Protein switches identified from diverse insertion libraries created using S1 nuclease digestion of supercoiled-form plasmid DNA

Abstract

We demonstrate that S1 nuclease converts supercoiled plasmid DNA to unit-length, linear dsDNA through the creation of a single, double-stranded break in a plasmid molecule. These double-stranded breaks occur not only in the origin of replication near inverted repeats but also at a wide variety of locations throughout the plasmid. S1 nuclease exhibits this activity under conditions typically employed for the nuclease's single-stranded nuclease activity. Thus, S1 nuclease digestion of plasmid DNA, unlike analogous digestion with DNaseI, effectively halts after the first double-stranded break. This property makes easier the construction of large domain insertion libraries in which the goal is to insert linear DNA at a variety of locations throughout a plasmid. We used this property to create a library in which a circularly permuted TEM1 β-lactamase gene was inserted throughout a plasmid containing the gene encoding Escherichia coli ribose binding protein. Gene fusions that encode allosteric switch proteins in which ribose modulates β-lactamase catalytic activity were isolated from this library using a combination of a genetic selection and a screen.

Figure 1

S1 nuclease converts supercoiled plasmid DNA to nicked circular and linear forms through digestion at a variety of locations in the plasmid. (A) Two μg of supercoiled pRH04.152-rbsB was incubated at 22°C (top) or 37°C (bottom) with different amounts of S1 nuclease for different lengths of time and then analyzed by agarose gel electrophoresis. The sixth and seventh lanes contain undigested supercoiled control and SpeI-digested linear control, respectively. The band that runs slower than the linear band was presumed to correspond to nicked circular plasmid. In the thirteenth lane, after the DNA was digested with 10 units for 24 hours an additional 10 units of S1 nuclease was added and the DNA was incubated for an additional hour. The first and last lanes contain the λDNA/HindIII molecular weight standards. (B) Plasmids with or without the f1 origin (pDIMC8-pfMBP and pRH04-pfMBP, respectively) were incubated with or without S1 nuclease, followed by incubation with or without NcoI, EagI, and SacII and then analyzed by agarose gel electrophoresis. (C) Plasmids with or without the f1 origin (pDIMC8-rbsb and pRH04-rbsb, respectively) were incubated with S1 nuclease or DNaseI, followed by incubation with XmnI and analyzed by agarose gel electrophoresis.

Figure 2

Schematic description of the library creation method. Step 1: S1 nuclease digests supercoiled plasmid DNA. Step 2: DNA with a single, double-stranded break is isolated by agarose gel electrophoresis. Step 3: Incubation with T4 polymerase and T4 ligase repairs any single strand overhands and seals any nicks. Step 4: Dephosphorylation prevents recyclization during the subsequent ligation step. Step 5: An insert DNA (created by PCR using phosphorylated primers) is incorporated using T4 ligase. Step 6: The library DNA is transformed into E. coli cells. Step 7: The library is selected for ampicillin resistance and the selected clones screened for switching ability.

Figure 3

S1 nuclease digestion of supercoiled DNA results in a diverse set of products. a) To scale linear diagram of plasmid pDIMC8-rbsB with rbsB gene (including the rbsB signal sequence), tac promoter (tacP/O), p15A origin, f1 origin, and chloramphenicol resistance gene (Cm^R) labeled. Arrows indicate the location of S1 nuclease digestion sites identified from sequencing of 30 naïve library members. Triangles indicate the location of S1 nuclease digestion sites in the switches discovered in the library, with four distinct switches discovered at the location marked by the *. b) Distribution of deletion lengths at the site of S1 nuclease-created double stranded breaks in the same 30 naïve library members.