Final Report Summary - TRANSYS (A systems approach to defining membrane protein networks and applications)
The overall aim of the TranSys project was to understand how membrane proteins insert into bacterial membranes, assemble into functional complexes, and how they form interactive networks with other membrane-bound and soluble proteins. The network employed 9 Early Stage Researchers and involved 5 full partners plus one Associate partner.
Important results from the 5 full partners can be summarised as follows:
1, University of Warwick.
Jacopo Baglieri carried out a detailed study of the structure and assembly of the twin-arginine translocation (Tat) protein transport pathway. He characterised intermediates in the assembly of the TatABC and TatA complexes and these data are being prepared for publication. He purified the TatE complex from E. coli for the first time and published an important paper on the structure of this complex, which yielded several surprises (Baglieri et al, J. Biol. Chem, see below). He then collaborated with ESR Monteferrante from partner 2 to understand the role of the TatAc complex in Bacillus subtilis, and this joint work was published in Appl Env Micro in 2012. Finally, the two ESRs collaborated closely to generate the first detailed structure of the important TatAd complex from B. subtilis, and these results are in the process of being published in Biochim Biophys Acta (paper is accepted subject to minor revision).
Anna Albiniak investigated the Tat pathway from a different angle. she studied the B. subtilis TatAdCd system's properties when expressed in E. coli, and found that it partially complemented an E. coli tat null mutant strain. Importantly, it was able to translocate some proteins, including heterologous proteins, if they bear specific Tat signal peptides, but the TatAdCd system did not complement a key property of the E. coli delta tat strain: the presence of large holes in the outer membrane. Expression of heterologous proteins with a Tat signal peptide therefore leads to initial export into the periplasm, but the protein is then released into the culture medium. This is a major breakthrough that offers a new means of producing and harvesting recombinant proteins, and a patent application has been filed. The results have also been submitted for publication.
2, University Medical Centre, Groningen.
Carmine G. Monteferrante has investigated the twin-arginine protein transport (Tat) system of Bacillus subtilis at all levels that were in the focus of the TranSys ITN research programme. Together with ESR Goosens from partner 2 and ESR MacKichan from partner 3, he studied the Tat interactome of B. subtilis at the network level. This resulted in the identification of several novel factors that are needed for productive protein translocation via Tat, as well as novel Tat-dependent phenotypes of B. subtilis (published in Appl. Environ. Microbiol., Proteomics, and J. Proteome Res.). He then investigated the formation of B. subtilis Tat complexes with ESR Baglieri from partner 1 at the biochemical and structural levels (published in Appl. Environ. Microbiol. and in revision for Biochim. Biophys. Acta). Next, he investigated the assembly, structure and biochemical properties of a novel iron transport system of B. subtilis that includes the Tat-dependent hemoprotein EfeB (in revision for Mol. Microbiol). At the same time, he discovered and characterized a novel Tat-dependent phosphodiesterase of B. subtilis, YkuE, for which he showed that it is specifically targeted to the cell wall (published in J. Biol. Chem.). Lastly, he investigated properties of the B. subtilis Tat system that set limits to its commercial exploitation for protein production (published in PLoS ONE and Appl. Environ. Microbiol.). Altogether, Carmine Monteferrante has authored 9 publications of which 7 have already appeared in peer-reviewed journals, and 2 are currently in revision.
Vivianne J. Goosens investigated different aspects of the membrane proteome of B. subtilis. In collaboration with ESR Monteferrante from partner 2 and ESR MacKichan from partner 3, she investigated the Tat-dependent membrane proteome of B. subtilis. This generated novel insights in the diverse processes in which this system plays important roles, including iron transport and biofilm formation (published in J. Biol. Chem. and J. Proteome Res., and one manuscript in preparation). During a three-months internship in the laboratory of the associate partner DSM, Vivianne Goosens studied the thiol-disulphide oxidoreductase functions of B. subtilis at the membrane proteome level. This resulted in a joint publication with researchers at DSM (published in Antioxid. Redox Signal.). In addition, she developed novel protocols for membrane proteome analyses in industrial Bacillus strains. A third project addressed the roles of membrane proteases in the regulated turnover of integral membrane proteins based on quantitative analyses of metabolically labeled B. subtilis cells (two manuscripts in preparation). Altogether, Vivianne Goosens is the author of 6 publications on interactive membrane protein networks of B. subtilis and their interactions with other membrane-bound and soluble proteins; 3 of these publications have already appeared in peer-reviewed journals, and 3 will shortly be submitted for publication.
3. INRA Jouy.
Calum Mackichan studied the locations of key proteins within the bacterial membrane. The bacterial membrane has been the subject to localization studies of both proteins and lipids, characterising a number of localization patterns; homogenous, heterogeneous spots, helical, polar, septal, and dynamic membrane foci. Many of these have been performed with ectopically expressed GFP fusions using epifluorescence microscopy, therefore an outstanding question remains over the organization of natively expressed proteins in the membrane. Here we have investigated the localization of functional fusions representing the important membrane processes of secretion, membrane protein insertion and phospholipid synthesis. The general secretory pathway, the Sec pathway, has previously been reported to form large static domains interpreted to form a helix-like array in the lateral membrane, and phospholipid synthesis has been reported to localize at septal regions. We characterized natively expressed functional GFP fusions to SecA, the membrane insertase YidC1, and the essential phospholipid synthase PgsA in live Bacillus subtilis cells using near-TIRF microscopy. Our studies revealed membrane-associated foci that transiently follow random trajectories, and are capable of fusing and dissociating from one another, forming and reforming dynamically in the membrane. The foci are visible throughout growth and their density correlates with cell length. Therefore, rather than taking place at specific sites in the cell as previously reported, these processes are localized in dynamically distributed foci throughout the membrane. Finally, we performed interaction studies on proteins involved in secretion. A network centred around the twin-arginine translocation (Tat) pathway was analyzed, revealing three proteins with roles in translocation. Localization studies were performed on two of these proteins, HemAT and CsbC, revealing polar and homogeneous profiles respectively. The polar localization of HemAT was found to be dependent on expression of the sigma factor D regulon.
Michal Ferens's project examined the role of membrane protein networks in bacterial cell division. In bacteria, initiation of DNA replication is tightly regulated to occur once per cell cycle. In the spore forming bacteria B. subtilis, mechanisms responsible for initiation control are poorly understood. We found that the replication initiator DnaA interacts with YkjA, an unknown transmembrane protein. B. subtilis encodes 5 five paralogs of ykjA, members of the ydfR-family, which indicates a strong functional redundancy. To better understand the role of this novel family of membrane protein in B. subtilis, we first looked at their interaction context by building a protein-protein interaction network centered on YkjA and other members. Using a yeast two hybrid approach, we found that these proteins were all associated with a group of highly connected membrane proteins call hubs. Genomic screens using their cytosolic domains revealed that they mostly interact with transcriptional regulators and other DNA-binding proteins. We undertook the biological characterisation of the YkjA/DnaA complex and found that YkjA is expressed during sporulation and co-localises with the asymmetric septum of the forespore. Sporulation is strongly delayed in a ykjA and other ydrR-related mutant strains, while excess of YkjA in vegetative cells is responsible to growth inhibition. Dosage of origin-proximal sequences suggests that the growth inhibition correlates with a down regulation of replication initiation. Using a reverse yeast two-hybrid approached, we identified residues important for interaction. These residues were mapped on the 3D structures of the proteins and delineated interacting surfaces.
4. Stockholm University.
Pilar Lloris Garcera worked on the archetypal dual-topology protein EmrE. Together with another PhD student (Susanna Seppälä) and a postdoc (Joanna Slusky), she has published two very important studies. In the first (Science 328:1698) she could show that the orientation of EmrE in the inner membrane of E. coli can be influenced by a single C-terminal positively charged residue. This was totally unexpected, since it implies that membrane protein topology can be determined post-translationally. In her second publication (JBC 287:26052), she was able to sort out a contentious issue concerning whether the active form of EmrE is a parallel or anti-parallel dimer. In as yet unpublished work, she has done a very carful mutagenesis scan of the entire EmrE protein, and the results are perfectly consistent with a published - but controversial - medium-resolution X-ray structure. She has also finished work on a series of EmrE-EmrE fusions that further illuminates the properties of the anti-parallel active dimer. These two latter studies are currently being written up for publication.
Marcin Skwark worked on the development of methods for protein structure prediction. His most important contributions have been the development of a number of new methods for structure predictions, in particular PconsC and PconsD. Both these methods use contact based information to improve structure predictions. In addition to his work on the development of novel methods he has been the main (only) person responsible for Stockholm University's protein structure prediction servers at CASP9 and CASP10.
5. Novozymes.
Sara Lundström worked on characterizing the gene cluster required for functional expression of a bacteriocin from Bacillus licheniformis ATCC 14580. This bacteriocin, with the proposed name formosin (ForD) are encoded in a gene located in the chromosome of B. licheniformis ATCC 14580 with three adjacent genes, proposed to be named forE, forF, and forG, respectively. Genetic in silico analysis showed that these four genes are arranged in an operon situated in a genomic island with host defensive properties. Structural and functional studies demonstrated that ForE and ForG constitute an ABC transporter required in both secretion of and immunity to formosin. ForF is an accessory protein to the ForEG ABC transporter containing an N-terminal transmembrane domain. No function could be linked to ForF. Secretion analysis revealed formosin to have two secretion signals; one N-terminal sec-dependent signal peptide and one C-terminal ABC transporter signal. However, only when secreted through the ForEG ABC transporter could formosin be detected in the medium. Characterization of formosin showed that it is a 9.6kDa heat-labile bacteriocin belonging to the lactococcin 972 family with an observed bacteriolytic effect on Bacillus subtilis. ForG is a structural homolog of a previously described “immunity” protein associated with this protein family. Investigations of lactococcin 972-like protein and adjacent genes, in addition to the results obtained within this project, concluded that the members of the lactococcin 972 family are associated with ABC transporters and not transmembrane immunity proteins as previously predicted.
Overall summary
The project significantly increased our understanding of membrane proteins at many levels, through (i) new computational methods to analyse them, (ii) the identification of new networks of interacting partners, (iii) new means of exploiting them for biotechnological purposes and (iv) new models to describe theior biogenesis and assembly. These data have been published in high-impact journals, with numerous publications still in preparation, and the data have been further disseminated through multiple events attended by the ESRs and their supervisors (see next section).