Browsing by Subject "Sequenzanalyse <Chemie>"

Now showing 1 - 2 of 2
  • Thumbnail Image
    Publication
    Bioinformatische Analyse und funktionelle Charakterisierung von strukturellen Genvarianten in ADME-Genen in humaner Leber
    (2016) Tremmel, Roman; Zanger, Ulrich M.
    Pharmacogenetics is the study about inter-individual genetic variation that influences the response to drugs and other xenobiotics. A major part of this variation is due to hepatic drug metabolism with enzymes, transporters and receptors involved in the ab-sorption, distribution, metabolism and the excretion of drugs, xenobiotics and endoge-nous substances and collectively defined as ADME-genes. Genetic factors along with environmental and endogenous factors, including gender, age, inflammation processes and others are known to influence the expression and activity of ADME-genes. These influences can affect drug response, side effects or toxicity. According to newly published data, the human genome of any subject differs from a reference genome at 4.1 to 5.0 million positions. More than 99.9% of these differences are single nucleotide polymorphisms (SNP) or short insertions or deletions. Further-more, a person carries up to 2,500 structural variants, including copy number variations (CNV) affecting ~20 million bases (1000 Genomes Project Consortium et al., 2015). Thus structural variants affect more bases than SNPs. Per definition the CNVs are du-plicated or deleted DNA segments greater than 1kb and it was shown that they cover at least 12-30% of the human genome. Genome-wide studies investigating the function-ality of CNVs in the fruit fly, the mouse and in humans showed that there are genes whose expression is clearly affected by CNVs (dosage-sensitve), but also genes show-ing lower expression with increased copy number (dosage reversed) or genes without any expression alterations despite different copy number (dosage-insensitive). A prominent example of CNVs influencing drug metabolism is the phase I gene CYP2D6. Carriers of reduced or amplified gene copies show significantly altered ex-pression and enzyme activity levels and also a different drug metabolism of substrates like codeine (opioid) or tamoxifen (selective estrogen receptor antagonist) in compari-son to carriers with normal copy status of two. Genotyping of CYP2D6 gene copy num-ber may thus help to adjust drug dosage in a genotype dependent manner. In this work I investigated if further ADME-genes are affected by CNVs and if these variants have a functional impact on the expression phenotype and drug metabolism. The distribution of CNVs in the most important ADME-genes (n=340) was investigated in three independent cohorts using CNV data in a public accessible database of ge-nomic variants (DGV; dgv.tcag.ca), processed SNP microarray data of paired samples of healthy (n=269) and tumor (n=351) liver tissue of the TCGA project (http://cancergenome.nih.gov/) and ADME-panel based exon next generation sequenc-ing (NGS) applied on 150 well documented human liver samples of an in-house cohort (IKP148). For the NGS data analysis a method was developed and optimized to esti-mate the relative copy number of the ADME genes or every single exon via the read depth. The results were validated using qPCR with specific TaqMan assays. RNA-sequencing data of 50 healthy TCGA liver samples, and normalised expression data from microarray experiments applied to lymphoblastoid cell lines (LCL) from the HapMap samples and the 150 human liver samples (IKP148) were used to analyse the association between CNVs and the mRNA expression. Furthermore, in the IKP148 liver samples protein and enzymatic activity levels were available or measured using West-ernBlot and mass spectrometry for selected ADME-genes. All pharmacologically important CNVs of phase I and phase II genes, including CYP2A6, CYP2D6, GSTM1, GSTT1, SULT1A1 and UGT2B17 could be confirmed in all datasets. CNVs which were known, but so far not functionally assessed were found in the phase I and II genes CES1, CYP2E1, CYP21A2, UGT2B15 and UGT2B28. In this work rare CNVs (<1%) were mainly found for transporters like ABCA2, SLC2A4 and SLC47A1. The analysis of the read depth in the IKP148 samples data revealed hybrid genes for CYP2A6 and CYP2D6 with their pseudogenes and allowed a fine mapping of the different alleles. The functional analysis further confirmed the positive association between CNVs and the mRNA expression of CYP2A6, CYP2D6, GSTM1, GSTT1, SULT1A1 and UGT2B17 in all three cohorts. The combination of all data from the NGS project in the IKP148 liver subjects, including SNP and CNV genotypes showed that 11% and 53% of the variability of CYP2A6 and CYP2D6 enzyme activity were explained by the genetic factors. In contrast the mRNA expression of the genes CES1 and CYP2E1 was not dependent of the CNV pattern in healthy liver tissue (IKP148 and TCGA) and lymphoblastoid cell lines. A detailed analysis of the protein and enzyme activity levels (chlorzoxazone-6-hydroxylation) of CYP2E1 confirmed the dosage-insensitivity in the IKP148 liver sub-jects. The dosage compensation can be principally explained by different mechanisms and could be tissue or tumor specific. Furthermore, CNV-linked genetic variants, altered miRNA regulation, incomplete inclusion of regulatory elements or coding sequences, hybridgenes, monoallelic expression, feedback loops or epigenetics could be factors which mask the CNV effect. In this work a haplotype analysis of the CYP2E1 region identified SNPs which were linked to the duplication and a reduced expression phenotype in persons with European ancestry. Using in silico prediction tools we found a relation of one of the linked SNPs in the 3’UTR with additional predicted miRNA bind-ing sites potentially regulating additional CYP2E1 gene copies. The CNV influence on the mRNA expression of the genes CYP21A2, UGT2B25 and UGT2B28 was inconsistent. Although CYP21A2 deletions were associated with a de-creased expression, gene duplications showed normal expression levels compared to samples with two copies. A significant influence of UGT2B28 CNVs was found in LCLs but not in human liver samples (IKP148 and TCGA). In total 7 of 17, 2 of 12 and 3 of 14 ADME genes showed a significant association between expression and CNV type in the IKP148, TCGA and LCLs of the HapMap samples, respectively. In the TCGA cancer tissue nearly all ADME-genes carry CNVs and in 30% of the genes a significant correlation was observed. With cooperation partners further polymorphisms and phenotypes of SULT1A1 and CYP2E1 were analyzed. CYP2E1: In this part of the thesis factors influencing the risk of developing differentiat-ed thyroid carcinoma (DTC) were investigated. Known risk factors for the progression of DTC are genetic and environmental factors, including ionizing radiations, previous thyroid diseases, and hormone factors. It has been speculated that dietary acrylamide intake correlates with the DTC formation. The acrylamide molecule is metabolized by CYP2E1 to the reactive carcinogenic glycidamide. The enzymatic reaction is probably dependent on the CYP2E1 genotype. Together with a cooperation partner (Prof. Dr. Landi, University of Pisa, Pisa, Italy) we investigated, whether CYP2E1 variants influ-ence the DTC risk. Prof. Landi and colleagues used a case-control-cohort and a haplo-type approach and observed a significant association between a tag-SNP rs2480258 (A allele), which covers variants in intron eight and the 3’UTR, and an increased DTC risk. In the human liver samples (IKP148) the rs2480258 genotypes were assessed using an imputation analysis and it was shown that particularly the A allele of the SNP reduce significantly the mRNA and protein expression and the enzyme activity. An in silico prediction for the molecular mechanism suggested that miR570 specifically down regulates the transcripts in carriers of the A allele. These results indicated that the inter-individual CYP2E1 activity as well as acrylamide (similar to glycidamide) influences the risk for DTC. SULT1A1: Methyleugenol, a secondary metabolite present in herbs such as basil or laurel is metabolized in humans by sulfation to a reactive product which can covalently bind to DNA. The resulting DNA adducts are mutagenic and can promote carcinogene-sis. Our cooperation partner Prof. Dr. Hans-Rudolf Glatt (German Institute for Human Nutrition Potsdam-Rehbruecke (DIfE), Nuthetal, Germany) had shown, that the meth-yleugenol metabolism takes place in the liver of mice and is mainly catalyzed by the phase II enzyme SULT1A1. To investigate these facts in humans, the methyleugenol DNA-adduct levels were measured by the cooperation partner in liver tissues (n=121; IKP148) using mass spectrometry. In this work the SULT1A1 protein levels were de-termined using western blot analysis and the relation between the DNA adducts as well as the SULT1A1 expression and the SULT1A1 CNVs was assessed. The SULT1A1 mRNA and protein expression were significantly correlated to the DNA adducts, e.g. higher SULT1A1 expression resulted in higher adduct levels. This emphasized the role of SULT1A1 in the in vivo metabolism in human liver samples. As mentioned above, there were individuals (IKP148) carrying one, two, three, four and five copies of SULT1A1. Deletions were found less frequent (4%) than duplications (36%). The CNVs were significantly associated with the SULT1A1 mRNA and protein expression. This result was consistent to previous studies investigating the association between SULT1A1 CNVs and enzyme activity. The methyleugenol DNA adduct levels were also significantly associated to the SULT1A1 copy number. Carriers of at least three gene copies exhibited a 2.8-fold higher DNA adduct level compared to donors carrying only one SULT1A1 gene copy. As a consequence this could mean that individuals with mul-tiple SULT1A1 copies reach faster, more often and more easily critical and ultimate adduct levels which increase the risk for developing cancer. Future studies should clari-fy whether methyleugenol intake as well as the individual SULT1A1 CNV make-up in-fluences the risk of cancer.
  • Thumbnail Image
    Publication
    Identification and analysis of a transcriptome of Douglas-fir (Pseudotsuga menziesii) and population structure inference using different next-generation sequencing techniques
    (2015) Müller, Thomas; Schmid, Karl J.
    Predictions assume severe changes in the climatic conditions in Central Europe in the coming decades. Longer periods of drought and less precipitation during summer are expected. Plants cannot change their habitat and have to adapt to the new conditions or their offspring has to colonize new ecological niches. Due to the long generation times in trees it is important to know if and how trees can cope with the expected climatic conditions. Forest managers already give thought to the composition of future forests, because they have to choose species and populations which have no or only few problems with the changed climate. Douglas-fir (Pseudotsuga menziesii) is a promising tree species for this purpose, because it is adapted to different habitats and climate zones in its natural distribution range in North America. The two main varieties, coastal and interior Douglas-fir, differ genotypically and phenotypically, e.g. in drought tolerance. Douglas-fir trees, mainly of the coastal variety, showed good growth performances in field trials in Germany. Hence, a research project called "DougAdapt" was designed to analyze and to link genotypic and phenotypic differences in several coastal and interior Douglas-fir provenances. In this project, trees from field trials and from greenhouse experiments were sampled. To analyze the genetic diversity of the provenances we first generated reference sequences. Even with modern and cost-efficient next-generation sequencing technologies it would be very expensive to decipher the ~ 19 gigabases of the Douglas-fir genome completely. An alternative to whole genome sequencing is transcriptome sequencing, in which only the coding regions of a genome are sequenced. The transcriptome sequencing, which was performed for the first time in Douglas-fir, resulted in a large number of putative unique transcripts (PUTs). Comparisons with published transcriptomes of other plant species showed that the PUTs represented the transcriptome of Douglas-fir comprehensively. As the sampled seedlings were part of a drought stress experiment and grew under controlled conditions, we were able to identify drought related candidate PUTs, which may be part of the trees response to drought. Furthermore, more than 27,000 previously unknown single nucleotide polymorphisms (SNPs) in Douglas-fir could be identified. SNPs can influence the phenotype of individuals, and they can be used for instance as markers or to analyze genetic diversity. The analysis of genetic diversity of Douglas-fir provenances and the search for genes which may be part of local adaptation were performed with a sequence capture experiment. In sequence capture only predefined regions of a genome are sequenced. We showed that sequence capture based on PUTs as target regions is applicable in species with large and mostly unknown genomes. The polymorphic drought related candidate PUTs showed higher genetic differentiation than the remaining genes. Nevertheless, none of them was among the candidate PUTs for positive selection, which in turn are probably part of the local adaptation of the trees. Despite a high level of gene flow between coastal and interior provenances, the SNP data showed genetic differentiation between both varieties but only very low differentiation between the coastal provenances. We also investigated if genotyping-by-sequencing (GBS) is a suitable method to detect polymorphisms in Douglas-fir and compared the results of two GBS experiments with the sequence capture. The genome is digested with one or several restriction enzymes in GBS. Afterwards, only fragments with a specific length are sequenced, which considerably reduces the part of the genome that is sequenced as well as the costs. The advantage compared to sequence capture is the possibility to sample more individuals at the same time with less effort and costs. We showed that a digestion with two restriction enzymes results in more SNPs with less missing data, compared to using only one restriction enzyme. Both GBS methods returned considerably less SNPs than the sequence capture. Nevertheless, it was possible to distinguish between southern interior, northern interior, and coastal provenances using SNP data of the GBS experiments. GBS, especially with two restriction enzymes, seems to be a promising approach to genotype a large number of Douglas-fir trees and to obtain SNPs at low costs, which can be used in several tasks like genome-wide association studies. A large amount of sequence data and SNPs were analyzed in this thesis. Together with phenotypic information, these data will be crucial for the analysis of useful traits in Douglas-fir, like drought tolerance. Furthermore, the results concerning the Douglas-fir genome and the genetic diversity of different provenances will be beneficial in breeding programs and association studies, which in turn can be helpful to choose the optimal provenances for a given location.