Project Summary/Abstract
We have developed a technique for randomizing 23S ribosomal RNA structure, and for selecting modified
ribosomes which incorporate into proteins specific types of modified amino acids not ordinarily incorporated by
wild-type ribosomes. Species incorporated both in vitro and in vivo have included nucleobase amino acids,
dipeptides/dipeptidomimetics, beta-amino acids, phosphorylated amino acids and glycosylated amino acids.
The selected ribosomes enable study of two key biochemical regulatory processes, i.e. protein glycosylation
and phosphorylation. We will also modify regulatory proteins that interact with nucleic acids, enabling
predictable modulation or altered specificity of interactions. We will exemplify new strategies using proteins
containing unusual non-proteinogenic amino acids, whose incorporation requires our selected ribosomes.
The creation of proteins phosphorylated stoichiometrically at single or multiple positions affords new
opportunities. These include the ability to verify natural phosphorylations, and to study their effects. It permits
the introduction of (metabolically stable) phosphate groups in vivo, and has enabled new strategies for
identifying residues whose phosphorylation modifies function. We showed that phosphorylation of IB- Tyr42
not only relaxes NF-B inhibition, but facilitates the rate of binding to a gene whose expression NF-B
regulates. We plan to study two other known phosphorylation sites in IB-, and three in NF-BWhile some
sites of serine phosphorylation in NF-B are known, this is not true for Tyr and Thr. Using a new strategy, we
have identified four sites of Tyr phosphorylation, and plan to study new Thr and Ser phosphorylation sites.
Most mammalian proteins are glycosylated, but understanding/altering carbohydrate functions is challenging.
Using a selected ribosome in a cell free system, we prepared murine interferon- (IFN-) containing GlcNAc-
Asn at position 29, which can confer antiviral activity. This intermediate acceptor substrate should enable
carbohydrate cluster transfer, producing IFN- fully glycosylated at position 29, and expected to have antiviral
activity. We have recently introduced the same glycosylated amino acid, and its peracetylated analogue, into a
model protein in very good yield in cellulo. This strategy can potentially produce the same intermediate as that
prepared in vitro, but in much larger quantities. We wish to prepare fully glycosylated proteins with natural
carbohydrate clusters, to simplify these clusters, and to focus on the inclusion of residues such as sialic acid.
Cell regulation often involves protein–DNA interactions; low nanomolar affinities and impressive selectivity
are typical. Many X-ray crystal structures could guide design changes, but attempts to alter these regulatory
processes by changing affinity or specificity have failed. Using Rob proteins having nucleobase amino acids,
we modified binding to their micF DNA partners, realizing stronger binding, and enhanced phenotypic cellular
responses. We propose to study the Rob–micF DNA interactions further to refine our design techniques. The
principles developed will be used to address protein–DNA recognition in a Zn finger system.