De novo protein design for catalysis of any desired chemical reaction is a long standing goal in protein engineering, due to the broad spectrum of technological, scientific and medical applications. Currently, mapping protein sequence to protein function is, however, neither computationionally nor experimentally tangible 1,2 . Here we developed ProteinGAN, a specialised variant of the generative adversarial network 3 that is able to 'learn' natural protein sequence diversity and enables the generation of functional protein sequences. ProteinGAN learns the evolutionary relationships of protein sequences directly from the complex multidimensional amino acid sequence space and creates new, highly diverse sequence variants with natural-like physical properties. Using malate dehydrogenase as a template enzyme, we show that 24% of the ProteinGAN-generated and experimentally tested sequences are soluble and display wild-type level catalytic activity in the tested conditions in vitro , even in highly mutated (>100 mutations) sequences.ProteinGAN therefore demonstrates the potential of artificial intelligence to rapidly generate highly diverse novel functional proteins within the allowed biological constraints of the sequence space.

De novo protein design for catalysis of any desired chemical reaction is a long standing goal in protein engineering, due to the broad spectrum of technological, scientific and medical applications. Currently, mapping protein sequence to protein function is, however, neither computationionally nor experimentally tangible 1,2 . Here we developed ProteinGAN, a specialised variant of the generative adversarial network 3 that is able to 'learn' natural protein sequence diversity and enables the generation of functional protein sequences. ProteinGAN learns the evolutionary relationships of protein sequences directly from the complex multidimensional amino acid sequence space and creates new, highly diverse sequence variants with natural-like physical properties. Using malate dehydrogenase as a template enzyme, we show that 24% of the ProteinGAN-generated and experimentally tested sequences are soluble and display wild-type level catalytic activity in the tested conditions in vitro , even in highly mutated (>100 mutations) sequences.ProteinGAN therefore demonstrates the potential of artificial intelligence to rapidly generate highly diverse novel functional proteins within the allowed biological constraints of the sequence space.

SummaryThe propensity of peptides and proteins to form self-assembled structures has very promising applications in the development of novel nanomaterials. Under certain conditions, amyloid protein α-synuclein forms well-ordered structures – fibrils, which have proven to be valuable building blocks for bionanotechnological approaches. Herein we demonstrate the functionalization of fibrils formed by a mutant α-synuclein that contains an additional cysteine residue. The fibrils have been biotinylated via thiol groups and subsequently joined with neutravidin-conjugated gold nanoparticles. Atomic force microscopy and transmission electron microscopy confirmed the expected structure – nanoladders. The ability of fibrils (and of the additional components) to assemble into such complex structures offers new opportunities for fabricating novel hybrid materials or devices.

The recombinant phage tail sheath protein, gp053, from Escherichia coli infecting myovirus vB_EcoM_FV3 (FV3) was able to self-assemble into long, ordered and extremely stable tubular structures (polysheaths) in the absence of other viral proteins. TEM observations revealed that those protein nanotubes varied in length (~10–1000 nm). Meanwhile, the width of the polysheaths (~28 nm) corresponded to the width of the contracted tail sheath of phage FV3. The formed protein nanotubes could withstand various extreme treatments including heating up to 100 °C and high concentrations of urea. To determine the shortest variant of gp053 capable of forming protein nanotubes, a set of N- or/and C-truncated as well as poly-His-tagged variants of gp053 were constructed. The TEM analysis of these mutants showed that up to 25 and 100 amino acid residues could be removed from the N and C termini, respectively, without disturbing the process of self-assembly. In addition, two to six copies of the gp053 encoding gene were fused into one open reading frame. All the constructed oligomers of gp053 self-assembled in vitro forming structures of different regularity. By using the modification of cysteines with biotin, the polysheaths were tested for exposed thiol groups. Polysheaths formed by the wild-type gp053 or its mutants possess physicochemical properties, which are very attractive for the construction of self-assembling nanostructures with potential applications in different fields of nanosciences.

Glycine betaine (GB) could be used by Arthrobacter globiformis cells as a sole carbon source. The cells took up this molecule in the low as well as in the high salinity medium. Addition of GB to the mineral medium with high salt concentration revealed that GB was also used as an osmoprotectant. Dimethylglycine oxidase (DMGO) was involved in the catabolism of GB. Two genes for DMGO were detected in a cloned 26267 bp fragment of A. globiformis DNA. The genes involved in the tetrahydrofolate-dependent assimilation of methyl groups were located nearby the two of DMGO genes. Both cloned A. globiformis DMGO were active. The activity of DMGO was detected in A. globiformis cells and it depended on the addition of GB and the salinity of the medium. Reverse transcription-PCR demonstrated that the addition of GB influenced the transcription of dmg genes.