If you have an ACS member number, please enter it here so we can link this account to your membership. (optional)

ACS values your privacy. By submitting your information, you are gaining access to C&EN and subscribing to our weekly newsletter. We use the information you provide to make your reading experience better, and we will never sell your data to third party members.


Biological Chemistry

Protein mirror images could generate long-lasting biologics

Peptides and proteins designed from D-protein helix database work longer in cells than natural biologic drugs

by Stu Borman
February 1, 2018 | A version of this story appeared in Volume 96, Issue 6

Schematic shows how mirror-image PDB and 2.8-million-helix database are created.
Credit: Adapted from PNAS
Technique generates a database of potentially protease-evading D-peptides in two easy steps.

Peptide- and protein-based biologic drugs have a problem. Enzymes in the body called proteases break down the linkages between L-amino acids in natural peptides and proteins. As a result, biologics typically have short lifetimes in the body and need to be injected or infused instead of taken as pills.

Synthetic analogs of natural peptides and proteins consisting of D-amino acids are safe from proteases. But drugmakers can’t just change L-amino acids to D-amino acids because this alters the orientation of the molecules’ sidechains, disrupting the way a peptide or protein drug binds to its target.

Two techniques, retroinversion and mirror-image phage display (MIPD), can create D-amino-acid analogs of natural bioactive peptides and proteins that bind targets effectively. But retroinversion—reversing a peptide’s sequence and changing its amino acids to D-versions—is ineffective for peptides with α-helical binding motifs because it doesn’t change the direction of helix rotation, causing helical binding features to be oriented differently than in natural peptides. And MIPD—screening libraries to find L-peptides that bind D-versions of drug targets and then synthesizing the corresponding D-peptides—doesn’t work for large protein targets such as G protein-coupled receptors because those targets are too hard to synthesize.

Philip M. Kim, postdoc Michael Garton, and coworkers at the University of Toronto have now devised a way to make D and D-protein analogs that can bind most biological targets (Proc. Natl. Acad. Sci. USA 2018, DOI: 10.1073/pnas.1711837115).

They computationally generated a D version of every protein in the Protein Data Bank (PDB), creating the D-PDB, and extracted the D-proteins’ α-helices into more than 2.8 million separate database files. They then used known drug-target interactions to screen the helix database for D-helices with binding features positioned similarly to those of natural peptide and protein drugs. To create matches for drugs that bind in complex ways, the researchers made short D-strands by retroinversion and used the strands to link D-helices into three-part D-analogs.

Kim and coworkers used the method to create D-analogs for GLP-1, a diabetes and obesity treatment that targets the GLP-1 receptor, and parathyroid hormone, an osteoporosis medication that hits the parathyroid receptor. The D-analogs had about the same efficacy as their natural counterparts in cells, although the GLP-1 replacement required a higher dose. And the D-analogs withstood the cells’ proteases for longer than the natural peptides.

Danny Hung-Chieh Chou of the University of Utah notes that the technique is limited to drugs with a known binding mechanism but calls it “a great concept that solves a problem existing technology couldn’t handle.” Bradley L. Pentelute of MIT adds that he believes the approach “will be a very useful tool for the biotech community.”



This article has been sent to the following recipient:

Chemistry matters. Join us to get the news you need.