Early Work
Research in the Schepartz
laboratory focuses on the chemical
biology of protein–protein and protein–DNA interactions
inside the cell. We are interested in the structural and energetic
factors that distinguish specific protein interfaces and how the
assembly (and disassembly) of these complexes define biology.
Specific
questions our lab focuses on include how cells effectively use a
limited number of transcription factors to achieve a precisely controlled
and robust gene-regulatory network, how this network is usurped
when cells succumb to viral attack, and how, inspired by the viral
hijackers, we can design miniature proteins that mimic (and sometimes
surpass) the functional properties of proteins found in Nature.
For a description of very early work in the Schepartz lab, click
here.
We are very interested in understanding the detailed mechanisms
by which transcription factors identify their preferred target sites
within a sea of genomic DNA and unrelated protein. These studies
began with Bernard
Cuenoud’s Science report of a series of molecules in which
the relative orientations of two peptides from the DNA-binding region
of GCN4 (a basic region-leucine zipper transcription factor) were
varied with a structurally tunable metal complex.
The properties of these molecules suggested that the sequence specificities
of bZIP proteins might depend on the orientations of their DNA-binding
regions.
Steven
Metallo later verified this hypothesis, which represented a
fundamentally new mechanism for controlling the sequence specificity
of DNA-binding proteins.
Another novel mechanism for controlling sequence specificity was
reported by David Paolella
in Science. He discovered that an intrinsic bend could pre-organize
DNA for binding by some proteins and not others. 
Steven Metallo and Jennifer
Kohler recently described a third mechanism by which specificity
— in this case, kinetic specificity — can be achieved.
They
discovered that many bZIP and bHLHzip proteins bind DNA as monomers,
dimerizing while DNA-bound, and that this monomer-binding pathway
represents the most rapid search method for locating specific DNA.

Our group is also very interested in how bZIP proteins are recognized
and discriminated by viral proteins that hijack cellular transcriptional
machinery for viral replication. Anne
Baranger reported in Nature
that the HTLV-I Tax protein increases the affinity of a cellular
bZIP protein for viral DNA by altering the specificity of DNA binding,
not by enhancing bZIP dimerization.
Subsequent work by
Rodgers Palmer demonstrated that the HBV X protein employs an
identical mechanism, despite the evolutionary distance between HBV
and HTLV-I and the lack of sequence homology between Tax and pX.
Though viral proteins have evolved to commandeer cellular transcription
factors for viral replication, the resultant side effects on cellular
transcription are responsible for producing diseased states. Tanya
Schneider showed that pX diminishes the inherent specificity
of the monomer-binding pathway of bZIPs, providing a molecular explanation
for aberrant transcriptional activation during HBV infection.
Most recently, Mary Kay Pflum
reported that pX bypasses the need for phosphorylation to activate
certain bZIP proteins, short-circuiting a natural signal transduction
pathway linking elevated cAMP levels and transcriptional activation.
Her work provides a direct link between HBV infection and the development
of hepatocellular carcinoma. 
Guided
creatively by insights gleaned from our work on Tax and pX, our
most recent efforts focus on the design of molecules that bind protein
surfaces and inhibit protein-protein interactions inside the cell.
Our interest in this area stems from the over-arching challenge
and opportunity posed by the sequencing of the human
genome. The actions and interactions of the proteins encoded
by these 35,000 genes — whose numbers could reach 350,000
or more when post-translational modifications, alternative splicing,
and proteolytic cleavage are considered — determine the full
range of cellular activities in normal cell processes and in disease.
Our laboratory has described a general solution to the design of
molecules capable of deconvoluting the role of transient protein–protein
interactions inside the cell. Building on the observation that proteins
generally recognize each other using extended, shallow, hydrophilic
interfaces, we proposed that the most appropriate ligands for protein
surfaces might be a miniature (but well folded) protein that presents
a preorganized functional epitope in a very
small package. Initially Neal
Zondlo and Jason Chin
reported in JACS
the successful use of protein grafting in the design and evolution
of DNA-binding miniature proteins based on the structure of GCN4.
These
molecules, containing only 36 residues, bind specific DNA under
physiological conditions with comparable affinities and higher specificities
that GCN4 itself. This work was highlighted in both Science
and Nature.
More recently Jin Montclare
reported in JACS that it is possible to miniaturize both the recognition
surface and the structural framework of a globular protein fold
through the design of a miniature homeodomain that binds specific
DNA without benefit of an “N-terminal arm.”
Jason Chin most recently
has translated this research to the area of protein recognition
by the design and discovery of miniature proteins that bind with
nanomolar affinity to the anti-apoptotic proteins Bcl-2 and Bcl-XL.
Stacey Rutledge
and Heather Volkman
have shown that miniature proteins possessing the elusive goal of
high affinity and specificity can be achieved even in cases where
the binding site is exceptionally featureless and when the ligand
is phosphorylated.
Dasantila Golemi
has shown that miniature proteins posessing paralog specificity
can be designed de novo and can distinguish members of
a protein family that are 60 percent identical and structurally
superimposable.
The newest miniature proteins use the PPII helix to distinguish
between EVH1-domains and induce a unique phenotype in mammalian
cells. 
The observation that miniature protein ligands achieve the elusive
goal of high affinity and (even more importantly) specificity suggests
that could have enormous utility as probes of protein–protein
interactions inside the cell, achieving a phenotypic precision that
is difficult to achieve any other way. Many current projects in our
lab are focused in this area and in the use of proteome chips to
evaluate and aid miniature protein design.
|