Chemical Glycobiology
Directed Evolution / in vitro Selection of Glycoligands
Many events in host-pathogen recognition, cell signaling, and cell adhesion are mediated by carbohydrate-protein interactions. We are therefore highly interested in the discovery of carbohydrate-containing ligands that bind tightly to a carbohydrate-binding protein of interest. Typically, such interactions are multivalent, requiring multiple copies of carbohydrate to achieve strong binding. We create highly diverse libraries in which a carbohydrate is clustered on trillions of different backbones (made of peptide, DNA or fluoro-RNA, see below). After multiple rounds of selection and amplification, tight binders of the target feature highly optimized clustering of carbohydrates, or cooperative binding of the carbohydrate and scaffold with the target. A major application of this work is to develop HIV-mimetic structures for investigation as HIV vaccines.
Carbohydrates Clustered on Peptide Backbones through mRNA Display
mRNA display is a technique that uses in vitro translation to create a library of peptides that are covalently labeled with their encoding mRNAs. Peptide-mRNA fusions that “win” a round of selection and bind to a target protein can be amplified by PCR to regenerate DNA/RNA for the next round of selection. We utilize unnatural amino acid incorporation techniques to make peptide libraries containing alkynyl amino acids, which serve as attachment points for carbohydrates using azide-alkyne click chemistry. This method has yielded multivalent glycopeptide ligands with low nM or pM dissociation constants and up to 360,000-fold binding enhancement versus the monovalent glycan.
Carbohydrates Clustered on DNA Aptamers (SELMA, SELection of Modified Aptamers)
Libraries of single stranded random-sequence DNA can be subjected to selection to isolate sequences that fold and bind to targets of interest. We have developed a technique (SELMA), in which alkynes are incorporated into the DNA and glycosylated by click chemistry with carbohydrate azides. The chemically modified glyco-DNA strand is covelently displayed on an unmodified natural DNA strand of analogous sequence, enabling PCR to retrieve selection winners despite the presence of chemical modifications. Selection with this method has yielded multivalent glycoDNA ligands with ~100,000-fold binding enhancement versus the monovalent glycan.
Carbohydrates Clustered on Nuclease-Stabilized Aptamers
We are using SELMA and similar newer display techniques to conduct in vitro selection of glycoligands with serum-stable modified oligonucleotide backbones, such as fluorinated RNA. In our newer method, RNA is transcribed off of a DNA template and anneals to a “capture strand”, remaining associated with the DNA template, which can be amplified by PCR. RNA structures (or modified RNA) that survive selection carry along their DNA template strands. Amplification of the DNA template circumvents the need for reverse transcription, which is difficult with modified RNA.
HIV Vaccine Studies
We are testing carbohydrate cluster glycopeptides derived from in vitro selection (see above projects) for the ability to raise antibodies against HIV.
Synthesis of Carbohydrates for Preparation of Libraries and Vaccines
The construction of libraries and vaccine materials involves the preparation of carbohydrates with chemical handles for facile attachment to peptide- or oligonucleotide backbones.
