|(übersetzt):||The archaeal transcription machinery resembles a primitive version of the RNAP II system with its three promoter elements, the transription factors aTBP and TFB, and a 10-12 subunit RNAP. In contrast to initiation, only little is known about elongation of transcription or potential elongation factors. In this theses appropriate investigations were done, using a cell-free transcription system for Methanococcus thermolithotrophicus and employing the method of immobilised transcription. In particular, a TFIIS homologue protein was characterised as an archaeal elongationfactor, differences between early and mature elongation complexes were presented, and RNA-protein interactions within the RNA channel of a mature complex were investigated. Additionally, the analysis of the function of the elongation factor in transcriptional proofreading was analysed. During this work it was also investigated whether the RNAP of Mc. thermolithotrophicus was able to incorporate DNA-monomers as a substrate for transcription. Employing immobilised in vitro transcription, we demonstrated that the TFIIS homologue protein stimulated the intrinsic nuclease activity in paused Methanococcus elongation complexes. After removal of nucleotides, a constant release of dinucleotide hydrolysis products was observed, combined with backsliding of the RNAP on the DNA. Additionally, arrested complexes were reactivated in the presence of this protein. Due to the sequence homology to TFIIS and its similar biochemical properties, the protein was named transcriptionfactor S, TFS. TFS induces a hydrolysis-resynthesis mechanism in halted complexes with the release of a dinucleotide from the 3´-end of the nacent RNA, followed by fast resynthesis of the shortened transcript. Hydrolysis induction occured about every 30 seconds. In contrast to TFIIS, TFS was also able to stimulate the intrinsic nuclease activity of the RNAP in the abortive phase of transcription at registers 4 or 5. This led to the conclusion that TFS-mediated RNA hydrolysis may be different, because neither a complete RNA-DNA hybrid nor a 13 nt RNA were needed. Archaeal proofreading occured in a pre- and postincorporation manner, due to TFS-induced hydrolysis-resynthesis. Misincorporation of ribonucleotides was dependent on the chosen nucleotide, its concentration, temperature, and incubation time. Methanococcus RNAP quantitatively incorporated a false UMP instead of CMP, with a slowed incorporation rate of about 350 to 500 times. After misincorporation occured, these complexes arrested. RNase-footprinting of arrested complexes showed that, in contrast to RNAP II, they were not backtracked. TFS was able to reactivate these complexes, accompanied by the release of the misincorporated nucleotide within a dinucleotide hydrolysis product. This led to the conclusion that the 3´-end of the nascent RNA remained in the catalytic centre of the RNAP. In the course of these investigations we could demonstrate that DNA-monomers were used as substrate for transcription. Following dCMP- instead of CMP-incorporation, complexes did not arrest. Instead, the rate of extension was slowed down by ~15 to 20 times (when one dCMP was incorporated) to ~60 to 80 times (when two dCMP were incorporated in a series). TFS was able to minimize the amount of extended RNA when two dCMP were incorporated. Using Exo III-, KMnO4-, RNase A-, and T1-footprinting differences in early and mature elongation or elongation complexes could be determined. In contrast to mature complexes, the early complexes had upstream-translocated Exo III-boundaries even though they where transcriptionally competend. They also had expanded RNase A- and T1-protection patterns, an increased intrinsic nuclease activity and were more sensitiv to the lack of Mg2+. Following TFS-induced cleavage and backsliding, the Exo III-downstream boundaries were enlarged. This was most probably due to DNA binding of TFS. Using the technique of UV-crosslinking, the interactions between RNAP subunits and the first 5 to 7 nt RNA within the RNA-DNA-hybrid were determined in halte d Mc. jannaschii complexes. KMnO4-footprints revealed that the hybrid consisted of maximal 8 bp. Within the gradual transcription, the interactions between the RNA and RNAP sunbunits shifted from A´- to A´´-subunit. In halted RNAP II complexes four positions in the RPB1-subunit, which were in close contact with the two 3´-terminal RNA-position, were determined. An alignment with the A´-and RBP1-subunit revealed that the contact amino acids were conserved in Mc. jannaschii. This underlines the evolutionary and structural conservation of the catalytic centres of all RNAP.