The unexpected structure of the designed protein Octarellin V.1 forms a challenge for protein structure prediction tools.

Despite impressive successes in protein design, designing a well-folded protein of more 100 amino acids de novo remains a formidable challenge. Exploiting the promising biophysical features of the artificial protein Octarellin V, we improved this protein by directed evolution, thus creating a more stable and soluble protein: Octarellin V.1. Next, we obtained crystals of Octarellin V.1 in complex with crystallization chaperons and determined the tertiary structure. The experimental structure of Octarellin V.1 differs from its in silico design: the (αßα) sandwich architecture bears some resemblance to a Rossman-like fold instead of the intended TIM-barrel fold. This surprising result gave us a unique and attractive opportunity to test the state of the art in protein structure prediction, using this artificial protein free of any natural selection. We tested 13 automated webservers for protein structure prediction and found none of them to predict the actual structure. More than 50% of them predicted a TIM-barrel fold, i.e. the structure we set out to design more than 10years ago. In addition, local software runs that are human operated can sample a structure similar to the experimental one but fail in selecting it, suggesting that the scoring and ranking functions should be improved. We propose that artificial proteins could be used as tools to test the accuracy of protein structure prediction algorithms, because their lack of evolutionary pressure and unique sequences features.

Two abundant proteasome subtypes that uniquely process some antigens presented by HLA class I molecules.

Most antigenic peptides presented by MHC class I molecules result from the degradation of intracellular proteins by the proteasome. In lymphoid tissues and cells exposed to IFNγ, the standard proteasome is replaced by the immunoproteasome, in which all of the standard catalytic subunits ß1, ß2, and ß5 are replaced by their inducible counterparts ß1i, ß2i, and ß5i, which have different cleavage specificities. The immunoproteasome thereby shapes the repertoire of antigenic peptides. The existence of additional forms of proteasomes bearing a mixed assortment of standard and inducible catalytic subunits has been suggested. Using a new set of unique subunit-specific antibodies, we have now isolated, quantified, and characterized human proteasomes that are intermediate between the standard proteasome and the immunoproteasome. They contain only one (ß5i) or two (ß1i and ß5i) of the three inducible catalytic subunits of the immunoproteasome. These intermediate proteasomes represent between one-third and one-half of the proteasome content of human liver, colon, small intestine, and kidney. They are also present in human tumor cells and dendritic cells. We identified two tumor antigens of clinical interest that are processed exclusively either by intermediate proteasomes ß5i (MAGE-A3(271-279)) or by intermediate proteasomes ß1i-ß5i (MAGE-A10(254-262)). The existence of these intermediate proteasomes broadens the repertoire of antigens presented to CD8 T cells and implies that the antigens presented by a given cell depend on their proteasome content.

Silencing of cancer-germline genes in human preimplantation embryos: evidence for active de novo DNA methylation in stem cells.

Several human germline-specific genes rely principally on DNA methylation for repression in somatic tissues. Many of these genes, including MAGEA1, were qualified as cancer-germline (CG), as they become activated in tumors, where losses of DNA methylation are common. The developmental stage at which CG genes acquire DNA methylation marks is unknown. Here, we show that in human preimplantation embryos, transcription of CG genes increases up to the morula stage, and then decreases dramatically in blastocysts, suggesting that CG gene silencing occurs in blastocyst stem cells. Consistently, transfection studies with MAGEA1 constructs in embryonal carcinoma (EC) cells, which represent a malignant surrogate of blastocyst-derived stem cells, revealed active repression and marked de novo methylation of MAGEA1 transgenes in these cells. Active repression of the endogenous MAGEA1 gene in human EC cells was evidenced by its rapid re-silencing following prior induction with a DNA methylation inhibitor. Moreover, de novo DNA methyltransferases DNMT3A and DNMT3B appeared to contribute to the silencing of MAGEA1 and other CG genes in EC cells. Altogether our data indicate that CG genes like MAGEA1 are programmed for repression in the blastocyst, and suggest that de novo DNA methylation is a key event in this process.

DNA methylation-associated repression of cancer-germline genes in human embryonic and adult stem cells.

Cancer-germline (CG) genes are a particular group of germline-specific genes that rely primarily on DNA methylation for repression in somatic tissues. In a wide variety of tumors, the promoter of these genes is demethylated, and their transcription is activated. The mechanism underlying this tumor-specific activation is still unclear. It was recently suggested that CG gene expression may be a hallmark of stem cells, and that expression of these genes in several tumors may reflect the expansion of constitutively expressing cancer stem cells. To clarify this issue, we carefully evaluated the expression of several CG genes in human stem cells of embryonic and adult origin. We found no or very weak expression of CG genes in these cells. Consistently, the promoter of CG genes was highly methylated in these cells. We conclude that CG genes do not qualify as "stemness" genes, and propose that their activation in cancers results from a tumor-specific activation process.

Epigenetic hierarchy within the MAGEA1 cancer-germline gene: promoter DNA methylation dictates local histone modifications.

Gene MAGEA1 belongs to a group of human germline-specific genes that rely on DNA methylation for repression in somatic tissues. Many of these genes, termed cancer-germline (CG) genes, become demethylated and activated in a wide variety of tumors, where they encode tumor-specific antigens. The process leading to DNA demethylation of CG genes in tumors remains unclear. Previous data suggested that histone acetylation might be involved. Here, we investigated the relative contribution of DNA methylation and histone acetylation in the epigenetic regulation of gene MAGEA1. We show that MAGEA1 DNA hypomethylation in expressing melanoma cells is indeed correlated with local increases in histone H3 acetylation (H3ac). However, when MAGEA1-negative cells were exposed to a histone deacetylase inhibitor (TSA), we observed only short-term activation of the gene and detected no demethylation of its promoter. As a more sensitive assay, we used a cell clone harboring a methylated MAGEA1/hph construct, which confers resistance to hygromycin upon stable re-activation. TSA induced only transient de-repression of the transgene, and did not lead to the emergence of hygromycin-resistant cells. In striking contrast, transient depletion of DNA-methyltransferase-1 in the reporter cell clone gave rise to a hygromycin-resistant population, in which the re-activated MAGEA1/hph transgene displayed not only marked DNA hypomethylation, but also significant reversal of histone marks, including gains in H3ac and H3K4me2, and losses of H3K9me2. Collectively, our results indicate that DNA methylation has a dominant role in the epigenetic hierarchy governing MAGEA1 expression.