Machine-learning methods for structure prediction of β-barrel membrane proteins

Different types of proteins exist with diverse functions that are essential for living organisms. An important class of proteins is represented by transmembrane proteins which are specifically designed to be inserted into biological membranes and devised to perform very important functions in the cell such as cell communication and active transport across the membrane. Transmembrane β-barrels (TMBBs) are a sub-class of membrane proteins largely under-represented in structure databases because of the extreme difficulty in experimental structure determination. For this reason, computational tools that are able to predict the structure of TMBBs are needed. In this thesis, two computational problems related to TMBBs were addressed: the detection of TMBBs in large datasets of proteins and the prediction of the topology of TMBB proteins. Firstly, a method for TMBB detection was presented based on a novel neural network framework for variable-length sequence classification. The proposed approach was validated on a non-redundant dataset of proteins. Furthermore, we carried-out genome-wide detection using the entire Escherichia coli proteome. In both experiments, the method significantly outperformed other existing state-of-the-art approaches, reaching very high PPV (92%) and MCC (0.82). Secondly, a method was also introduced for TMBB topology prediction. The proposed approach is based on grammatical modelling and probabilistic discriminative models for sequence data labeling. The method was evaluated using a newly generated dataset of 38 TMBB proteins obtained from high-resolution data in the PDB. Results have shown that the model is able to correctly predict topologies of 25 out of 38 protein chains in the dataset. When tested on previously released datasets, the performances of the proposed approach were measured as comparable or superior to the current state-of-the-art of TMBB topology prediction.

Molecular basis oh herpes simplex virus entry into the cell and retargeting of the viral tropism for the design of oncolytic herpesviruses

Herpes simplex virus 1 (HSV-1) infects oral epitelial cells, then spreads to the nerve endings and estabilishes latency in sensory ganglia, from where it may, or may not reactivate. Diseases caused by virus reactivation include mild diseases such as muco-cutaneous lesions, and more severe, and even life-threatening encephalitis, or systemic infections affecting diverse organs. Herpes simplex virus represents the most comprehensive example of virus receptor interaction in Herpesviridae family, and the prototype virus encoding multipartite entry genes. In fact, it encodes 11-12 glycoproteins and a number of additional membrane proteins: five of these proteins play key roles in virus entry into subsceptible cells. Thus, glycoprotein B (gB) and glycoprotein C (gC) interact with heparan sulfate proteoglycan to enable initial attachment to cell surfaces. In the next step, in the entry cascade, gD binds a specific surface receptor such as nectin1 or HVEM. The interaction of glycoprotein D with the receptor alters the conformation of gD to enable the activation of gB, glycoprotein H, and glycoprotein L, a trio of glycoproteins that execute the fusion of the viral envelope with the plasma membrane. In this thesis, I described two distinct projects: I. The retargeting of viral tropism for the design of oncolytic Herpesviruses: • capable of infecting cells through the human epitelial growth factor receptor 2 (HER2), overexpressed in highly malignant mammary and ovarian tumors and correlates with a poor prognosis; • detargeted from its natural receptors, HVEM and nectin1. To this end, we inserted a ligand to HER2 in gD. Because HER2 has no natural ligand, the selected ligand was a single chain antibody (scFv) derived from MAb4D5 (monoclonal antibody to HER2), herein designated scHER2. All recombinant viruses were targeted to HER2 receptor, but only two viruses (R-LM113 and R-LM249) were completely detargeted from HVEM and nectin1. To engineer R-LM113, we removed a large portion at the N-terminus of gD (from aa 6 to aa 38) and inserted scHER2 sequence plus 9-aa serine-glycine flexible linker at position 39. On the other hand, to engineer R-LM249, we replaced the Ig-folded core of gD (from aa 61 to aa 218) with scHER2 flanked by Ser-Gly linkers. In summary, these results provide evidence that: i. gD can tolerate an insert almost as big as gD itself; ii. the Ig-like domain of gD can be removed; iii. the large portion at the N-terminus of gD (from aa 6 to aa 38) can be removed without loss of key function; iv. R-LM113 and R-LM249 recombinants are ready to be assayed in animal models of mammary and ovary tumour. This finding and the avaibility of a large number of scFv greatly increase the collection of potential receptors to which HSV can be redirected. II. The production and purification of recombinant truncated form of the heterodimer gHgL. We cloned a stable insect cell line expressing a soluble form of gH in complex with gL under the control of a metalloprotein inducible promoter and purified the heterodimer by means of ONE-STrEP-tag system by IBA. With respect to biological function, the purified heterodimer is capable: • of reacting to antibodies that recognize conformation dependent epitopes and neutralize virion infectivity; • of binding a variety cells at cell surface. No doubt, the availability of biological active purified gHgL heterodimer, in sufficient quantities, will speed up the efforts to solve its crystal structure and makes it feasible to identify more clearly whether gHgL has a cellular partner, and what is the role of this interaction on virus entry.

From "wet biology" to statistical analysis of structural features with bioinformatics tools

Many new Escherichia coli outer membrane proteins have recently been identified by proteomics techniques. However, poorly expressed proteins and proteins expressed only under certain conditions may escape detection when wild-type cells are grown under standard conditions. Here, we have taken a complementary approach where candidate outer membrane proteins have been identified by bioinformatics prediction, cloned and overexpressed, and finally localized by cell fractionation experiments. Out of eight predicted outer membrane proteins, we have confirmed the outer membrane localization for five—YftM, YaiO, YfaZ, CsgF, and YliI—and also provide preliminary data indicating that a sixth—YfaL—may be an outer membrane autotransporter.

Use of E. coli OMVs as delivery system to elicit neutralizing antibodies against Chlamydia infectivity: the HtrA protein example

The obligate intracellular pathogen Chlamydia trachomatis is a gram negative bacterium which infects epithelial cells of the reproductive tract. C. trachomatis is the leading cause of bacterial sexually transmitted disease worldwide and a vaccine against this pathogen is highly needed. Many evidences suggest that both antigen specific-Th1 cells and antibodies may be important to provide protection against Chlamydia infection. In a previous study we have identified eight new Chlamydia antigens inducing CD4-Th1 and/or antibody responses that, when combined properly, can protect mice from Chlamydia infection. However, all selected recombinant antigens, upon immunization in mice, elicited antibodies not able to neutralize Chlamydia infectivity in vitro. With the aim to improve the quality of the immune response by inducing effective neutralizing antibodies, we used a novel delivery system based on the unique capacity of E. coli Outer Membrane Vesicles (OMV) to present membrane proteins in their natural composition and conformation. We have expressed Chlamydia antigens, previously identified as vaccine candidates, in the OMV system. Among all OMV preparations, the one expressing HtrA Chlamydia antigen (OMV-HtrA), showed to be the best in terms of yield and quantity of expressed protein, was used to produce mice immune sera to be tested in neutralization assay in vitro. We observed that OMV-HtrA elicited specific antibodies able to neutralize efficiently Chlamydia infection in vitro, indicating that the presentation of the antigens in their natural conformation is crucial to induce an effective immune response. This is one of the first examples in which antibodies directed against a new Chlamydia antigen, other than MOMP (the only so far known antigen inducing neutralizing antibodies), are able to block the Chlamydia infectivity in vitro. Finally, by performing an epitope mapping study, we investigated the specificity of the antibody response induced by the recombinant HtrA and by OMV-HtrA. In particular, we identified some linear epitopes exclusively recognized by antibodies raised with the OMV-HtrA system, detecting in this manner the antigen regions likely responsible of the neutralizing effect.

B-type cytochromes of the DOMON domain superfamily (Novel redox elements of plant plasma membrane)

The aim of the present study is understanding the properties of a new group of redox proteins having in common a DOMON-type domain with characteristics of cytochromes b. The superfamily of proteins containing a DOMON of this type includes a few protein families. With the aim of better characterizing this new protein family, the present work addresses both a CyDOM protein (a cytochrome b561) and a protein only comprised of DOMON(AIR12), both of plant origin. Apoplastic ascorbate can be regenerated from monodehydroascorbate by a trans-plasma membrane redox system which uses cytosolic ascorbate as a reductant and comprises a high potential cytochrome b. We identified the major plasma membrane (PM) ascorbate-reducible b-type cytochrome of bean (Phaseolus vulgaris) and soybean (Glycine max) hypocotyls as orthologs of Arabidopsis auxin-responsive gene air12. The protein, which is glycosylated and glycosylphosphatidylinositol-anchored to the external side of the PM in vivo, was expressed in Pichia pastoris in a recombinant form, lacking the glycosylphosphatidylinositol-modification signal, and purified from the culture medium. Recombinant AIR12 is a soluble protein predicted to fold into a β-sandwich domain and belonging to the DOMON superfamily. It is shown to be a b-type cytochrome with a symmetrical α-band at 561 nm, to be fully reduced by ascorbate and fully oxidized by monodehydroascorbate. Redox potentiometry suggests that AIR12 binds two high-potential hemes (Em,7 +135 and +236 mV). Phylogenetic analyses reveal that the auxin-responsive genes AIR12 constitute a new family of plasma membrane b-type cytochromes specific to flowering plants. Although AIR12 is one of the few redox proteins of the PM characterized to date, the role of AIR12 in trans-PM electron transfer would imply interaction with other partners which are still to be identified. Another part of the present project was aimed at understanding of a soybean protein comprised of a DOMON fused with a well-defined b561 cytochrome domain (CyDOM). Various bioinformatic approaches show this protein to be composed of an N-terminal DOMON followed by b561 domain. The latter contains five transmembrane helices featuring highly conserved histidines, which might bind haem groups. The CyDOM has been cloned and expressed in the yeast Pichia pastoris, and spectroscopic analyses have been accomplished on solubilized yeast membranes. CyDOM clearly reveal the properties of b-type cytochrome. The results highlight the fact that CyDOM is clearly able to lead an electron flux through the plasmamembrane. Voltage clamp experiments demonstrate that Xenopus laevis oocytes transformed with CyDOM of soybean exhibit negative electrical currents in presence of an external electron acceptor. Analogous investigations were carried out with SDR2, a CyDOM of Drosophila melanogaster which shows an electron transport capacity even higher than plant CyDOM. As quoted above, these data reinforce those obtained in plant CyDOM on the one hand, and on the other hand allow to attribute to SDR2-like proteins the properties assigned to CyDOM. Was expressed in Regenerated tobacco roots, transiently transformed with infected a with chimeral construct GFP: CyDOM (by A. rhizogenes infection) reveals a plasmamembrane localization of CyDOM both in epidermal cells of the elongation zone of roots and in root hairs. In conclusion. Although the data presented here await to be expanded and in part clarified, it is safe to say they open a new perspective about the role of this group of proteins. The biological relevance of the functional and physiological implications of DOMON redox domains seems noteworthy, and it can but increase with future advances in research. Beyond the very finding, however interesting in itself, of DOMON domains as extracellular cytochromes, the present study testifies to the fact that cytochrome proteins containing DOMON domains of the type of “CyDOM” can transfer electrons through membranes and may represent the most important redox component of the plasmamembrane as yet discovered.

The Domon Family of Plant Plasma Membrane B-Type Cytochromes: Heterologous Expression, Biochemical Characterization and Physiological Roles in Vivo

The DOMON domain is a domain widespread in nature, predicted to fold in a β-sandwich structure. In plants, AIR12 is constituted by a single DOMON domain located in the apoplastic space and is GPI-modified for anchoring to the plasma membrane. Arabidopsis thaliana AIR12 has been heterologously expressed as a recombinant protein (recAtAIR12) in Pichia pastoris. Spectrophotometrical analysis of the purified protein showed that recAtAir12 is a cytochrome b. RecAtAIR12 is highly glycosylated, it is reduced by ascorbate, superoxide and naftoquinones, oxidised by monodehydroascorbate and oxygen and insensitive to hydrogen peroxide. The addition of recAtAIR12 to permeabilized plasma membranes containing NADH, FeEDTA and menadione, caused a statistically significant increase in hydroxyl radicals as detected by electron paramagnetic resonance. In these conditions, recAtAIR12 has thus a pro-oxidant role. Interestingly, AIR12 is related to the cytochrome domain of cellobiose dehydrogenase which is involved in lignin degradation, possibly via reactive oxygen species (ROS) production. In Arabidopsis the Air12 promoter is specifically activated at sites where cell separations occur and ROS, including •OH, are involved in cell wall modifications. air12 knock-out plants infected with Botrytis cinerea are more resistant than wild-type and air12 complemented plants. Also during B. cinerea infection, cell wall modifications and ROS are involved. Our results thus suggest that AIR12 could be involved in cell wall modifying reactions by interacting with ROS and ascorbate. CyDOMs are plasma membrane redox proteins of plants that are predicted to contain an apoplastic DOMON fused with a transmembrane cytochrome b561 domain. CyDOMs have never been purified nor characterised. The trans-membrane portion of a soybean CyDOM was expressed in E. coli but purification could not be achieved. The DOMON domain was expressed in P. pastoris and shown to be itself a cytochrome b that could be reduced by ascorbate.