as
well as geomorphological field mapping the development of permafrost in glacier forefields under climate change conditions is analyzed with cooperation partners from the SPCC project. Ground temperature simulations forced with long-term climatological data are modeled to assess future permafrost development in glacier forefield Murtèl. Results suggest that permafrost is stable as long as the ice-rich layer between the active layer and
the permafrost table exists. After a tipping point is reached, the disintegration of frozen ground starts to proceed rapidly from the top.
In this work, high-energy observables arising during different phases of SN explosions are studied with respect to their potential for allowing conclusions on suggested explosion scenarios and physical mechanisms that are thought to influence the evolution of SNe in a major way. The focus on selected observables at keV and MeV energies is motivated by the appearance of large degeneracies that can even be found for disparate scenarios in many wavelength regimes. Since the discussed emission in the high-energy regime is directly linked to nuclear processes being usually very distinct for different suggested physical models, the signatures at keV and MeV energies allow for meaningful comparisons of simulations with observations.
The experimental technique predominantly employed within the scope of this Thesis constitutes one subarea of femtochemistry: the time-resolved spectroscopy of photoin-
duced chemical reactions in the liquid phase by means of molecular signatures in the mid-infrared (MIR) spectral range. Probing transient vibrational states, i.e., dynamic
changes in the vibrational motion of specic molecular subunits or functional Groups allows for a distinct separation and assignment of measured signals to emerging molecular species. For this purpose, one key building block is indispensable, which most of the investigations carried out within the eld of femtochemistry have in common: a coherent light source delivering ultrashort laser pulses with a temporal duration that matches the femtosecond time scale on which molecular motions typically occur. This instrumentation enables the observation of photoinduced chemical reactions from the
starting point|the excitation event to the appearance of intermediates to the nal formation of stable photoproducts after several pico- or nanoseconds.
This work comprises the acquisition and presentation of time-resolved spectroscopic data related to promising molecular systems upon photoexcitation as well as the im-
plementation and testing of experimental optical techniques both for the presented experiments but as well for experiments conceivable in the future. In addition, linear spectroscopy measurements and quantum-chemical simulations on the emerging chemical species have been carried out. In so doing, the primary processes and subse-
quently emerging reaction products of two compounds on a timescale of several nanoseconds after photoexcitation have been elucidated in great detail. Both compounds, the
[Mn(CO)3(tpm)]+ (tpm = tris(2-pyrazolyl)methane) CO-releasing molecule (CORM) and the 5-diazo Meldrum's acid (DMA), are of academic interest but in addition belong
to molecular classes that might be utilized in the near future as dark-stable prodrugs under physiological conditions or that are already utilized in industrial chemistry procedures, respectively. The ndings of both studies gave rise to implement and examine two techniques for prospective transient absorption experiments, namely the shaping and characterization of ultraviolet (UV) laser pulses and the recording of two-photon excitation spectra. Beyond that, since each of the depicted experiments is based on the detection of weak transient absorption signals in the MIR spectral region, two dif-
ferent detection schemes, via chirped-pulse upconversion (CPU) on the one hand and via direct multichannel MCT detection on the other hand, have been juxtaposed at the
conclusion of this work. Since both techniques are suitable in femtosecond pump-probe measurements but thereby exhibit individual strengths and weaknesses, a comparative study provides clarication of the respective pros and cons. The first study introduced within this work investigates the complex photochemistry
of DMA, a photoactive compound used in lithography and industrial chemistry. By femtosecond MIR transient absorption spectroscopy covering several nanoseconds, the
light-induced dynamics and ultrafast formation of several photoproducts from the manifold of reaction pathways have been disclosed to form a coherent picture of the overall
reaction scheme. After UV excitation of DMA dissolved in methanol to the second excited state S2, 70% of excited molecules relax back to the S0 ground state. In compet-
ing processes, they can either undergo an intramolecular Wolff rearrangement to form ketene, which reacts with a solvent molecule to an enol intermediate and further to carboxylate ester, or they rst relax to the DMA S1 state, from where they can isomerize to a diazirine. The third competing reaction channel, having the lowest quantum efficiency with respect to the rst two channels, is the formation of a singlet carbene out of the S1 state. From there an ylide can arise or, via an intersystem crossing, the triplet form of the carbene follows. Whereas the primary reaction steps occur on a picosecond timescale, the subsequently arising intermediates and stable photoproducts are formed
within a few hundreds to thousands of picoseconds. For a reliable identication of the involved compounds, density functional theory calculations on the normal modes and
Fourier-transform infrared spectroscopy of the reactant and the photoproducts in the chemical equilibrium accompany the analysis of the transient spectra. Additional experiments in ethanol and isopropanol led to slight spectral shifts as well as elongated time
constants due to steric hindrance in transient spectra connected with the ester Formation channel, further substantiating the assignment of the occurring reaction pathways and photoproducts.
The study demonstrated that the combination of linear and time-resolved spectroscopic measurements in conjunction with quantum-chemical calculations constitutes a powerful tool to unravel even highly complex photoreactions exhibiting multiple consecutive intermediate states within parallel reaction pathways. Although some of the individual reaction steps, for example the ketene formation via Wolff rearrangement, have been observed on ultrashort time scales before, this work encompassed the Observation of the whole set of appearing photoproducts of DMA in different alcohol solutions within several nanoseconds. In this sense, the ultrafast photochemistry of DMA represents a prototype example for a multisequential reaction scheme, elucidated by the capabilities of femtosecond MIR spectroscopy.
With a modified instrumentation concerning amongst others the system delivering the fundamental laser pulses or the generation of the UV pump pulses, the next ob-
jective within this work was to elucidate the primary processes upon UV Irradiation of a manganese tricarbonyl CORM in aqueous environment. The time-resolved
experiment was performed with two different pump wavelengths and furthermore supported by linear spectroscopy methods and time-dependent density functional theory (TDDFT) calculations on the excited states as well as DFT calculations on the ground
states. The measurements revealed that irradiating the compound with UV excitation pulses primarily leads to ultrafast photolysis of one CO ligand. Geminate recombination may occur within one picosecond but it remains a minor process as the photolyzed CO
group is liberated and the unoccupied coordination site is predominantly filled by an incoming solvent molecule. There was no evidence for hot CO bands, i.e., the remaining CO ligands|in the dicarbonyl photoproduct as well as in the intact CORM are not vibrationally excited through the UV excitation of the CORM. According to this, the excess energy merges into low-frequency vibrational modes associated with the molecule as a whole. Since studies on a macroscopic scale at irradiation times of several minutes prove that UV irradiation eventually leads to the release of two or even all three CO
ligands, further loss of CO most likely necessitates manganese oxidation or another interaction with light. To clarify the latter, a consecutive UV pulse was employed in order to excite the photoproducts subsequent to the initial pump interaction. However, the data obtained was not instructive enough to denitely exclude the manganese oxidation being responsible for the loss of further CO groups. Besides the exchange of a CO Group by a solvent molecule or the geminate recombination, the employment of two different excitation wavelengths in combination with ndings derived from the TDDFT calculations suggested another reaction process, namely the possibility that the excitation does not lead to any bond cleavage at all. As the CORM under investigation is tissue-selective and cytotoxic against cancer cells, knowledge of these rst photoinduced reaction steps is essential for a full understanding of its biological activity. Inspired by these two studies, experimental techniques for prospective transient absorption measurements have been implemented and tested within preparative measure-
ments. First, in the course of a UV-pump-MIR-probe experiment with specically tailored excitation pulses, one could pursue the aim of coherently controlling the outcome of a photoreaction in the liquid phase. Out of the rich photochemistry of DMA the vibrational signature of a particular molecular species might thereby serve as a feedback signal, which is a central part of a learning loop that adaptively determines the pulse shape that steers the quantum mechanical system upon photoexcitation into a desired direction. This motivated the installation and testing of devices by means of which the shaping and characterization of ultrashort laser pulses in the UV could be performed. Second, motivated by the biological applications of CORMs, one can imagine a scenario where a certain amount of CORMs is deposited inside cancerous tissue. Since the activation of CO loss by means of UV pulses is not possible due to the absorption characteristics of biological tissue, the simultaneous excitation via two photons from the visible spectral regime seems appealing. However, success or failure of such an application depends on whether the deposited compound efficiently absorbs two photons simultaneously, i.e., whether the two-photon absorption cross section is large enough. Therefore, a setup to record two-photon excitation spectra under full consideration of
the crucial laser pulse parameters like the pulse duration, energy and central wavelength was arranged and tested. The rst results were obtained with a commercially available reference system (Mn2CO10) but the setup as well as the described measurement and
data analysis procedure can easily be applied to record the two-photon absorption cross section of more promising molecular systems. Third, as the detection of probe pulses
in the MIR spectral region is part of each time-resolved measurement throughout this thesis, a comparison between the newly established technique of CPU and direct multi-
channel MCT detection is presented by means of pump{probe experiments on Mn2CO10 and Co4CO12 with a 1 kHz shot-to-shot data acquisition. It was shown that the CPU detection technique scores with its high spectral resolution and coverage of the easy-to-handle and more cost-effective CCD detectors. On the other hand, in the course of the additional nonlinear upconversion process intensity fluctuations of the chirped fundamental pulses are transferred to the probe spectrum in the visible regime. This entails a lower signal-to-noise ratio than the direct MCT detection, which can be compensated by an additional normalization procedure applied to the CPU probe pulses. As a consequence, the CPU detection scheme offers more flexibility for future investigations
employing MIR probe pulses. This is of great importance for many applications within the presented eld of femtochemistry as a huge variety of time-resolved investigations on a multitude of systems in the liquid phase is based on the detection of weak transient
absorption signals in the MIR spectral region.
In this work, femtosecond laser pulses are used to launch optical excitations on different nanostructures. The excitations are confined below the diffraction limit and propagate along the nanostructures.
Fundamental properties of these ultrashort optical near fields are determined by characterizing the far-field emission after propagation with a setup developed for this task. Furthermore, control of the nanooptical excitations' spatial and temporal evolution is demonstrated for a designed nanostructure.
Bacterial mastitis is caused by invasion of the udder, bacterial multiplication and induction of
inflammatory responses in the bovine mammary gland. Disease severity and the cause of disease are
influenced by environmental factors, the cow’s immune response as well as bacterial traits. Escherichia coli (E. coli) is one of the main causes of acute bovine mastitis, but although pathogenic E. coli strains can be classified into different pathotypes, E. coli causing mastitis cannot unambiguously be distinguished from commensal E. coli nor has a common set of virulence factors
been described for mastitis isolates. This project focussed on the characterization of virulence-
associated traits of E. coli mastitis isolates in comprehensive analyses under conditions either
mimicking initial pathogenesis or conditions that E. coli mastitis isolates should encounter while entering the udder. Virulence-associated traits as well as fitness traits of selected bovine mastitis or faecal E. coli strains were identified and analyzed in comparative phenotypic assays. Raw milk whey was introduced to
test bacterial fitness in native mammary secretion known to confer antimicrobial effects.
Accordingly, E. coli isolates from bovine faeces represented a heterogeneous group of which some
isolates showed reduced ability to survive in milk whey whereas others phenotypically resembled
mastitis isolates that represented a homogeneous group in that they showed similar survival and
growth characteristics in milk whey. In contrast, mastitis isolates did not exhibit such a uniform phenotype when challenged with iron shortage, lactose as sole carbon source and lingual
antimicrobial peptide (LAP) as a main defensin of milk. Reduced bacterial fitness could be related to LAP suggesting that bacterial adaptation to an intramammary lifestyle requires resistance to host
defensins present in mammary secretions, at least LAP.
E. coli strain 1303 and ECC-1470 lack particular virulence genes associated to mastitis isolates. To find out whether differences in gene expression may contribute to the ability of E. coli variants to cause mastitis, the transcriptome of E. coli model mastitis isolates 1303 and ECC-1470 were analyzed to
identify candidate genes involved in bacterium-host interaction, fitness or even pathogenicity during bovine mastitis.
DNA microarray analysis was employed to assess the transcriptional response of E. coli 1303 and
ECC-1470 upon cocultivation with MAC-T immortalized bovine mammary gland epithelial cells to
identify candidate genes involved in bacterium-host interaction. Additionally, the cell adhesion and invasion ability of E. coli strain 1303 and ECC-1470 was investigated. The transcriptonal response to the presence of host cells rather suggested competition for nutrients and oxygen between E. coli and MAC-T cells than marked signs of adhesion and invasion. Accordingly, mostly fitness traits that may also contribute to efficient colonization of the E. coli primary habitat, the gut, have been utilized by the mastitis isolates under these conditions. In this study, RNA-Seq was employed to assess the bacterial transcriptional response to milk whey.
According to our transcriptome data, the lack of positively deregulated and also of true virulence-associated determinants in both of the mastitis isolates indicated that E. coli might have adapted by other means to the udder (or at least mammary secretion) as an inflammatory site. We identified traits that promote bacterial growth and survival in milk whey. The ability to utilize citrate promotes fitness and survival of E. coli that are thriving in mammary secretions. According to our results, lactoferrin has only weak impact on E. coli in mammary secretions. At the same time bacterial determinants involved in iron assimilation were negatively regulated, suggesting that, at least during the first hours, iron assimilation is not a challenge to E. coli colonizing the mammary gland. It has been hypothesized that cellular iron stores cause temporary independency to extracellular accessible iron. According to our transcriptome data, this hypothesis was supported and places iron uptake
systems beyond the speculative importance that has been suggested before, at least during early
phases of infection. It has also been shown that the ability to resist extracytoplasmic stress, by oxidative conditions as well as host defensins, is of substantial importance for bacterial survival in mammary secretions.
In summary, the presented thesis addresses important aspects of host-pathogen interaction and
bacterial conversion to hostile conditions during colonization of the mastitis inflammatory site, the mammary gland.