42 It can be noted that the viability of cells in the uninfected spinner flask also continuously decreased throughout the experiment. This decrease in cell viability was unexpected. Next, for the SDS- PAGE analysis, 10 μl of cell lysates from different spinner flasks and time points were loaded and separated, followed by the western blot procedure. Fig.3.7 (B) shows that there was no detection of any band for the period of 0 - 96 h.p.i for the MOI= 0.04 and for the MOI= 0.8 infection until 48 h.p.i. The bands of GFP protein (27 kDA) for the samples, taken at 72 hours and 96 hours post- infection of both MOI= 0.4 and MOI= 0.8 can be seen near the 30 kDA position of a pre-stained marker in fig. 3.7 (C). A band in the GFP lysate used as a positive control also can be seen in both images which confirms the that the western blot procedure had worked technically. For the quantification of proteinexpression, GFP signals were normalized to intensities obtain from analysed PVDF membranes using Image J software, in which the same area of the position of each sample was considered. The chart of GFP intensities of each sample can be seen in fig.3.7 (D). It can be seen a significant increase in GFP expression at 72 - 96 hours post-infection for all MOIs. For MOI= 0.8, GFP expression increases considerably at 96 hours compared to other samples. It needs to be noted that as the virus amount increases, so does the GFP expression signal. However, during the experiment, a low level of whitish material was observed in all spinner flasks, including uninfected flask. There was no specific indication for a biological contamination detected. Furthermore, it was also clear from the uninfected cell profile of 0 - 96 hours that cells seem not good in the environment condition of spinner flask either due to shear forces or some other reasons.
Some experiments have also been conducted using animal knock-out models. Batias et al. demonstrated in jun-d-/-mice that sterile mutants possessed Sertoli-cell-only tubules in which connexin43 content was either reduced or undetectable (Batias et al., 1999). The testes of connexin43-null mutant fetuses did not display normal proliferation and differentiation of germ cells (Roscoe et al., 2001). In 2007, Sridharan et al. generated a Sertoli cell-specific connexin43 knockout (SC-Cx43 KO) mouse in which connexin43 was not present in the Sertoli cells but was expressed normally in organs such as the heart. SC-Cx43 KO mice showed continuous Sertoli cell proliferation and delayed maturation in adulthood, indicating that connexin43 played key roles in Sertoli cell development (Sridharan et al., 2007b). Defamie et al. suggested that connexin43 mRNA and proteinexpression may be a functional marker of undifferentiated Sertoli cells (Defamie et al. 2003). Sertoli cells may also reveal a positive signal for the gap junctional protein in some SCO syndrome phenotypes as long as the Sertoli cells are fully differentiated (Anniballo et al., 2011). Because the dysregulation of connexin43 has been related to the malfunction of spermatogenesis, it can be predicted that reduction in connexin43 expression could possibly impair male fertility. Because no testicular alteration was reported in men with connexin gene mutations (Lai-Cheong et al., 2007), it has been postulated that altered connexin expression is a consequence of impaired testicular function rather than the cause of such testicular pathology (Pointis et al., 2010).
may be present in rat hepatocytes [ 16 , 17 ]. Several studies have demonstrated that human melatonin receptors may not only be desensitized but also even internalized after exposure to physiological or supraphysiological levels of melatonin in vitro [ 19 – 21 ]. Furthermore, melatonin exposure may influence genetic expression patterns in rat livers [ 22 ]. However, the spatial allocation of melatonin receptors in liver lobules is unknown, and regulatory changes in the expression of hepatic melatonin receptors after melatonin receptor agonist exposure have, to our knowledge, not been investigated in an in vivo model. Therefore, this study was designed to determine the hepatocellular and regional expression pattern, as well as mRNA and proteinexpression, of hepatic melatonin receptor subtypes MT 1 and MT 2 in male rats exposed to either physiologically generated melatonin following hemorrhage
In order to produce a huge number of recombinant proteins, clone libraries are grown in parallel high-throughput cultivation processes. Clone libraries of type 1 (Table 1-1) are the most applied today. Hence, different reports using such clone libraries are reviewed (with a focus on process-related parameters) to evaluate the characteristics of typical high-throughput cultivation processes (refer to Appendix A). Such processes are predominantly performed in microtiter plates (MTPs). These are preferred because of their high parallelism, low costs and easy handling (Duetz 2007; Fernandes and Cabral 2006; Gräslund et al. 2008; Hermann et al. 2003; Kumar et al. 2004; Leemhuis et al. 2009). A typical MTP-based production process for recombinant proteins consists of multiple cultivation steps, applying clone libraries as starting material (Berrow et al. 2006; Busso et al. 2008b; Gräslund et al. 2008). Such clone libraries are established via transformation of competent cells with plasmids harboring different target genes. Then, the transformation mix is normally plated on selective agar plates. In a next step, single colonies from those agar plates are picked and transferred to a preculture MTP. Sometimes the transformation mix is also directly applied to the precultivation step. From the preculture MTP often cryocultures are established, which in turn serve as long- term storage of the clones and also as starting material for further proteinexpression studies. In summary, there are three possible sources that are used as starting material for inoculation of precultures in high-throughput production processes: cryocultures, freshly transformed cells or single colonies from agar plates (Fig. 1-1).
has been described previously [ 47 ]. Additionally, differ- ences of expression in exponential phase, initiated by varying promoters, decreased during prolonged fermen- tation. In stationary phase (after 22 h) mCherry levels of tested inducible promoters are aligned. Yet, depending on the recombinant protein (e.g. amino acid composi- tion and post-translational modifications) or experiment outlook (e.g. short time setting or production of cell toxic products) promoters, which are inducible by conven- tional sugars or well-established inducers are of particu- lar interest. Although general knowledge of recombinant proteinexpression (e.g. therapeutics or metabolites) in lactobacilli steadily increases, efficiency and expression levels are not yet comparable to E. coli based systems. Recombinant gene expression usually exerts additional metabolic burden for the host. This often results in unstable genetic constructs, inhibition of cell growth and/or plasmid loss. Therefore, inducible expression sys- tems where transcription of the target gene can be tightly controlled are preferable. The presented expression systems might behave different in other Lactobacillus strains, the adaption of new promoter/repressor systems,
Folge einer Rezeptor-Interaktion mit β-Arrestin-2 ist häufig die Bindung von den hier behandelten Faktoren an den Rezeptor [102-104]. Daher werden diese Faktoren häufig als „akzessorische Bindungspartner“ bezeichnet. Die Rekrutierung dieser Faktoren zum Rezeptor verursacht eine Internalisierung des Rezeptors, also einen Transport weg von der Zellmembran, in zytosolische Kompartimente [104-107]. Dort können die Rezeptor-Protein-Komplexe zunächst gelagert werden [104, 119, 120]. Anschließend kann ein Abbau (Degradation) des Rezeptors erfolgen, was einer posttranslationalen Down-Regulation der Protein-Expression entspricht [104, 119, 120]. Alternativ kann jedoch ein Rücktransport und eine Reaktivierung des Rezeptors erfolgen („Recycling“) [104, 119, 120].
detecting the health status of the host. E. coli perceive even minimal changes in the host environment and adapt to the changed conditions by modifying their proteinexpression. In the case of severe intestinal inflammation, this adaptation is characterised by a strong inhibition of E. coli‘s central energy metabolism and an induction of several stress response proteins. Special emphasis should be placed on the upregulation of the NfuA protein, which is a general adaptation to severe inflammatory conditions. Upregulation of this protein may contribute to the high E. coli cell numbers found in IBD patients. Effective colonisation is also a prerequisite for E. coli Nissle’s probiotic function and therefore it is important to know, which factors promote this colonisation efficiency under in vivo conditions. Two, so far unknown fitness factors (YggE and Ivy) of E. coli Nissle were identified that enable this probiotic strain to effectively colonise the intestine under hostile inflammatory conditions. Moreover, indole was identified as a potential new probiotic agent of E. coli Nissle. Since the caecal indole concentrations in healthy individuals were higher than in inflamed mice, indole is hypothesised to contribute to the main therapeutic effect of E. coli Nissle: Maintenance of the remission state in patients with UC.
For all experiments, the strain E. coli Tuner(DE3) pRhotHi-2-EcFbFP was used. The Tuner(DE3) strains are lacY deletion mutants of BL21; therefore, no lactose permease is produced. The entry of IPTG is solely regu- lated by concentration-dependent diffusion which leads to adjustable levels of homogenous proteinexpression throughout the entire population [ 35 ]. A copy of the T7 polymerase gene controlled by the lacUV5 promoter is chromosomally integrated. The plasmid pRhotHi-2 is provided with a T7 promoter to allow inducible high- level expression of heterologous proteins [ 23 ]. The recombinant protein EcFbFP is an (FMN)-binding fluo- rescent protein, which is codon-optimized for the expres- sion in E. coli and whose maturation is, compared to GFP, independent of oxygen supply [ 36 ].
Interestingly, all genetically modified AML1-ETO mouse models remained healthy during their life-time (Buchholz et al., 2000; Higuchi et al., 2002; Rhoades et al., 2000) except for a knock-in strategy, which used Sca-1 regulatory elements for expressing AML1-ETO. In this Sca-1 knock-in model, AML1-ETO-expressing animals showed a myeloproliferative-like disorder (Fenske et al., 2004). Additional information has been provided by other approaches using reconstitution of lethally irradiated mice with bone marrow (BM) or hematopoietic stem/progenitor cells that were retrovirally transduced to express AML1-ETO. In these experiments, constitutive AML1-ETO activation led to perturbed erythro- and lymphopoiesis and induced the selective amplification of LKS (lineage - /c-Kit + /Sca-1 + ) cells, containing immature blood progenitors and HSCs (de Guzman et al., 2002; Schwieger et al., 2002). A major conclusion of this work was that AML1-ETO-expressing HSCs or blood cell progenitors no longer seemed to be restricted by the normal genetic controls regulating HSC pool size. Complementary experiments with AML1-ETO transduced CD34 + human bone marrow cells also confirmed the increase of immature progenitors seen with murine cells (Bäsecke et al., 2005; Mulloy et al., 2003). Although the above studies showed for the first time that aberrant AML1-ETO expression specifically increased LKS and CD34 + populations, the effect of AML1-ETO
revealed that Spry2 suppresses HGF-induced ERK activation and acts as inhibitor of proliferation (Lee et al. 2008). In a screening for candidate TSGs in human HCC Spry2 was identified as a potential TSG. Data obtained from array-based comparative genomic hybridization uncovered that Spry2 is downregulated and frequently deleted in HCC. Further investigastions in mice showed that in tumor cells stably expressing mutated Spry2 and β- Catenin the levels of ERK were elevated and genes involved in proliferation, apoptosis and angiogenesis were deregulated. Moreover, dominant negative Spry2 (Spry2Y55F) was able to induce liver tumorigenesis in this mouse model (i.e. in cooperation with activated β-Catenin- signalling) supporting the proposed role of Spry2 as TSG in HCC. Corroborating, in malignant tissue of HCC patients Spry2 was clearly decreased and loss of Spry2 expression was found in a subset of HCC with aggressive clinical behaviour (Lee et al., 2010). Promoter hypermethylation, LOH, as well as altered posttranscriptional regulation by Nedd4 accounted for the observed reduction in Spry2. Furthermore, concomitant activation of c-Met was measured in HCC samples, and in mouse models the function of Spry2 as regulator of c-Met signalling was demonstrated. Loss of Spry2 resulted in elevated ERK and AKT pathway activation via c-Met, accompanied with increased proliferation and sustained angiogenesis, leading to HCC development in vivo. Recent publications confirm the essential role of Spry2 in hepatic cancerogenesis. Spry2 inactivation was shown to contribute to AKT-driven hepatocarcinogenesis in the mouse (Wang et al. 2012). Tumorigenesis was significantly accelerated in absence of functional Spry2, likely via enhanced proliferation as well as glycolysis (caused by enhanced MAPK and pyruvate kinase M2 (PKM2) induction) In a large-scale study Spry2 was identified to be downregulated in about 90% of examined HCC cases (Song et al. 2012). Since patients whose primary tumors were negative for Spry2 displayed a higher probability of recurrence, the use of Spry2 as an independent prognostic predictor for the progression of surgically resected primary HCC was proposed.
Germany). The blots were blocked with blocking buffer containing 5% non-fat dry milk and 1% Tween in PBS at room temperature for 60 min and then incubated at room temperature for 120 min with monoclonal antibodies anti-GAPDH (Cat. # 5G4, Hy test Turku, Finland) or mouse Anti-β-Actin (Cat. #A5441, Sigma-Aldrich, Munich, Germany) (1: 100,000 dilution), mouse Anti-Hsp70 (Cat. # SPA-810Stressgen Bioreagents, Ann Arboor, USA) (1: 1000 dilution), and goat Anti-Rat IL-6 (Cat. # AF506, R&D System, Wiesbaden-Nordenstadt, Germany) (1: 200 dilution). After washing with blocking buffer, the membranes were incubated overnight at 4 °C with HRP-conjugated goat Anti-Mouse IgG (W402B, Promega, Mannheim, Germany) (1:5000 dilution) and HRP-conjugated donkey Anti-Goat IgG (sc-2033, Santa Cruz, Heidelberg, Germany) (1:2000 dilution). Immunodetection was visualized with Amersham ECL™ (Enhanced chemiluminescence) Western Blotting Detection Reagents (RPN2109, Amersham Bioscience, Freiburg, Germany) on Hyperfilm ECL autoradiography (RPN3103K, Amersham Bioscience, Freiburg, Germany). The bands were quantified by densitometric analysis software (Gelscan Standard V5.01) and the expression level was calculated by normalization to GAPDH or β-actin in the same sample.
Die Integration von Genexpressionsdaten aus Transkript- und Proteinhochdurchsatzmessungen hilft, funktionelle Beziehungen zwischen Genen, Transkripten und Proteinen zu verstehen. Ein bestimmter Ansatz, im Feld auch als Koexpressionsanalyse bezeichnet, nutzt verschiedene statistische Methoden, um paarweise Assoziationsmetriken zwischen Genen und Proteinen zu generieren. Bislang stützen sich Koexpressionsanalysen zumeist auf Transkriptionsdaten, da insbesondere dieser Typ Messdaten generiert und öffentlich verfügbar gemacht wurde. Jüngste Forschungsergebnisse legen jedoch nahe, dass die Expression von Proteinen stärker an die betreffende Genfunktion gebunden sind, als bisher angenommen. Diese kumulative Dissertation behandelt von mir untersuchte nicht-funktionale, genomische Effekte auf die Koexpression von mRNA, welche sich nicht auf die zu regulierenden Proteine auswirken. Diese Effekte beruhen zum überwiegenden Teil auf spezifischen genomischen Eigenschaften, wie der dreidimensionalen Chromatinstruktur und epigenetischer Zustände. Die genomische Architektur scheint direkte, weitreichende Effekte auf die mRNA-Koexpression zu haben, die beispielsweise aus stochastischen Fluktuationen zwischen offenen und geschlossenen Zuständen des Chromatins oder der Replikation von DNA hervorgehen könnte. Ein großer Anteil koexprimierter mRNAs proximal-liegender Gene besitzt keinen funktionalen Zusammenhang und wird auf Proteinebene gepuffert, wahrscheinlich aufgrund verschiedener posttranskriptioneller Mechanismen. Ich zeige diesen Effekt in menschlichen lymphoblastoiden Zelllinien und in differenzierten murinen Geweben durch Integration von öffentlich vorhandenen Omics-Datensätzen. Außerdem lege ich dar, wie ein Random Forest -Algorithmus Kovariationsmuster mitochondrialer Proteinen aus hochdimensionalen Interphasen-Chromatin-Daten extrahieren kann, um mögliche neue mitochondriale Proteine vorherzusagen. Schließlich zeige ich wie maschinelles Lernen die Analyse von Proteinkoexpression im Vergleich zu traditionellen statistischen Methoden, wie beispielsweise der Pearson Korrelationsanalyse, verbessern kann. Ich integriere 294 SILAC-Experimente, die im PRIDE -Archiv hinterlegt wurden und kalkuliere eine
Since MS2 peptide database search is the most commonly used method, and also the only one performed in the studies presented in this paper, the following description of identification algorithms will cover only this workflow. Several peptide database search engines have been developed to identify peptides in a high throughput manner, such as Mascot (Perkins et al., 1999), SEQUEST (Eng et al., 1994), and X!Tandem (Craig and Beavis, 2004). The overall designs of the different identification algorithms are similar; they all attempt to match experimental spectra with theoretical spectra generated from the sequences in a protein database. Before submitting the data to the search engine, several parameters must be defined, including database, mass accuracy, modification, etc. The algorithm usually has a particular built-in approach to evaluate the match between experimental and theoretical spectra. Users normally can optimize the output on the basis of different filtering thresholds. False discovery rate (FDR) (Balgley et al., 2007; Cargile et al., 2004; Jones et al., 2009; Wang et al., 2009) is usually employed as the criterion for the success rate of the identification. However, the algorithmic difference behind these platforms will not be discussed here; more detailed information can be found in the following informative review articles on this topic (Balgley et al., 2007; Kapp and Schutz, 2007; Martens and Apweiler, 2009; McHugh and Arthur, 2008; Nesvizhskii, 2007; Shadforth et al., 2005).
Based on previous studies, oligomerization is regarded as an important mechanism or exporting signal directing membrane proteins without a known cytosolic export signal into COPII vesicles for the specific destinations (Sato and Nakano, 2003; Springer et al., 2014). Normally, misfolded proteins tend to be retained in the ER by ER quality control (ERQC) (Hurtley and Helenius, 1989; Araki and Nagata, 2011), which could be assisted by chaperones via proper folding and targeting to the right localization. Therefore, this raises the hypothesis that chaperones are required to mediate the interaction of PIP2s (especially the highly abundant PIP2;1 and PIP2;2) with PIP1s lacking a known cytosolic export signal to exit the ER and reach the plasma membrane in certain cell types (in PIP2;1- and PIP2;2-expressing cells). Such an interaction possibly induces an allosteric functional conformation change via oligomerization and then results in the proper folding of PIP1s or vice versa, and forms the hetero-oligomer for exporting from the ER by facilitating the interaction of PIPs with a cargo receptor protein or other trafficking partners. For instance, the SNARE family (ZmPIP2;5 with SYP121, AtPIP2;7 with SYP121 and SYP61) is known to mediate vesicular trafficking and impact PIP2 targeting (Geelen et al., 2002; Besserer et al., 2012; Hachez et al., 2014). It would also be interesting to know whether these trafficking partners were also involved in the trafficking and/or stability of PIP1s in Arabidopsis.
fects and allow stable expression of the transgene. Indeed, ubiquitous chromatin opening elements (UCOEs) ( 16–18 ), scaffold/matrix attachment regions (S/MARs) ( 19 , 20 ) and antirepressor ( 21 ) elements have been reported to have a beneficial effect on proteinexpression levels and stability. Alternatively, it is possible to insert the gene of interest (GOI) into a pre-defined locus in the host cell by using re- combinase mediated cassette exchange (RMCE) methods ( 22 ). A well-chosen locus or so-called ‘hot-spot’ contain- ing euchromatin can insure long-term stable protein pro- duction levels. However, only a single copy is integrated us- ing this method which may limit the maximum achievable expression levels, although high antibody yields (up to 1 g /l) have been reported using the RMCE technology ( 23 ). In contrast to the previously described strategies, we aimed to create a system where we can combine the advantages of targeted integration in a hot-spot and the flexibility of ran- dom integration methods. We reasoned that large expres- sion vectors harboring whole loci containing euchromatin (hot-spots) will not be affected by positional effects and will confer high and stable expression levels. To this end, we ex- plored Bacterial Artificial Chromosomes (BACs) as expres- sion vectors for recombinant protein production in CHO cells. BACs have a large cloning capacity (200–300 kilobase (kb)) and therefore they can accommodate an entire locus with most if not all of the elements that control the expres- sion of a gene. Indeed, BACs have been widely used in the mouse transgenic field because they ensure positional effect independent and copy number dependent expression of a transgene ( 24–26 ). According to this, BAC vectors should be ideal tools applied to heterologous protein production in mammalian cells. BAC-based expression vectors containing carefully chosen loci should combine the beneficial effects of a stable genetic environment with the possibility to in- tegrate several vector copies in the cell host, thus boosting the transgene expression and making transgene amplifica- tion unnecessary.
When transfection of IVT-RNA is performed, one should consider upfront, whether the antiviral response will be helpful or detrimental to achieve the intended aim. For example, immunostimulatory reactions to IFNs can be helpful to boost the development of an immune response to RNA-based vaccines (Conry et al., 1995; Leitner et al., 1999; Kreiter et al., 2010), whereas the PKR-mediated translational shutdown apparently limits proteinexpression. The induction of apoptosis as a response to IVT-RNA precludes mRNA transfection of cells from any experiment that requires living cells. However, cell death after IVT-RNA transfections could be avoided by siRNA-mediated knock-down of the IFN response genes, which enabled repetitive RNA transfections (Angel and Yanik, 2010). Another promising way to reduce the IFN response to IVT-RNA is make it “look self” by incorporation of modified nucleotides (Karikó et al., 2005; Hornung et al., 2006; Nallagatla and Bevilacqua, 2008). In higher organisms nucleotide modifications are introduced into RNA as an adaption to distinguish between self and foreign RNA. Thus, when simple NTPs are substituted with modified nucleotides during in vitro transcription, the resulting modified RNA is no longer recognized as “non-self” by cells. A reduced TLR3, 7 and 8 activation as well as a reduced cytokine production was reported when modified RNAs were used for transfection (Karikó et al., 2005). Even a partial substitution of only 25% of cytidine and uridine by 5-methylcytidine (m5C) and thiouridine reduces significantly the binding to TLR3, 7 and 8 as well as to RIG-I and thereby minimizes the IFN response of modified IVT-RNA (Kormann et al., 2011). Similarly, when uridine is totally substituted by pseudouridine (ΨU) in mRNA, activation of PKR is prevented and thereby translation is increased (Anderson et al., 2010). In line with that, a total replacement of cytidine and uridine by 5 methylcytidine and pseudouridine was effectively blocking the IFN response, together with the anti IFN recombinant protein B18R in the medium. Thus they could use IVT-RNA to reprogram fibroblasts to induced pluripotent stem cells (Warren et al., 2010).
It is obvious that the folding process should ideally be studied under native conditions. However, the folding mechanism is experimentally difficult to address in the complex context of a living cell. Recent developments of cell-free proteinexpression systems circumvent these constraints. These cell-free systems comprise the essential components for transcription and translation [ 20 – 22 ] of membrane proteins [ 23 – 25 ]. A proper folding milieu for integral mem- brane proteins is provided by nanodiscs, which are discoidal lipid bilayers wrapped by two amphiphilic membrane scaffold proteins in a belt-like configuration ( Fig 1b ). Nanodiscs repre- sent the unique advantage of investigating one folding experiment in two different approaches synchronously, in batch and on the surface. This allows the quantitative and qualitative control of the cell-free proteinexpression for every experiment on a level, which cannot be provided by classical bilayer models like liposomes or lipid monolayers. Elucidating the folding mechanism of membrane proteins requires a method that owes not only molecular sensitivity to resolve the structural changes of the nascent polypeptide chain but also temporal resolution to trace the folding dynamics. Within these boundary conditions, IR spectroscopy has a proven record for molecular studies where structural changes have been monitored at utmost temporal reso- lution and spatial sensitivity. Furthermore, exploiting plasmonic effects adds selectivity to IR spectroscopy. Here, Surface-Enhanced Infrared Absorption Spectroscopy (SEIRAS) [ 26 – 30 ] exclusively monitors processes that take place in the biomimetic membrane because the enhancement exerted by a rough gold surface is limited to only about 10 nm from the plasmo- nic gold layer [ 31 ] to which the membrane is tethered to (solid-supported membrane). This length scale competes with the typical thickness of 5 nm of a biological membrane. In the pres- ent work, we combine a cell-free
Die Menge an MBP wird durch die Induktion nicht beeinflusst. Die Banden der Pellet- Fraktion stellen alle Proteine des Cytoplasmas sowie in der Membran verankerte und unlösliche Proteine dar. Im Gegensatz dazu befinden sich in der Fraktion des osmotischen Schock-Fluids alle löslichen Bestandteile des periplasmatischen Raums, die durch den Zellaufschluss freigesetzt wurden. Diese Proteine wurden anschließend auf die Chromatografiesäule gegeben. Der Hauptanteil der degradierten Proteine befindet sich nach dem Zellaufschluss im Zellpellet. In den meisten Fällen (vergl. S. 60, S. 61) befindet sich in der Fraktion des osmotischen Schock-Fluids eine einzige Bande bei 64 kDa, die dem Fusionsprotein entspricht. Die Konzentration ist jedoch sehr gering und das Protein wird nur durch die sensitive Silberfärbung sichtbar gemacht. Der Hauptanteil der E. coli-Proteine befindet sich im Zellpellet, wodurch der Anteil an Gesamtprotein im Periplasma sehr gering ausfällt. In allen Fällen weisen die Elutionsfraktionen trotz Silberfärbung kein Fusionsprotein auf. Die Detektion wurde erst durch eine Aufkonzentrierung der Proben ermöglicht (vergl. 3.2.3).
In our microarray analysis, we used RNA-extractions from whole-body adult flies raised under common lab conditions. This has several implications for the detection of expression variation. First, genes whose expression is limited to certain tissues can escape detection simply due to their low expression levels compared to other more broadly expressed genes. While this problem is inherent to the approach of studying whole-body gene expression, higher resolution of tissue-specific expression variation can be tackled by performing expression analyses of single tissues only. Second, using four-to-six day-old flies limits our analysis of expression to young adult flies, and does not allow for inferences about other developmental stages. Gene expression varies greatly between developmental stages, such that expression variation during other stages will differ from our findings for adult flies. In addition, immature fly stages can also be targets of selection (Sgrò et al. 2010, Frankel et al. 2011), and adaptation could at least in part occur through gene expression changes that occur over the course of development. Such expression changes could be missed when focusing on adult flies. Gene expression changes in specific developmental stages can be unraveled by analyzing expression within and between the populations for larval or pupal stages. Third, using flies raised under common lab conditions, we do not detect genotype-by-environment interactions that affect gene expression. The African and the European population are presumably adapted to their respective natural environments and gene expression results from a combination of genetic and environmental factors. Thus, gene expression levels might differ depending on the environment under which the flies have been raised. However, using common conditions for both populations ensures that all measured expression variation is due to genetic components. Measuring the environmental component of gene expression can be accomplished by analyzing the same strains kept under differing environmental conditions, such as different temperatures, levels of humidity, or exposure to different pathogens.