Group Leader: Matthias Wilmanns

Staff Scientist: Young-Hwa Song

Post-doctoral fellows: Christos Colovos*, Olivier Goncalves*, Anni Linden, Johan Nissen*, Christina Vega-Fernandez, Martina Walker, Peijian Zou.

Predoctoral fellows: Alice Douangamath, Fracisco Fernandez-Perez, Simone Müller, Attila Remenyi, William Stanley.

Research technician and trainees: Larissa Consani-Santos*, Katrin Lochner*, Gavin Murphy, Nikos Pinotsis*, Annette Roeben*, Katja Schirwitz, Jana Schroth*, Husseyin Uysal*.

Our group is interested in research in molecular structural biology, using the synchtrotron radiation beam lines at EMBL Hamburg and employing our structural data for exploring the functional roles of these biomacromolecules within the cell or the organism. We are attempting to link our research interests with the novel possibilities emerging from the knowledge of sequenced genomes. We are currently investigating six research areas, of which two are highlighted in this report. Recent results on other projects are described within the references.

• Domains and domain/ligand complexes from the muscle protein titin.

• Structural basis of the protein import into peroxisomes via the PTS1 receptor.

• Regulation of activity of transcription factors.

• Signalling by members of the calcium/calmodulin dependent protein Ser/Thr kinases.

• Signalling by modular SH3 domains in yeast.

• Evolution and catalytic mechanisms of proteins involved in metabolic pathways.

The functional properties of a wide range of transcription factors are altered by conformational changes induced by activators, repressors and specific receptor ligands. At the structural level, these changes often lead to the formation of additional helices, reorientation of loops and rearrangements of hydrophobic cores. We chose to study the structural mechanisms of regulation of transcription of the prototypic factor Oct-1 that belongs to the POU transcription factor family. Members of this family are involved in a broad range of biological processes ranging from house-keeping gene functions (Oct-1) to programming of embryonic stem cells (Oct-4) and the development of immune responses (Oct-1, Oct-2). However, according to the latest global sequencing reports, human, fly, and worm genomes encode only 15, five, and four POU factors, respectively. Therefore, members of this transcription factor family need to rely on multilevel control mechanisms such as post-translational modification, interaction with heterologous transcriptional regulators, and flexible DNA binding, to perform these multiple tasks. The flexibility inherent to members of the POU factor family, is conferred by a linker joining the POU-specific (POUS) and the POU-homeo domain (POUH). This linker is variable both in sequence and length (15-56 residues). Since both domains are structurally and functionally autonomous in DNA-binding, various arrangements on DNA are possible.

Figure 1: An artistic cartoon, based on the X-ray structures of the Oct-1/MORE and Oct-1/PORE complexes, how this transcription factor could associated with specific DNA response motifs. The image has been prepared by Ansgar Phillipsen, University of Basel.

The transcriptional activity of Oct-1 on the octamer motif in B-cells is regulated by the lymphoid-specific co-activator OBF-1 by clamping the POU_H and POU_S sub-domains together and thus enhancing their DNA binding affinity. Recent data from the collaborating group (Hans Schöler, Univ. Pennsylvania, previously EMBL) revealed the Oct-1 dimer formed within immunoglobulin heavy chain promoters (V_H) fails to interact with OBF-1. In contrast, the Oct-1 dimeric complex, formed by recognition element for osteopontin expression, can interact and synergize in transcriptional activation with this co-activator. These findings established the paradigm of differential transcriptional regulation mediated by two distinct POU dimer configurations [Tomilin et al. (2000) Cell 103, 853-864].

To elucidate the structural basis of this phenomenon, we have solved two crystal structures of the Oct-1 DNA binding segment bound to two different DNA binding elements, called MORE and to the PORE, that mimick the V_H and osteopontin recognition sites [4,7]. By direct comparison, these structures demonstrate how the same polypeptide chain can form two different dimer arrangements with two distinct protein/protein dimerization interfaces. These data introduce the concept of distinct transcription factor dimerization that depends on the sequence and the spacing of the protein domain binding motifs of the DNA response element. Thus, it extends previous models of protein-DNA complex formation mediated by ligand induced allosteric effects. Our results demonstrate how transcriptional activity can be regulated by swapping protein/protein interfaces between different quaternary arrangements of POU factor/DNA complexes (Figure 1).

Multiple signalling pathways in bifunctional glutaminase-synthase complexes

Histidine biosynthesis is carried out by eleven enzymatic reactions using ATP and PRPP. In the previous report, we described the (ba)₈-barrel structures of HisA and HisF [Lang et al. (2000) Science 289, 1546-1550]. Our sequence and structure analysis allowed to propose a common evolutionary pathway of these two enzymes, comprising three steps: (1) duplication of half-barrel genes, (2) diversification into an ancestral barrel with broad functionality, (3) a second gene duplication step leading to diversification of the extant (ba)₈-barrel enzymes HisA and HisF. The knowledge of these structures provided the basis for new approaches in biotechnology. The group of Reinhard Sterner at the University of Cologne succeeded to engineer TrpF activity in a TrpF knock out strain from Thermotoga maritima by a random mutagenesis approach. The directed evolution experiment demonstrated that one single residue mutation is sufficient to generate TrpF catalytic activity from HisA. These data have been reviewed and summarised in [6,7].

More recently we turned to the elucidation of the catalytic mechanism of one of three two enzymes, HisF or imidazole glycerol phosphate synthase, which forms a bifunctional complex with yet another enzyme of the histidine biosynthesis pathway, the glutaminase HisH. The HisF-HisH complex catalyses a cyclisation reaction of a pathway intermediate, called 5'-PRFAR, which includes the addition of an ammonia group. Since reactive ammonia is not available under physiological conditions, in a number of metabolic reactions the amino acid glutamine is used as a source for the incorporation of nitrogen. These reactions are catalysed by so-called glutamine amidotransferases (GATases), in which the glutaminase activity is coupled with a subsequent synthase activity that is specific for each member of the enzyme family. In the case of the ImGP synthase, it consists of a class-I HisH glutaminase and a (ba)₈-barrel HisF cyclase, which receives ammonia from HisH.

In order to elucidate the molecular mechanism of this coupled bienzyme-reaction, we have recently solved the X-ray structure of the ImGP synthase bienzyme from the hyperthermophile Thermotoga maritima at 2.4 Å resolution [10]. Its structure shows that the HisH glutaminase subunit is associated with the N-terminal face of the central b-barrel of the HisF cyclase subunit (Figure 2). The complex reveals a putative tunnel for the transfer of ammonia over a distance of 25 Å, of which the larger part is formed by the interior of the central b-barrel of HisF. The structure of the ImGP synthase suggests a mechanism for coupling of the glutaminase and cyclase activities and for the tunnelling of ammonia within the bienzyme complex. Although ammonia tunnelling has been reported for other GATases, the ImGP synthase has evolved a novel mechanism, which extends the known functional properties of the versatile (ba)₈-barrel fold.

Figure 2: Overall structure of the ImGP synthase bienzyme complex, in which the HisF cyclase subunit is in yellow/orange and the HisH glutaminase subunit is in blue/cyan. In the right panel, the putative path for ammonia between the active sites of HisH and HisF is indicated by a number of red spheres. A number of catalytic key residues and the two phosphate ions bound to the HisF active site are indicated by balls and sticks.

Recent data suggest that in the ImGP synthase complex there are probably three signalling pathways linking the catalytic activities of the two active sites (Figure 3), which, despite the knowledge of its three-dimensional structure, are only poorly understood. The most obvious pathway is for the transport of nascent and reactive ammonia from the glutaminase (HisH) active site to the cyclase (HisF) active site. Despite the structure only suggests a single plausible pathway across the HisF b-barrel, in the available inactive conformation it is blocked by a salt-bridge cluster within this b-barrel. Furthermore, there is accumulating evidence from spectroscopy and mutagenesis data (R. Sterner, personal communication) that there is cross-activation of HisF and HisH catalytic activities, by binding of the HisH (glutamine) and HisF (5'-PRFAR) substrates, respectively. Our future aim is to provide a molecular mechanism for the triple signalling pathways within the HisH-HisF bienzyme complex.

Figure 3: Structural model of the reaction cycle of ImGP synthase. (A) In the inactive conformation of the bienzyme complex, the HisH subunit is docked to the N-terminal face of the HisF subunit in a conformation that allows access of glutamine to the active site. Binding of the substrate PRFAR to the HisF active site triggers a signal that is transmitted to the HisH-HisF interface, leading to sequestration of the HisH active site from solvent and opening of the ammonia tunnel across the HisF (ba)₈-barrel. Ammonia, which is the product of the HisH glutaminase reaction, is then transferred via the putative (ba)₈-barrel tunnel to the active site of HisF, where it reacts with PRFAR to yield ImGP and AICAR. After release of the HisF products, the conformation of the HisH-HisF complex returns to its inactive state. The following symbolic objects are used: HisH subunit, pink rectangle; (ba)₈-barrel HisF subunit, blue barrel; glutamine (HisH substrate) and glutamate (HisH product), orange rectangle; PRFAR, cyan dump-bell; ammonia, green sphere. Red arrows represent signal transfer from HisF to HisH (B), and ammonia transfer from HisH to HisF (C). The figure has been prepared by Petra Riedinger, OIPA, EMBL.

References:

[1] Purification, characterisation and crystallization of anthranilate phosphoribosyltransferase from Sulfolobus solfataricu. Ivens, A., Mayans, O., Szadkowski, H., Wilmanns, M. & Kirschner, K. (2001), Eur J. Biochem. 268, 2246-2252
[2] Atomic resolution structure of a Src Homolgy 3 domain mutant of the spectrin SH3 domain. Berisio, R., Viguera, A., Serrano, L. & Wilmanns, M. (2001), Acta Cryst.D57, 337-340.
[3] Stabilisation of indoleglycerol phosphate synthase from Escherichia coli by an engineered disulfide bond. Ivens, A., Szadkowski, H., Mayans, O., Wilmanns, M. & Kirschner, K., (2001) Eur. J.Biochem, in press.
[4] Structural evidence for a possible role of reversible disulfide bridge formation in elasticity of the muscle protein titin. Mayans, O., Wuerges, J., Gautel, M. & Wilmanns, M. (2001), Structure9, 331-340.
[5] Crystallisation of redox-insensitive Oct1 POU domain with different DNA response elements A. Remenyi, E. Pohl, H. Schöler & M. Wilmanns (2001), Acta Cryst. D57, 1634-1638.
[6] Divergent evolution of (ba)₈-barrel enzymes. M. Henn-Sax, B. Höcker, M. Wilmanns & R. Sterner (2001), Biol. Chem. 382, 1315-1320.
[7] Stability, evolution and applications of (ba)_8-barrel enzymes. Birte Höcker, C. Jürgens, M. Wilmanns & R. Sterner (2001), Curr. Opin. Biotechnology 12, 376-381,
[8] Differential transcriptional activity of Oct-1 by DNA-motif induced dimerization interface swapping. Attila Reményi, Alexey Tomilin, Ehmke Pohl, Katharina Lins, Ansgar Philippsen, Rolland Reinbold, Hans R. Schöler and Matthias Wilmanns (2001), Mol. Cell 8, 569-58.
[9] Matti Saraste (1949-2001). Matthias Wilmanns (2001). Structure9, 767-770.
[10] Unusual binding properties of the SH3 domain of the yeast actin binding protein Abp1: Structural and functional analysis. Barbara Fazi, Jamie Cope, Alice Douangamath, Silvia Ferracuti, Katja Schirwitz, Adriana Zucconi, David Drubin, Matthias Willmanns, Gianni Cesareni and Luisa Castagnoli (2001). J. Biol. Chem., in press.
[11] Structural evidence for ammonia tunnelling across the (ba)8-barrel of the imidazole glycerol phosphate synthase bienzyme complex. Alice Douangamath, Martina Walker, Silke Beismann-Driemeyer, M. Cristina Vega-Fernandez, Reinhard Sterner and Matthias Wilmanns, Structure, in press.