 Hi, I'm Chris Beistroff, last author on green lighting, green fluorescent protein, faster and more efficient folding by eliminating a cis-trans peptide isomerization event published in Protein Science. In brief, my lab computationally designed and characterized a variant of GFP that has some new and useful properties. Peptide isomerization is the rotation of a peptide bond from cis to trans, a slow process with a halftime in the range of hundreds of seconds. Cis-trans isomerizations occur mainly at pre-proline positions. GFP is an 11-stranded beta-barrel with a distorted alpha helix going through the center. The barrel structure positions internal side chains to catalyze the cyclization and oxidation of the backbone of three central residues in that helix to create the fluorescent chromophore. The folding of GFP is multi-phasic, having fast, medium, and slow phases implying parallel folding pathways. The slow phase was shown by Kuajima to be a peptide bond isomerization event in the unfolded state associated with a proline. Basework by Jennings in 2009 found that the refolding of GFP by guanidine dilution left 61% of the protein in a trapped, non-fluorescent state. In the trapped state, the chromophore appeared to be mobile as judged by the lack of splitting of NMR resonances on the chromophore. Based on the effects of point mutations, they suggested that three highly conserved prolings at positions 58, 75, and 89 were required for the efficient maturation of the GFP chromophore. We reasoned that one of these three prolings must be responsible for the non-fluorescent trapped state, and we set out to test that hypothesis. Only Pro 89 is in the cis confirmation in GFP, therefore we hypothesized that Pro 89's cis-peptide bond isomerization was responsible for the trapped state and also for the slowest step in GFP's folding kinetics. Since others had already attempted point mutations at Pro 89 and found them to be unstable, we decided to eliminate the cis-peptide by rebuilding the loop. We added two residues, then we searched the database for likely confirmations for the new loop. We then used in-house software to design the side chains of the new loop, having no cis-peptides in it. The protein was synthesized in the lab and, unlike most experiments in my lab, it worked like a charm, producing a brightly glowing green protein. Its only drawback was a 40% decrease in brightness due to both lower chromophore maturation efficiency and lower quantum yield. We then solved the crystal structure. The crystals had five molecules in the asymmetric unit, and in each copy we could see clearly that the newly designed GFP variant had no cis-peptide bonds in the region around Pro 89, and since it has no cis-peptide bonds elsewhere either, we called it all-trans GFP or AT GFP. The structure is available from the PDB as 4LW5. Then we refolded the protein by quantity dilution and compared it to the GFP variant from which it was derived, a super-efficient folder generated by the Waldo lab called OPT GFP. AT GFP recovered much more of its original fluorescence than OPT GFP. And when the data was fit to multiple kinetic phases, it was found to have the same three fast phases of folding, but the slow phase was faster and decreased in its relative amplitude. If we refolded the protein by pH jump instead of quantity dilution, we saw that the slow phase was completely eliminated. Read the paper if you want to know how we explain this discrepancy. Then we tested to see if the protein still fell into the long-lived trapped state. We selected the midpoints of unfolding using both pH and quantity in separate experiments. GFPs were titrated to the midpoint from either the folded state or from the unfolded state and left to equilibrate for at least 96 hours. Then we measured the fluorescence. The difference between the protein that was titrated from the unfolded state versus the one that was titrated from the folded state is a measure of hysteresis which quantifies how much of the protein is misfolding into a trapped non-fluorescent state. Different with the previously published results, OptGFP showed a significant loss of fluorescence due to hysteresis, but to our amazement we saw that all the hysteresis was eliminated in ATGFP. This was true whether we used guanidine or pH. We conclude that the trapped state must be a near-native conformation with a transpeptide bond at Pro 89. It cannot isomerize to the native cis conformation due to its kinetic stability in the folded state. An image of the folding landscape of GFP is emerging from these studies in which the trapped trans Pro 89 state cannot reach the native state without passing through the unfolded state and the native state folding pathway consists of two channels corresponding to two folding intermediates. The removal of the cis trans isomerization at Pro 89 leaves a simpler energy landscape. ATGFP may prove useful wherever the robust folding of GFP is important for its application. Thanks for watching.