Selection and evolution of enzymes from a partially randomized non-catalytic scaffold

Abstract

Enzymes are exceptional catalysts that facilitate a wide variety of reactions under mild conditions, achieving high rate-enhancements with excellent chemo-, regio- and stereoselectivities. There is considerable interest in developing new enzymes for the synthesis of chemicals and pharmaceuticals and as tools for molecular biology. Methods have been developed for modifying and improving existing enzymes through screening, selection and directed evolution. However, the design and evolution of truly novel enzymes has relied on extensive knowledge of the mechanism of the reaction. Here we show that genuinely new enzymatic activities can be created de novo without the need for prior mechanistic information by selection from a naive protein library of very high diversity, with product formation as the sole selection criterion. We used messenger RNA display, in which proteins are covalently linked to their encoding mRNA, to select for functional proteins from an in vitro translated protein library of >10(12 )independent sequences without the constraints imposed by any in vivo step. This technique has been used to evolve new peptides and proteins that can bind a specific ligand, from both random-sequence libraries and libraries based on a known protein fold. We now describe the isolation of novel RNA ligases from a library that is based on a zinc finger scaffold, followed by in vitro directed evolution to further optimize these enzymes. The resulting ligases exhibit multiple turnover with rate enhancements of more than two-million-fold.

Figure 1. In vitro selection of enzymes by mRNA-display

a, General selection scheme for enzymes for bond-forming reactions. A DNA library is transcribed into RNA, cross-linked to a 3′-puromycin oligonucleotide, and in vitro translated. The library of mRNA-displayed proteins is reverse transcribed with a primer bearing substrate A. Substrate B, which carries an anchor group, is added. Proteins that join A and B attach the anchor group to their encoding cDNA. Selected cDNA sequences are then amplified by PCR, and used as input for the next round. b, Selection of enzymes that perform template-dependent ligation of a 5′-triphosphate-activated RNA (PPP-substrate) to a second RNA (HO-substrate). The PPP-substrate is ligated to the primer and then used in the reverse transcription reaction. The cDNA of the catalytically active molecules is immobilized on streptavidin-coated beads via biotin, washed, and released by UV-irradiation of the photocleavable linker (PC). c, The scaffolded library is based on a two zinc finger domain with two loop regions (light blue) that are replaced by segments of 12 or 9 random amino acids.

Figure 2. Progress of the selection

The fraction of ³²P-labelled cDNA that bound to streptavidin agarose (SA) and eluted after photocleavage at each round of selection is shown. The input DNA into rounds 9*, 10* and 11* was subjected to mutagenic PCR amplification and, in addition, a recombination procedure was performed before rounds 9* and 11*. The selection pressure was increased by decreasing the time of the reaction as indicated. Asterisks indicate selection rounds after mutagenesis and recombination.

Figure 3. Sequences of the starting library and selected ligases

Loop regions are highlighted in light blue. The cysteines highlighted in orange constitute the two pairs of CX_nC (n = 2 or 5) motifs that coordinate zinc ions in the original hRXR domain. Randomized amino acids in the library are shown as x. Dashes indicate amino acids that are the same as in the starting library, whereas periods highlighted in grey symbolize deletions. The underlined flanking regions were not part of the hRXRα domain but were added to contain a Flag epitope tag, a hexahistidine tag and a linker region.

Figure 4. Characterization of ligase enzyme

a, The RNA ligation reaction. b, Reaction catalyzed by ligase #4 after 1, 3 and 10 hrs (lanes 1,2,3). Lanes 4–7: 10 hrs with no splint, 5′-monophosphate instead of PPP-substrate; 5′-hydroxyl instead of PPP-substrate, and wild-type hRXRα protein domain instead of ligase #4. c, Release of inorganic pyrophosphate during ligation. Ligation reactions with γ-³²P GTP-labelled PPP-substrate were separated by thin-layer chromatography. A mixture of inorganic ³²P-phosphate (Pi), ³²P-pyrophosphate (PPi) and 5′-γ-³²P-labelled PPP-substrate was run for reference (Ref.). d, 3′-5′ regiospecificity of ligation. Ligation of α-³²P GTP body-labelled PPP-substrate yielded product with ³²P at the indicated (*) positions. The product was digested to nucleoside monophosphates with ribonuclease T2 (which does not efficiently digest 2′,5′ RNA linkages) in the presence of a 22-nt chemically synthesized RNA identical in sequence to the predicted ligation product but which contains a 2′-5′ linkage at the ligation junction (5′-CUAACGUUCGC^2′p^5′GGAGACUCUUU). Digestion products were separated by two-dimensional thin-layer chromatography. Ultraviolet shadowing revealed the carrier RNA digestion products (Ap, Cp, Gp, Up), including the 2′-linked GpCp dinucleotide (encircled spots). Black spots represent the overlaid autoradiograph. The small dashed circle indicates the origin. e, Multiple turnover ligation. Substrate oligonucleotides and splint (each 20 μM) were incubated with ligase #4 (1 μM) for the indicated times and the ligation product was quantified. Error bars indicate s.d. f, Thermal unfolding of ligase #6 monitored by circular dichroism spectroscopy.