• Core Facilities
• Research Support
• Campus Science Support Facilities (CSF)

• International PhD Programme
• Summer School
• Scientific Meetings & Seminars

• Scientific Advisory Board (SAB)
• Annual Report
• Publication Overview

Enzymatic activities to phosphorylate and ligate RNA molecules were reported in the late nineteen-seventies and early nineteen-eighties, but the actual genes have remained elusive. Combining protein purification, mass spectrometry, RNAi and phyletic patterns, we identified the human RNA-kinases CLP1 and NOL9 which are able to bind ATP and phosphorylate RNA-5' termini, and the human tRNA ligase complex in charge of tRNA splicing and probably other non-canonical splicing events. We aim to dissect molecular mechanisms and identify RNA targets for all of these enzymatic activities. In addition, we are generating mouse models to observe in vivo functions and potential links to disease. Discovering new enzymes and pathways is exciting, especially when it permits renewed discoveries and introduces fresh perspectives into phenomena identified some 30 years ago.

•  a) Subunit composition and molecular mechanisms: (Jennifer Jurkin and Johannes Popow).
  In 2011 we identified HSPC117 (also known as RTCB) as the catalytic subunit of the long-time elusive human tRNA ligase complex. This opened new and exciting perspectives at our laboratory. In addition to RTCB, the tRNA ligase complex is composed of the DEAD-box ATP-dependent RNA helicase DDX1 and three extra subunits of unknown function, i.e. CGI-99, FAM98B and Ashwin. DEAD box proteins are capable of unwinding double-stranded RNAs as well as assembling/disassembling RNA-protein complexes and remodeling RNA molecules within ribonucleoprotein particles. They display a helical core composed of at least eleven conserved motifs. To study the function of DDX1 within the tRNA ligase complex, we generated stable cell lines that express mutant versions of DDX1 containing a FLAG-tag. We mutagenized RNA and ATP binding domains and assayed RNA ligase activity of FLAG-immunopurified tRNA ligase wild-type and mutant complexes. The results have been encouraging: we observe reduced ligation activity in those mutants when using different types of RNA substrates. In the near future we will try to reconstitute the entire tRNA ligase complex with recombinant proteins and study the catalytic activity and substrate-binding ability of recombinant DDX1.

  b) A potential function in non-canonical splicing events during the Unfolded Protein Response (UPR): (Theresa Henkel, Jennifer Jurkin, Anne Nielsen, Johannes Popow and Stefan Weitzer).
  We suspect that the tRNA ligase performs functions in addition to tRNA splicing. This speculation has been confirmed by PAR-CLIP experiments (Photo Activatable Ribonucleoside Enhanced Crosslinking and Immunoprecipitation), which revealed that the tRNA ligase binds to a multitude of mRNAs including Xbp1, an mRNA encoding a transcription factor required during UPR (Figure 1). In unstressed cells XBP1 encodes a cytoplasmic unstable protein. However, upon stress Xbp1-mRNA is cleaved twice by the ER-membrane endoribonuclease IRE1, leading to the removal of a 26-nucleotide intron. Joining the neighboring exons introduces a frame shift that, upon translation, generates a larger protein that moves to the nucleus and acts as a crucial transcription factor (Figure 2). The question we would like to answer is: what is the ligase that joins the exons? Interestingly, Peter Walter's group at the UCSF revealed, back in 1996, that the yeast tRNA ligase TRL1 splices Hac1-mRNA (homologous to Xbp1-mRNA) during yeast UPR. Our unpublished data strongly suggest that the human tRNA ligase complex is the long sought UPR ligase in human cells.

  c) In vivo function: generation of knockout mice: (Jennifer Jurkin) Transferring our biochemical results to mouse models, we have generated mice carrying a conditional HSPC117 allele in which exon 4 is flanked by two loxP sites. After removal of the floxed region a premature stop codon is generated, and the resulting transcript is subjected to non-sense-mediated decay. Crossing these mice to a ubiquitously expressed Cre line (i.e. ß-actin-Cre, which mediates deletion in all body cells from early embryonic development onwards) leads to embryonic lethality between implantation and E10.5. We are currently trying to determine the precise time point and the reason for this early lethality. Furthermore, we are generating tissue-specific knockouts. Protein extracts obtained from these tissues lack HSPC117 and are therefore impaired in respect of tRNA ligation activity in vitro, confirming successful targeting. We are currently characterizing the phenotype of these tissue-specific knockouts. We have also replaced the wild type HSPC117 gene with a hypomorphic allele (knock-in) and will analyze mice carrying this allele in the near future.

  Cellular functions of the 5’-polynucleotide kinase NOL9:

  (Sabrina Bandini)

  Ribosome biogenesis is a finely regulated process that allows the cell to coordinate cell growth with cell proliferation. To synthesize the required number of ribosomes the cell relies on efficient and accurate generation of ribosomal RNAs (rRNAs). The 18S, 5.8S and 28S rRNAs are transcribed in the nucleolus by RNA polymerase I as one polycistronic precursor RNA, and are liberated by a complex series of endo- and exonucleolytic cleavage events. Eventually they are assembled together with the 5S rRNA and a plethora of ribosomal proteins to form the ribosomes.

  We recently identified NOL9, the first human nucleolar polynucleotide kinase. The catalytically active enzyme is required for processing the 32S rRNA precursor into 28S and 5.8S rRNAs. Processing of rRNA is almost unexplored in human because of its complexity and redundancy. Our main purpose is to determine the reasons why the human rRNA pathway requires an RNA kinase, and to identify the key enzymes required for specific steps in the processing of the 32S rRNA.

  We stably expressed FLAG-tagged NOL9 wild-type and ATP-binding mutants in HEK293 cells to study its nucleolar localization, its potential interaction with co-purified proteins, and reveal specific NOL9-binding sites on the rRNA by PAR-CLIP. In Figure 3 we show that NOL9 binds the 5' end of the 5.8S rRNA in vivo - a predictable result in view of the role NOL9 plays in establishing ratios between the short and long versions of the 5.8S rRNA.

•