Posted as an aid to your understanding.

mRNA Display Design of Fibronectin-based Intrabodies That Detect and Inhibit Severe Acute Respiratory Syndrome Coronavirus Nucleocapsid Protein*

Hsiang-I. Liao‡,1,
C. Anders Olson§,1,
Seungmin Hwang‡,
Hongyu Deng‡,
Elaine Wong‡,
Ralph S. Baric¶,
Richard W. Roberts‖ and
Ren Sun‡,**,2

+
Author Affiliations
From the ‡Department of Molecular and Medical Pharmacology and
the **California Nano System Institute, UCLA, Los Angeles, California 90095,
§Biochemistry and Molecular Biophysics Option, California Institute of Technology, Pasadena, California 91125,
the ¶Department of Epidemiology, University of North Carolina, Chapel Hill, North Carolina 27599, and
the ‖Department of Chemistry, Chemical Engineering, and Biology, University of Southern California, Los Angeles, California 90089-1211
↵2To whom correspondence should be addressed: 650 Charles E. Young Dr. South, CHS 23-120, UCLA, Los Angeles, CA 90095. Fax: 310-825-6267; E-mail: rsun@mednet.ucla.edu.

↵1 Both authors contributed equally to this work.

From http://www.jbc.org/content/284/26/17512.full

…The ability to detect and inhibit protein function is central to molecular and cellular biology research. To date, phage display and monoclonal antibody production have been the most common routes to design reagents for protein detection and inhibition, antibodies and antibody-like reagents that serve as high affinity, high specificity molecular recognition tools (1). Totally in vitro selection methods using alternative scaffolds are becoming more common to produce affinity reagents with improved and expanded functionality (2, 3). For example, ribosome display and mRNA display enable creating 1–100 trillion-member peptide and protein libraries that surpass immunological and phage display diversities by 3–5 orders of magnitude (4).

Antibodies or antibody-like molecules are important because they can serve as diagnostics, probes for studying proteins in vivo, and potential therapeutics (or surrogate ligands for therapeutic design/screening). Regarding biology, antibodies used inside living cells, denoted “intrabodies,” are appealing because they provide an alternative to genetic knock-outs, dominant negative mutations, and RNA interference strategies, enabling targeting proteins in a domain-, conformation-, and modification-specific fashion as well as identifying hot spots for protein interaction (5, 6). For example, green fluorescent protein-labeled intrabodies can act as molecular beacons to determine real time, live cell localization of endogenous target proteins rather than non-native expression of green fluorescent protein target fusions (7).

Although antibodies often demonstrate laudable affinity and selectivity, these proteins are likely to be suboptimal as a general approach to create intracellular reagents. Most notably, antibodies contain disulfide bonds that are likely to be reduced in the cytosol, thus impeding their proper folding and function (8). To overcome the paucity of functional intrabodies generated by in vitro selection methods, in vivo screens may be employed at the expense of combinatorial diversity (9). On the other hand, it has been demonstrated that intracellular antibodies can generate aggresomes, which may inhibit the ubiquitin-mediated degradation pathway and promote apoptosis (10–12).

Ideally, intrabodies would be as follows: 1) easy to produce in a broad variety of cells; 2) stable; 3) specific; 4) high affinity; 5) highly selective; 6) functional in intracellular environments; and 7) noninterfering with normal cellular processes. Recently, ribosome display has been used to generate protein affinity reagents based on ankyrin domains (DARPins), which detect and inhibit kinase or proteinase function in vivo (13, 14). Although this scaffold is powerful, it is structurally very different from antibodies as it utilizes a discontinuous binding surface rather than the continuous surface generated by the CDR loops in antibody VH and VL domains.

Our approach here has been to use mRNA display to design disulfide-free antibody-like proteins that can be used to create general protein targeting tools. To do this, we used a protein library based on the 10th fibronectin type III domain of human fibronectin (10Fn3)3 (15, 16). The 10Fn3 domain was developed as an antibody mimetic by Koide et al. (16) because of the following: 1) it is topologically analogous to the immunoglobulin VH domain; 2) it is exceptionally stable; 3) it presents a continuous protein interaction surface; and 4) it expresses well in both eukaryotic and bacterial cells (16). We recently described construction and characterization of a 3 × 1013 member 10Fn3 library (15) and validated this library by developing proteins and fluorescence resonance energy transfer sensors that recognize IκBα in a phosphoserine-specific fashion (17). There the selected 10Fn3 functioned in vivo, blocking proteasome-mediated degradation of full-length IκBα efficiently.

Here we have targeted the severe acute respiratory syndrome (SARS-CoV) nucleocapsid protein (N). SARS-CoV is a unique member of the Coronaviridae family with only 20% sequence identity to the closest homolog (18). There is a need for reagents and methods that can be used to detect new infectious entities as they arise. Indeed, the recent SARS epidemic was unexpected, reaching an 8% fatality rate despite the fact that coronaviruses typically are involved in ∼30% of common cold infections. N protein is 422 amino acids long, phosphorylated, and composed of two structured domains linked by a nonstructured domain. The N-terminal domain (NTD) is a putative RNA binding domain, and the C-terminal domain (CTD) mediates self-association (Fig. 1A) (19, 20). The unstructured middle domain interacts with the membrane (M) protein, anchoring M protein to the viral core. The two structured domains act in concert to bind genomic RNA, oligomerize, and form the final packaged ribonucleoprotein complex.

We chose N as our target for several reasons. First, the N protein is the most abundant protein produced by SARS virus. Second, N plays multiple roles in vivo, including binding/packaging the viral genomic RNA, mediating interactions with the viral membrane (via the M protein), acting in genome replication, and exerting control over host cell processes (21, 22). Finally, no therapeutic reagents currently target N protein; therefore, new inhibitory N-directed ligands represent an important potential new route for developing anti-SARS drugs.

After six rounds of selection, we were able to generate molecules that detect SARS N protein in vitro and modulate its SARS replication in vivo in a domain-specific manner. The selection yielded six high affinity molecules that recognize the CTD and two molecules that require the NTD for binding. We confirmed the interaction between the selected 10Fn3 proteins and N protein both in vitro and in vivo by pulldown, co-immunoprecipitation, and immunofluorescence microscopy. Seven of the 10Fn3-based intrabodies inhibit replication, ranging from 11- to 5900-fold, recognizing at least two nonoverlapping epitopes/hot spots in a synergistic manner. These molecules represent new tools for detecting SARS virus, assessing N function in living cells, and identifying regions of N critical for virus proliferation.

Abstract follows…