expression was found in matched samples from the kid- ney and the prostate (Table 2), while it was absent in most samples from the other seven tissues. Lipophilin B expres- sion was also detectable in one out of 25 gastric, one out of 24 lung, and two out of 25 rectal tumors, but not in the corresponding normal samples (data not shown). The expression pattern of mammaglobin A in malignant and normal samples from different tissues was in general highly concordant to that of lipophilin B (compare e.g. Figure 2A and 2B for breast tissue), except for kidney and prostate samples, where only lipophilin B but not mam- maglobin A was expressed. The two genes exhibited an identical pattern of differential expression (i.e. both down-, up- or non-regulation) in the majority of matched pairs from the breast (78%), uterus (78%), ovaries (56%), and the uterine cervix (100%) (Table 2), without marked disparities (i.e. no cases with one gene up- and the other down-regulated) in the remaining cases. According to the Spearman rank correlation test, co-expression of mamma- globin A and lipophilin B was highly significant in breast, uterine and ovarian tissues (each p < 0.001) but failed to reach significance in cervical tissues due to the small sam- ple size (n = 2). A very interesting finding was that mam- maglobin A and lipophilin B were both up-regulated in the same one out of 25 gastric, and the same two out of 25 rectal tumor samples, in which expression of the two genes was detectable. However, this was not the case with lung tumors, in which mammaglobin A and lipophilin B were each expressed in one out of 24, but not in the same sample (data not shown). No correlation was found between the (co)expression pattern of mammaglobin A and lipophilin B in various tumors and available clinico- pathological data.
Several studies have used r-value cut-offs ranging between 0.6 and 0.8 to depict co- expression correlations (for example van Noort et al., 2004). However, different genes have different distributions of r-values, i.e. at a given cut-off some genes may correlate significantly with hundreds of genes while other genes may not correlate with any. Despite this, it is still possible that the latter may hold biologically relevant relationships. For example, the two transcription factors MYB33 (At5g06100) and MYB65 (At3g11440) regulate pollen and anther development, are expressed similarly, and are functionally redundant (Millar and Gubler, 2005). However, an r-value cut-off of 0.8 did not associate these genes transcriptionally (r-value 0.7; data not shown; Mutwil et al., 2008). To minimize this problem we chose to normalize the r-value distributions in the calculated Pearson correlation networks by using highest reciprocal rank (HRR) as they define the mutual co- expression relationship between two genes of interest. Using this approach the MYB33 and MYB65 were readily transcriptionally linked (average rank=2 using GeneCAT; Mutwil et al., 2008). With this approach we were also able to define a connection cut-off, or maximum number of connections, for a given gene. The importance of defining such cut-off is apparent when looking at the distribution of r-values among the data. For example, approximately 1500 genes are only expressed in pollen (estimated from GeneCAT; Mutwil et al., 2008). All of these genes are correlated with each other with an r-value of 0.8 and should therefore be connected to each other in a Pearson-based correlation network (Mentzen and Wurtele, 2008). However, it is virtually impossible to retain any information from such network structure through manual inspection. Instead we argue that displaying these genes in close network vicinities, which is achieved by the HRR-based network, is more useful. In addition, recent results indicate that correlation ranked networks produce sounder results than networks based on correlation co-efficients (Obayashi and Kinoshita, 2009).
Expression analysis using Cancer Profiling Arrays (CPAs) Mammaglobin A (SCGB2A2) and lipophilin B (SCGB1D2) expression were analyzed by dot blot analysis using Clontech's "Matched Tumor/Normal Array" (MTNA) and "Cancer Profiling Array I" (CPA) for each gene. The two expression arrays together contained 630 cDNAs synthesized from 309 human tumor and 309 matched normal tissue specimens, and 12 cDNAs from human metastases corresponding to 12 of the tumor/nor- mal pairs. The specificity of the mammaglobin A and lipophilin B hybridizations on the arrays was established by co-hybridization of two dot blots, containing spotted plasmid cDNAs of either mammaglobin A or lipophilin B (see Figure 2). The radiolabelled mammaglobin A probe efficiently hybridized only to mammaglobin A cDNA (not Abundant (co)expression of mammaglobin A (SCGB2A2) (1A) and lipophilin B (SCGB1D2) (1B) in breast tumors
12 fulfilled criterion (i)). To assemble GOI1, GOI2 and the BIDI lib with pBasic in a single step, we positioned the BpiI site on the DNA parts such that its recognition site was cleaved off and the compatible ends were exposed on the restricted DNA fragments. Neighbouring fragments hybridized via their compatible ends and a ligase sealed the nicks. By our design, the resulting assembly was seamless (Figure 2a). Since the final vector construct lacked the BpiI recognition site, restriction and ligation could be performed simultaneously (‘one-pot, one-step’ assembly), which greatly accelerated the procedure. The assembly of the co-expression library from the four DNA fragments in the presence of BpiI and T4 ligase was accomplished in 2.5 hours. The protocol is outlined in detail in the BIDI Promoter Kit-Manual, section 8. Separate digestion and purification steps of the vector backbone or the inserts were not necessary (fulfilled criterion(ii)) because an accidental re-ligation of a BpiI recognition site with a compatible DNA fragment restored a functional BpiI restriction site that could be cleaved again. By using typeIIS RE BpiI, the assembly of the co-expression library did not involve any PCR steps (fulfilled criterion (iii)), which, again, minimized the chance of accidental sequence errors. Thus, to summarize the one-pot assembly, the circular pBasic plasmid for the ‘one-pot, one-step’ assembly (pBasic_BpiI), the BIDI lib, and accordingly designed GOI1 and GOI2 are applied together with BpiI and the T4 ligase in one pot. By this, one generates the final co-expression library of GOI1 and GOI2 in one cloning step only. A detailed description of and instructions for the design of the GOIs and the BIDI lib can be found in the BIDI Promoter Kit-Manual, sections 5 and 6, as well as Table
The observed sensitization of GC-A by CNP/GC-B signaling in pituitary αT3-1 cells raised the significant question whether such a regulation may take place also in other cell types co-expressing GC-A and GC-B. A particularly interesting cell type in this regard is the vascular smooth muscle cell. ANP signaling plays a pivotal role for regulating vascular contraction/relaxation via cGMP-mediated pathways (Pandey 2005), and several reports suggested the expression of both ANP (GC-A) and CNP (GC-B) receptors in smooth muscle cells of the aorta (Nagase, Katafuchi et al. 1997; Steinmetz, Potthast et al. 2004; Potter, Abbey-Hosch et al. 2006). The latter conclusion, however, was primarily based on investigations of receptor gene expression, and convincing data indicating co-expression at the protein level have not yet been provided.
We reasoned that A. niger gene co-expression networks, which we made publicly available at the data repository FungiDB [ 6 , 10 ], could be mined for genes which are transcriptionally coupled with protein secretion and cit- ric acid production. Accordingly, we retrieved candidates which are co-expressed with genes involved in the TCA cycle (citrate synthase citA, fumarate reductase fumR, and isocitrate dehydrogenase idh2), and vesicle traffick- ing at the Golgi (alpha/beta subunits of the coat protein complex (COPI) copA/sec26, and COPII subunit sec13). The COPI and COPII complex mediate retrograde and anterograde vesicle trafficking between the Golgi and endoplasmic reticulum, respectively [ 26 , 27 ]. Note that all six query genes were also selected due to evidence of function based on wet-lab experimentation conducted in either A. niger or A. nidulans [ 28 ]. Interrogation of co- expression networks above the stringent 0.5 Spearman correlation coefficient threshold revealed that 259 genes were co-expressed with all six query genes. GO enrich- ment of this multi-gene sub-network suggested the TCA cycle and Golgi vesicle transport are transcriptionally coupled with various processes in A. niger, including oxoacid acid/carboxylic acid metabolism, microtubule cytoskeleton organization, hyphal growth, and responses to pH, amongst others (Additional file 1 ). A notable observation from the GO analysis was enrichment of genes for regulation of Arf protein signal transduction (p = 0.01) which included orthologues for S. cerevisiae Arf GEFs SEC7 (An07g02190) and GEA2 (An18g02490, Additional file 1 ). Manual interrogation of the sub- network also revealed a gene predicted to encode the orthologue for S. cerevisiae Arf GTPase activating pro- tein Age2 (An11g02650) co-expressed with all 6 query genes (according to A. niger nomenclature, we name these genes secG, geaB and ageB in A. niger, respectively). Based on the co-expression network, we hypothesised A. niger ageB, secG, and geaB genes may concomitantly impact protein secretion and organic acid synthesis in A. niger.
In case of E2, expression in E. coli was reported for crystal structure determination (Larsson et al., 2005). Expression of E3 was performed for characterization of the enzyme rather than enzyme production, as well (see above and Vuong and Wilson, 2009a; Vuong and Wilson, 2009b). For E4, in addition to reports where the E. coli expressed enzyme was used for protein characterization without statement of specific yields (Escovar-Kousen et al., 2004; Kostylev et al., 2012; Li et al., 2007b), one publication described yields of purified protein in the range of 10-25 mg/l (Zhou et al., 2004). Endoglucanase E5 was expressed in several experiments in E. coli, but only a recent study reports activities of approx. 18 U/ml*h on CMC for fermentation at lab scale (Yan et al., 2013). For exoglucanase E6, expression was performed without declaration of yields for protein characterization and implementation into cellulosomes (Caspi et al., 2008; Kostylev et al., 2014; Kostylev and Wilson, 2013; Kostylev and Wilson, 2011; Moraïs et al., 2010). Endoglucanase E1 is the only protein which has not been expressed successfully in E. coli to this day. Although the cellulases from T. fusca have been produced in E. coli before, most of the experiments were focused on fundamental studies of the enzymes rather than on production of high amounts of protein for cellulose degradation. This was solely reported for E5 by Yan and coworkers (2013). Furthermore, the examples mentioned above employed IPTG for expression of T. fusca cellulases. Thus, the constitutive system employing the prrn promoter described here could be interesting for further studies to estimate possible yields and activities and eventually provide an alternative route for cost efficient cellulase expression in E. coli at lower enzyme production expenses. This is in particular interesting for the expression of BglC, since low amounts of β-glucosidases are a bottleneck for more efficient cellulase mixtures (Juhász et al., 2005; Zhou et al., 2009a).
In order to better understand the differences in gene functions between the different condi- tions, we performed a gene enrichment analysis. We looked for GO [ 36 , 37 ], KEGG [ 38 ] and REACTOME [ 39 ] enriched terms in the significantly differentially expressed genes between the clinical conditions and various derived gene lists. This includes all gene lists used and gen- erated in this article (see in the Supplementary Archive tables/ and geneSets/ for all enrich- ments and gene lists). An overview of GO (BP) term enrichments of the comparisons of normal to earlyRA, arthralgia, OA and undifferentiated arthritis is shown in Fig 3 . More details about tools and strict filtering settings needed for a diagram fitting onto a single page are in the method section. The GO terms in a larger font therein were selected for their specifity for earlyRA and meaningfulness. For example, in the upper right cluster, the term ’vesicle-medi- ated transport’ might be interesting, but is enriched in the up-regulated genes of all four condi- tions. The term ’cell activation’ is specifically enriched in up-regulated genes in earlyRA, but the term itself is rather nonspecific. Taken the GO terms of DEGs in earlyRA together, there is specifically more expression for chromatin (lower right in Fig 3 ), coagulation factors (as also reported in several articles [ 40 – 42 ]), less activity of polymerase II (as can also be seen in section ’Different gene expression at different gene biotypes’), less muscle cell activity (see section ’Dif- ferent RA gene expression in men and women’ for a more detailed different view on that) and more antigen presentation (left side in Fig 3 ). Other patterns in this view are also interesting, like enrichment specifically for earlyRA and undifferentiated arthritis as these conditions are clinically quite close. For example, the Gene Ontology terms ’biological adhesion’, ’regulation of cell-cell adhesion’ and ’immunoglobulin production’ are enriched in earlyRA and UA, but not in OA and arthralgia (left in Fig 3 ). Unfortunately, there are only few samples for undiffer- entiated arthritis, which weakens the hints from these patterns. The complete enrichment lists
Livestock diets can be supplemented with dipeptides in order to promote opti- mal growth and wellbeing. Due to the dual role of arginine as building block for proteins and regulator of physiological functions, pronounced effects were ob- served after addition of β-aspartate-argi- nine (β-Asp-Arg) dipeptides to feed. Cur- rently, β-Asp-Arg is generated in vitro from the cyanobacterial storage polymer cyanophycin (CGP) via incubation with the cyanophycinase enzyme (CGPase) which are both produced in E. coli. Be- cause of the high costs and limited scala- bility, bioreactor-based production is commonly used for the synthesis of high- value but not for cost-sensitive products such as supplements for animal diets. Alternatively, recombinant low-value products can be produced in plants in an economic manner using existing agricul- tural infrastructure and farming prac- tices. We already established the produc- tion of CGP in plastids of tobacco and po- tato, yielding up to 9.4% of the dry weight (dw) in stably transformed plants. We also demonstrated that CGPases can be transiently co-expressed in the cytosol of
ERβ signifikant mit einem geringeren Gesamtüberleben in unifokalen Tumorfällen, nicht aber bei multifokalen Tumoren assoziiert war. Hervorzuheben bleibt, dass eine Co-Expression von LCoR, jedoch stärker von RIP140 und ERβ hoch signifikant mit einem geringeren Gesamtüberleben und kürzerem Rezidivfreien Überleben bei unifokalen Tumoren assoziiert war. Bezüglich RIP140 zeigten sich diese Ergebnisse konkordant mit den Daten der größeren Patientenkohorte der ersten vorliegenden Originalarbeit. In der Literatur wird hinsichtlich eines hoch exprimierten ERβ und den klinischen Auswirkungen kontrovers diskutiert. Zum einen konnte von Kim et. al. und Guo et. al. belegt werden, dass eine Überexpression von ERβ mit einem geringeren Rezidivfreien Überleben einhergeht (71), (71), wohingegen in Untersuchungen von Tan et. al. eine Assoziation mit einem günstigeren Rezidivfreien Überleben besteht (52). Die vorliegende Diskrepanz erklärt sich am ehesten durch verschiedene getestete ERβ-Isoformen, sowie durch unterschiedlich angewandte Methoden (Testung auf mRNA- oder Proteinebene). Der inhibierende Effekt auf Estrogen-vermittelte Signalkaskaden wird durch RIP140 stärker in Interaktion mit ERβ als mit ERα ausgeübt (51). Des Weiteren konnte beim Ovarialkarzinom beobachtet werden, dass eine ERβ-abhängige Induktion von RIP140 stattfindet (51).
the shape of proteins. As soon as the immune system produces neutralising antibodies, the target proteins have changed sufficiently to avoid antibody binding (MODROW et al. 2003). Nevertheless, studies of protecting vaccines against SIV indicate that antibody-mediated protection is possible (DESROSIERS et al. 1989). Sera from individuals infected with HIV have been analysed extensively for the presence of neutralising antibodies. Five human monoclonal antibodies 4E10 (STIEGLER et al. 2001), 1b12 (BURTON et al. 1994), F105 (POSNER et al. 1991), 2G12 (TRKOLA et al. 1996) and 2F5 (PURTSCHER et al. 1994) have been isolated out of the serum of inapparent patients and were found to be capable of neutralising a broad range of primary B-clade HIV isolates, however high titre are required for this protection (MASCOLA et al. 1997; LI et al. 1998; PARREN et al. 1999). Their epitopes include regions on gp41 (2F5 and 4E10), the CD4-binding site of gp120 (1b12) and parts of the carbohydrate moiety of gp120 (2G12; reviewed in CALARESE et al. 2003). Work in this thesis focused on the neutralising antibodies 2F5 and 2G12, which were generated and characterised by a team led by Hermann Katinger (TRKOLA et al. 1996). The neutralising antibody 2F5 of the isotype IgG3 was switched to an IgG1 by ligation of the 2F5 V H to an IgG1 constant region, since this isotype is known for longer half-live in humans (KUNERT et al. 2000). Both antibodies are produced by recombinant expression in Chinese
Co-voting could be implemented for any voting method, be it at the voting booth, by letter or electronically. As explained, however, the most simple way to proceed would be to have all Vote-holders vote electronically. This might also solve another problem, i.e. how to prevent the Vote-holders from being influenced if their identity is revealed. Chosen at random by an algorithm, the Vote-holders would ideally be informed of their right to vote by electronic messaging, be asked to retrieve the background information they need for their decision electronically, and finally to cast their vote electronically, too. Encryption should be planned in the same way as e-banking transactions, so that no one could find out who the Vote-holders are or interfere with their voting.
Despite the statistical, economic and financial importance of systemic co-jumps, the literature is still missing a formal test to be used as an effective tool for their detection. A vast literature 3 concentrates
on univariate jump tests applied to broad portfolios, such as stock indices. Progress on developing tests for common jumps in a pair of asset prices was started by Barndorff-Nielsen and Shephard (2003). They propose a way to separate out the continuous and co-jump parts of quadratic covariation of a pair of asset prices. Mancini and Gobbi (2012) develop an alternative threshold-based estimator of continuous covariation. Jacod and Todorov (2009) propose two tests for co-jumps, building on functionals which depend, asymptotically, on co-jumps only. Bandi and Ren` o (2016) propose a test for co-jumps based on non-parametric infinitesimal moments. Finally, Bibinger and Winkelmann (2015) develop a co-jump test using spectral methods. However, these methodologies apply to the case N = 2 assets only, and their generalization to the case N > 2 is non-trivial. Bollerslev et al. (2008a) propose a test for common jumps in a large panel (N → ∞) which is based on the pairwise cross-product of intraday returns. In empirical works, the detection of multivariate jumps is typically achieved with a simple co-exceedance rule (see, e.g., Gilder et al., 2014), according to which the co-jump test is the intersection of univariate tests. We fill this gap in the literature by introducing a novel testing procedure for “multi-jumps”, that is a simultaneous co-jump in a collection of assets, which naturally applies to the case N ≥ 2, with N finite. The proposed approach builds on the comparison of two types of suitably introduced smoothed power variations. High values of the test-statistics (which is asymptotically χ 2 (N ) under the null) signal
Zur Lehrbuchdiskussion Zu Felix & Co.
Mit Interesse habe ich die ebenso engagierte wie kurzweilige Diskussion im „Leserforum“ des Forum Classicum über die neuen Unterrichts- werke für den neuen bayerischen Lehrplan Latein verfolgt. Es ist gut zu sehen, dass die Schriftlei- tung der Zeitschrift einem für die Schulpraxis so elementaren Problemfeld entsprechenden Raum zugesteht. Im Leserforum des letzten Heftes (2/2005, 149ff.) ist allerdings ein Missgeschick unterlaufen, auf das ich Sie die Leser des Forum Classicum hinzuweisen bitte:
Im Netzwerk von Hempel & Co.
Von Ulrich Herrmann
Eine Sache wird eingefädelt
Mein Terminkalender hat am 11. März 1991 den Eintrag „14:30 Hempel“. Dieser 11. März ist in bester Erinnerung, weil Rüdiger Bubner (jetzt Heidel- berg) aus Anlass seines 50. Geburtstags zu einem festlichen Mittagessen eingeladen hatte, zu dem auch sein Lehrer Hans-Georg Gadamer und aus Tübingen Hans Mayer ihm die Ehre gaben. Aber „14:30 Hempel.“ Nach allem, was dann an Einträgen im zweiten Halbjahr 1991 und im Jahre 1992 folgt, muss es sich um dasjenige Treffen gehandelt haben, bei dem Wolfgang Hempel mich darüber informierte, dass sein Parteifreund Hinrich Enderlein – mir bekannt als dem langjährigen Tübinger FDP-Abgeordne- ten im Stuttgarter Landtag – als Wissenschaftsminister des Landes Bran- denburg in Potsdam eine Universität gründen bzw. die dortige Branden- burgische Landeshochschule (zu DDR-Zeiten Pädagogische Hochschu- le) und die Hochschule für Recht und Verwaltung (der DDR) in Babels- berg in eine Universität umgründen wolle.
Figure 2: Examples of highly reactive molecules that are isoelectronic
to carbon dioxide.
To overcome its thermodynamically low level, additional energy is required to activate the CO 2 molecule. The threefold reactivity (Figure 3) of CO 2 with a nucleophilic oxygen atom, an electrophilic carbon atom and a π system provides the chemist with many options. Likewise, a rich coordination chem- istry to metal centres has been reported for CO 2 [5,6]. A forth- coming path is the reaction of CO 2 to form energy-rich inter- mediates that can subsequently transfer the CO 2 molecule to target substrates . The use of efficient catalysts is often another requisite to direct the reaction pathways with high selectivity to yield the desired target products and to overcome kinetic limitations associated with certain slow elementary steps.
In einer zweiten, eigenständigen Zielsetzung soll im Rahmen der vorliegenden Arbeit die Dzyaloshinskii-Moriya-Wechselwirkung von epitaktisch gewachsenen Co/Ir- und Co/Pt- Grenzflächen untersucht werden. Wie ausführlich in dem anschließenden Abschnitt 4.2.1 dargelegt wird, stellen experimentelle Messungen der DMI an Proben, die durch Sput- terdeposition hergestellt wurden, die aus Berechnungen stammende Erkenntnis, dass die DMI der Co/Ir- und der Co/Pt-Grenzfläche ein unterschiedliches Vorzeichen aufweist, in Frage [43, 48–51]. So wird häufig auf eine abschwächende [43, 49–53] Wirkung beider Grenzflächen geschlossen, wenn Kobalt in einer Schichtsequenz zwischen Platin und Iridi- um aufgebracht wird. Im Gegensatz dazu liefern Ab-initio-Berechnungen der DMI eines idealen Co/Ir(111)- [36, 37, 54, 55] oder Co/Pt(111)-Systems [36, 38, 54, 55, 107–111] über- einstimmend eine negative DMI für Co/Ir(111) und eine positive DMI für Co/Pt(111), sodass eine verstärkende Wirkung in einer Ir/Co/Pt-Schichtsequenz abgeleitet wird [36, 37, 54, 55, 110]. Obgleich die Resultate der Ab-inito-Berechnungen methodenabhängigen Abweichungen unterliegen und auch die Experimente im Betrag variierende Werte mes- sen, welche durch von Anlage zu Anlage unterschiedliche Grenzflächenbeschaffenheiten der Proben erklärt werden können, lässt sich an dieser Stelle eine Unstimmigkeit ableiten. Es ist in einem naiven Bild zu erwarten, dass beispielsweise eine Materialdurchmischung die DMI im Betrag reduziert, nicht aber eine relativ große DMI mit umgekehrten Vorzei- chen verursacht. Da die Situation in der Literatur komplex ist, soll diese zunächst in den beiden folgenden Abschnitten für Messungen und Berechnungen der DMI aufgearbeitet werden. Aus dieser Aufarbeitung leitet sich die Notwenigkeit ab mit einem Experiment, das die DMI einer nahezu idealen Co/Ir- und Co/Pt-Grenzfläche untersucht, Klarheit zu schaffen.