Dominik Escher

Correlation between in Vitro Stability and in Vivo Performance of Anti-GCN4 Intrabodies as Cytoplasmic Inhibitors

A cellular assay system for measuring the activity of cytoplasmically expressed anti-GCN4 scFv fragments directed against the Gcn4p dimerization domain was established in the budding yeast Saccharomyces cerevisiae. The inhibitory potential of different constitutively expressed anti-GCN4 scFv intrabodies was monitored by measuring the activity of β-galactosidase expressed from a GCN4-dependent reporter gene. The in vivo performance of these scFv intrabodies in specifically decreasing reporter gene activity was related to their in vitro stability, measured by denaturant-induced equilibrium unfolding. A framework-engineered stabilized version showed significantly improved activity, while a destabilized point mutant of the anti-GCN4 wild-type showed decreased effects in vivo. These results indicate that stability engineering can result in improved performance of scFv fragments as intrabodies. Increasing the thermodynamic stability appears to be an essential factor for improving the yield of functional scFv in the reducing environment of the cytoplasm, where the conserved intradomain disulfides of antibody fragments cannot form.

Antigen-independent selection of stable intracellular single-chain antibodies.

The intracellular expression of single‐chain Fv antibody fragments (scFv) in eukaryotic cells has an enormous potential in functional genomics and therapeutics [Marasco (1997) Gene Ther. 4, 11–15; Richardson and Marasco (1995) Trends Biotechnol. 13, 306–310]. However, the application of these so‐called intrabodies is currently limited by their unpredictable behavior under the reducing conditions encountered inside eukaryotic cells, which can affect their stability and solubility properties [Wörn et al. (2000) J. Biol. Chem. 275, 2795–2803; Biocca et al. (1995) Bio/Technology 13, 1110–1115]. We present a novel system that enables selection of stable and soluble intrabody frameworks in vivo without the requirement or knowledge of antigens. This system is based on the expression of single‐chain antibodies fused to a selectable marker that can control gene expression and cell growth. Our results show that the activity of a selectable marker fused to well characterized scFvs [Wörn et al. (2000) J. Biol. Chem. 275, 2795–2803] correlates with the solubility and stability of the scFv moieties. This method provides a unique tool to identify stable and soluble scFv frameworks, which subsequently serve as acceptor backbones to construct intrabody complementarity determining region libraries by randomization of hypervariable loops.

An adenovirus vector with a chimeric fiber incorporating stabilized single chain antibody achieves targeted gene delivery

Adenovirus (Ad) vectors are of utility for many therapeutic applications. Strategies have been developed to alter adenoviral tropism to achieve a cell-specific gene delivery capacity employing fiber modifications allowing genetic incorporation of targeting motifs. In this regard, single chain antibodies (scFv) represent potentially useful agents to achieve targeted gene transfer. However, the distinct biosynthetic pathways that scFv and Ad capsid proteins are normally routed through have thus far been problematic with respect to scFv incorporation into the Ad capsid. Utilization of stable scFv, which also maintain correct folding and thus functionality under intracellular reducing conditions, could overcome this restriction. We genetically incorporated a stable scFv into a de-knobbed, fibritin-foldon trimerized Ad fiber and demonstrated selective targeting to the cognate epitope expressed on the membrane surface of cells. We have shown that the scFv employed in this study retains functionality and that stabilizing the targeting molecule, per se, is critical to allow retention of antigen recognition in the adenovirus capsid-incorporated context.

Conservation of Glutamine-Rich Transactivation Function between Yeast and Humans

ABSTRACT Several eukaryotic transcription factors such as Sp1 or Oct1 contain glutamine-rich domains that mediate transcriptional activation. In human cells, promoter-proximally bound glutamine-rich activation domains activate transcription poorly in the absence of acidic type activators bound at distal enhancers, but synergistically stimulate transcription with these remote activators. Glutamine-rich activation domains were previously reported to also function in the fission yeastSchizosaccharomyces pombe but not in the budding yeastSaccharomyces cerevisiae, suggesting that budding yeast lacks this pathway of transcriptional activation. The strong interaction of an Sp1 glutamine-rich domain with the general transcription factor TAFII110 (TAFII130), and the absence of any obvious TAFII110 homologue in the budding yeast genome, seemed to confirm this notion. We reinvestigated the phenomenon by reconstituting in the budding yeast an enhancer-promoter architecture that is prevalent in higher eukaryotes but less common in yeast. Under these conditions, we observed that glutamine-rich activation domains derived from both mammalian and yeast transcription factors activated only poorly on their own but strongly synergized with acidic activators bound at the remote enhancer position. The level of activation by the glutamine-rich activation domains of Sp1 and Oct1 in combination with a remote enhancer was similar in yeast and human cells. We also found that mutations in a glutamine-rich domain had similar phenotypes in budding yeast and human cells. Our results show that glutamine-rich activation domains behave very similarly in yeast and mammals and that their activity in budding yeast does not depend on the presence of a TAFII110 homologue.

Gene activation at a distance and telomeric silencing are not affected by yeast histone H1

Abstract Until recently, it was believed that the budding yeast Saccharomyces cerevisiae has no histone H1 gene. However, a search of the yeast genome database revealed a possible H1 homologue of 258 amino acids, termed yeast histone H1 (HHO1). The protein shows 36% identity to the human H1 core domain over a stretch of 93 amino acids. Unlike other H1 proteins, Hho1p has a second possible core domain which shows 43% identity to the first core domain. Since vertebrate H1 histone had been implied in gene repression as well as gene activation at a distance, we tested the effect of deleting the yeast H1-like gene on remote activation of a modified GAL1 promoter, which contains a synthetic GAL4 binding site close to the TATA box, and the natural UASG, consisting of four GAL4 binding sites. Different spacing up to 1.8 kb between the proximal binding site and the distal UASG enhancer revealed no differences in gene activation between wild-type and knockout strains. Overexpression of a heterologous histone H1 from sea urchin showed an overall inhibition of gene activation by the GAL1 promoter, whereas overexpression of the yeast histone H1 had no effect. Also, the expression of A1, ALPHA2 or SUC2 genes, all of which are known to be responsive to an altered chromatin structure, was unchanged in HHO1 knockout or HHO1-overexpressing strains when compared to wild-type cells. We also considered the possibility that HHO1 was involved in forming the heterochromatin at telomeres. On testing for telomeric silencing of a URA reporter gene introduced 1.3 kb away from the chromosomal end, we again observed no differences between wild-type and knockout strains. Thus, the yeast histone H1-like gene appears to have no role in gene activation at a distance or in silencing under the conditions tested. It remains to be seen whether the yeast H1 histone is a gene-specific regulator rather than a general chromatin-associated protein.

A trans ‐acting peptide activates the yeast a1 repressor by raising its DNA‐binding affinity

The cooperative binding of gene regulatory proteins to DNA is a common feature of transcriptional control in both prokaryotes and eukaryotes. It is generally viewed as a simple energy coupling, through protein–protein interactions, of two or more DNA‐binding proteins. In this paper, we show that the simple view does not account for the cooperative DNA binding of a1 and α2, two homeodomain proteins from budding yeast. Rather, we show through the use of chimeric proteins and synthetic peptides that, upon heterodimerization, α2 instructs a1 to bind DNA. This change is induced by contact with a peptide contributed by α2, and this contact converts a1 from a weak to a strong DNA‐binding protein. This explains, in part, how high DNA–binding specificity is achieved only when the two gene regulatory proteins conjoin. We also provide evidence that features of the a1–α2 interaction can serve as a model for other examples of protein–protein interactions, including that between the herpes virus transcriptional activator VP16 and the mammalian homeodomain‐containing protein Oct‐l.