Overexpressing and Purifying a Toxic Nuclease from Escherichia coli

Escherichia coli (E. coli) is a bacterium widely used to produce recombinant proteins. Numerous mutant strains, including BL21(DE3) and its derivatives, have been developed to express target proteins¹. The process is typically driven by an encoded bacteriophage T7 RNA polymerase and induced with isopropyl ß-D1-thiopalactopyranoside (IPTG), an analogue of lactose that will activate the lac operon by releasing the lac repressor^1,2,3. Because it is non-hydrolyzable, there is continuous mRNA expression and higher protein production³.

While this type of system is convenient and consistent for many proteins, some types of recombinant proteins induce toxicity in E. coli, leading to poor protein yields. Toxicity from the presence of recombinant protein negatively affecting the bacteria may be due to leaky expression before induction or stress to the bacteria upon induction^3,4. The result is a decrease in the growth rate and viability of bacteria⁵. Suppression of leaky expression can play an important role in successfully expressing toxic proteins. Expression strains of E. coli with pLysS plasmids (i.e., BL21(DE3)pLysS, C41(DE3)pLysS, etc.) suppress leaky expression by expressing a T7 lysozyme that inhibits T7 polymerase prior to induction^6,7. Other strains, such as C41(DE3) and C43(DE3) strains, contain mutations that reduce the activity of the T7 polymerase by reducing its mRNA levels, and remove the lon and ompT proteases^1,4,8,9,10. Suppressing leaky expression or reducing the rate of recombinant protein production (via T7 polymerase regulation or decreased temperatures) helps increase the yield of toxic proteins^1,9.

Nucleases are often toxic and hard-to-express proteins when using an E. coli expression system due to their enzymatic activity towards cellular DNA or RNA¹. For example, the endoribonuclease Nsp15 found in coronaviruses and other nidoviruses is toxic to the bacteria, resulting in a slower growth rate and low yields^{11,12,13,14,15,16,17}. Because of its catalytic role, cleaving 3' of uridines in viral RNA, Nsp15 is thought to act on both its own and cellular mRNA when expressed recombinantly, which results in dysregulation of metabolism and recombinant protein expression^11,14. For Nsp15, oligomerization is necessary for enzymatic activity. Previous studies of beta-coronavirus (such as SARS and SARS-CoV-2) Nsp15 demonstrated that the enzyme primarily oligomerizes into a homohexamer in solution, with the monomeric form being inactive^16,18,19,20. When mutated to only form monomers, significant increases in yield were seen¹⁴. Additionally, when mutating one of the catalytic residues within the catalytic triad to an alanine, improved expression under the same conditions as the WT was observed, further supporting that Nsp15's toxicity to E. coli is driven by its nuclease activity^13,16,17. In this protocol, we show the difference in yield between WT gamma-coronavirus (Infectious Bronchitis Virus) Nsp15 and a catalytic-dead version achieved by mutating one of the catalytic histidine residues (H223A).

Here we describe a step-by-step methodology for overexpressing recombinant Nsp15 in a C41(DE3) expression system, followed by protein purification through affinity and size exclusion chromatography. An N-terminal 6x His-tag with cobalt-based immobilized metal affinity chromatography (IMAC) is employed to isolate the protein. Further purification is performed via SEC to separate Nsp15 by oligomeric state and isolate the active hexameric species. This approach offers a method to overexpress and purify toxic nucleases for biochemical and structural studies.

1. Preparation for buffer and reagents

1. Overexpression reagents
   1. Prepare 1 L of 2xTY media by combining 16 g of tryptone, 10 g of yeast extract, 5 g of sodium chloride (NaCl), and 900 mL of deionized water in a 2 L baffled flask. Autoclave the flask for at least 30 min on a liquid cycle.
      NOTE: At least 2L of 2xTY media is needed per purification.
   2. Prepare appropriate antibiotic stock/s for the plasmid. For Nsp15-pET-14b, make 15 mL of 100 mg/mL ampicillin by dissolving 1.5 g of ampicillin with deionized water in a 15 mL conical tube. Store the antibiotic stocks at -20 ˚C until needed.
   3. Prepare 250 mL of Lysogeny Broth (LB) agar media by combining 2.5 g of tryptone, 1.25 g of yeast extract, 2.5 g of NaCl, 3.75 g of agarose, and 250 mL of deionized water in a 500 mL glass bottle. Sterilize the solution by autoclaving for at least 30 min on a liquid cycle and cool down until it is warm to the touch.
   4. Prepare LB agar plates supplemented with ampicillin by transferring 25 mL of the warm LB agar media into a sterilized 250 mL Erlenmeyer flask (25 mL per plate to be made). Add 25 µL of 100 mg/mL ampicillin for every 25 mL of media and swirl to combine the contents. Using a serological pipette, add 25 mL of media mix to a sterile plastic petri dish; avoid introducing any bubbles.
      1. Prevent contamination by wiping the lab bench with 20% ethanol and lighting a Bunsen burner to maintain sterility while pouring the plates.
      2. Allow the plates to cool before storing them at 4 ˚C until further use.
   5. Prepare 15 mL of 1 M IPTG by dissolving 3.574 g of IPTG with deionized water in a 15 mL conical tube. Store at -20 ˚C until needed. Avoid repeated freeze/thaw cycles by aliquoting into three 5 mL aliquots.
2. Purification reagents
   NOTE: All reagents should be prepared at least one day prior to purification
   1. Prepare 250 mL of lysis buffer (50 mM Tris pH 8.0, 150 mM NaCl, 5% glycerol, 5 mM imidazole) and store at 4 ˚C.
   2. Prepare 15 mL of elution buffer (50 mM Tris pH 8.0, 150 mM NaCl, 5% glycerol, 250 mM imidazole) and store at 4 ˚C until needed.
   3. Prepare 250 mL of cleavage buffer (50 mM HEPES pH 7.5, 150 mM NaCl, 5% glycerol) and store at 4 ˚C until needed.
   4. Prepare 1 L of SEC buffer (20 mM HEPES pH 7.5, 150 mM NaCl, 5 mM MnCl₂) and filter through a 0.22 µm PES filter using vacuum filtration. Store at 4 ˚C with FPLC.
   5. Prepare 15 mL of 1 M 4-(2-Aminoethyl)benzenesulfonyl fluoride hydrochloride (AEBSF) by dissolving 3.595 g of AEBSF with deionized water in a 15 mL conical tube. Make 1 mL aliquots in 1.5 mL microcentrifuge tubes and store at -20 ˚C until needed.
   6. Optional: Prepare Sodium Dodecyl Sulfate (SDS) gel-loading buffer master mix by diluting 200 µL of stock solution (250 mM Tris pH 6.8, 8% SDS, 0.1% bromophenol blue, 40% glycerol, 100 mM DTT) in 400 µL of deionized water.
   7. Prepare gel sample tubes to track purification. Transfer 30 µL of 1x loading buffer into each tube.
   8. Prepare 1 L of 10x Tris-Glycine running buffer by dissolving 30 g of Tris base, 144 g of glycine, and 20 g of SDS, in 1000 mL of deionized water. For a 1x solution, dilute 100 mL of the 10x solution in 900 mL of deionized water in a 1 L bottle. Store both reagents at room temperature until further use.

2. Nsp15 overexpression

NOTE: All reagents and materials should be kept as sterile as possible to prevent the risk of contamination.

1. Transformation
   1. Remove C41(DE3) E. coli cells from -80 ˚C to thaw for 10 min on ice. Pipette 1 µL of plasmid DNA into a tube containing the C41 cells. Let it incubate by sitting on ice for 30 min.
      NOTE: While the concentrations of plasmids may vary, for the data described here, 10-20 ng of DNA was added per transformation. The commercial protocol for C41(DE3) cells recommends 10-50 ng of DNA per transformation. This protocol is applicable to any plasmid using the T7 promoter system; however, here a previously published pET14b plasmid encoding codon-optimized Nsp15, with a TEV site following the N-terminal 6xHis-tag and thrombin cleavage site, was modified by replacing SARS-CoV-2 Nsp15 with the IBV Nsp15 sequence (GENBANK: WIL95322.1, 5989-6324)¹⁹.
   2. Place the tube in a heat block at 42 ˚C for 45 s to heat shock the cells to uptake the plasmids. Incubate the tube on ice for 2 min.
   3. Pipette 200 µL of recovery media (SOC or other rich media) into the tube containing the C41(DE3) cells and plasmids. Shake at 210 rpm for 1 h at 37 ˚C.
   4. Remove LB agar plates supplemented with ampicillin from 4 ˚C and place into a non-shaking incubator set to 37 ˚C.
   5. Pipette bacteria on the agar plate. Using a sterile spreader, spread the cells onto the agar plate. Place the plate in a non-shaking incubator set to 37 ˚C to grow for 14-16 h.

3. Growth

1. Single colony picking
   NOTE: Same-day single colony starters were used rather than an overnight starter culture in order to inoculate cells in the logarithmic growth phase. For overnight cultures, cells are typically in the stationary or death phase, especially with toxic proteins. Additionally, given longer culture times, contaminants or mutants can outcompete the toxic nuclease.
   1. Use one 50 mL sterilized Erlenmeyer flask for each L of culture plus 2-3 additional starters. Label one of the flasks with a star for optical density (OD) checks. The OD collected from this flask will represent the entire growth unless there is a visual difference between flasks.
      NOTE: Erlenmeyer flasks were found to provide better aeration than 15 mL starter tubes.
   2. Create a master mix of 2xTY plus appropriate antibiotics using a volume of 10 mL x N number of starters. For a 4 L growth, make 6 starter cultures. The master mix would be 60 µL of ampicillin stock (thawed from -20 ˚C) and 60 mL of 2xTY (1:1000 dilution). Swirl to mix and aliquot 10 mL into each Erlenmeyer.
   3. Remove the transformed plate from the incubator. Pluck a single, isolated colony from the plate using a sterile toothpick or a pipette tip. Transfer the toothpick with the colony into a flask with media. Repeat until each flask contains a toothpick.
   4. Place the inoculated starter flasks in the shaking incubator at 210 rpm, 37 ˚C. Allow the starter flasks to grow for ~5-7 h.
2. Culture inoculation
   1. After 5 h, check the OD at 600 nm (OD₆₀₀) of the flask marked with a star using a microcuvette and spectrometer. Remove and transfer 1 mL of the culture from the flask into a microcuvette. Repeat measurements periodically until the OD₆₀₀ reaches 0.8-1.0.
   2. 30 min to 1 h prior to inoculation (when OD₆₀₀ reaches 0.4), warm the 2 L 2xTY flasks at 37 ˚C. Warm media reduces lag time following inoculation with starter culture.
   3. Thaw a 100 mg/mL ampicillin stock from -20 ˚C. Once the target OD₆₀₀ is reached, prepare four 2 L flasks with 1 L of 2xTY media by adding 1 mL of 100 mg/mL of ampicillin (1:1000 dilution).
   4. To inoculate, pour the contents of one starter flask into one 2 L flask. Repeat until each 2 L flask is inoculated. Avoid pouring the toothpick into the flask for an easier time harvesting cells.
   5. Place the inoculated 2 L flasks into the shaking incubator at 210 rpm, 37 ˚C for 3 h before checking the first OD₆₀₀.
3. Induction with IPTG
   1. Check the OD₆₀₀ by removing 1 mL of media from a flask with a serological pipette. Repeat until the OD₆₀₀ reaches 0.8-1.0.
   2. Once the target OD₆₀₀ is reached, add 1 mL of 1 M IPTG for a final concentration of 1 mM (thawed from -20 ˚C) to induce overexpression.
   3. Transfer the 2 L flasks to a shaking incubator at 210 rpm, 16 ˚C for overnight induction (14-16 h).
      NOTE: The colder temperature will slow down the growth of the E. coli cells, reducing toxicity by slowing protein production and nuclease activity.
4. Harvest
   1. Remove the induced flasks from the shaking incubator and pour the contents into separate 1 L centrifuge bottles. Balance the bottles using deionized water. Spin down the cells at 4500 x g for 10 min at 4 ˚C.
   2. Remove the supernatant and transfer the pellet into a 50 mL conical tube using a spatula. Combine pellets into a single conical tube. Store at -80 ˚C until further use.

4. Nsp15 purification

NOTE: All buffers and tubes should be kept on ice. All centrifugation steps should be carried out at 4 ˚C. The following protocol assumes a His-tagged protein.

OPTIONAL: After each step, gel samples can be taken to follow the purification progress.

1. Cell lysis and lysate clarification
   1. Resuspend the bacterial pellet by adding 2 mL of lysis buffer for every 1.2 g of pellet. Add 100 µL of 1 M AEBSF for every 10 mL of lysis buffer as a protease inhibitor. If using a combined pellet from 4 L, add 100 µL of DNase I for a final concentration of 200 U.
      NOTE: One EDTA-free protease inhibitor cocktail tablet per L can be substituted as a protease inhibitor.
   2. Briefly vortex the contents of the tube for 30 s to allow the pellet and buffer to mix. Transfer the vortexed pellet into a Dounce tissue grinder. With a loose pestle, homogenize the contents (~10 strokes).
   3. Transfer the homogenized sample into a metal beaker for sonication.
      1. To ensure the maximum sample is obtained, rinse the tube containing the original pellet with 5 mL of lysis buffer and vortex the tube briefly. Pour the contents into the Dounce homogenizer, apply ~5 strokes, and then transfer to the sonication beaker.
   4. To the metal beaker, add 75 µL of Triton X-100 to assist in membrane lysis and solubilization. Place the metal beaker into an ice bath to ensure the sample stays cold during sonication.
   5. Place the sonication probe into the metal beaker and sonicate the sample for 6 min and 30 s with pulses every 2 s and amplification at 70%.
      CAUTION: Use hearing protection if the sonicator probe is not enclosed.
   6. Using a serological pipette, transfer the lysate into a centrifuge tube. To the same tube, add 1 M AEBSF at a 1:100 dilution as a fresh protease inhibitor. Clarify the lysate via centrifugation at 26,915 x g for 50 min.
2. Affinity chromatography by gravity filtration
   NOTE: Cobalt resin significantly decreases non-specific metal binders compared to nickel resin²¹.
   1. While sonicating, obtain 2 mL of cobalt resin slurry and resuspend with 25 mL of lysis buffer in a new conical tube. Centrifuge the resuspended resin for 5 min at 500 x g. After centrifugation, pour off the supernatant, being careful not to lose any resin.
   2. Repeat with two separate 25 mL lysis buffer washes followed by centrifugation. Do not pour off the supernatant of the last wash until ready to batch bind. When ready, pour out the lysis buffer from the last resin wash and resuspend the resin with 10 mL of fresh lysis buffer (10 mL of buffer per 1 mL of resin).
   3. Add the soluble fraction to the tube and allow the 6x His-tagged protein to batch bind to the resin on a rocker for 45 min at 4 ˚C.
      NOTE: If the batch bind volume exceeds 50 mL, perform this step in a 250 mL beaker with a stir bar at a very slow speed at 4 ˚C. Transfer all the resin into the beaker when performing this step using lysis buffer.
   4. After batch-binding, pour the contents down a gravity filtration column and collect the flow through in a 125 mL Erlenmeyer flask. The contents within the flask should be any protein that did not bind to the resin; His-tagged Nsp15 will be bound to the resin. Using a separate flask for collection, wash the resin with 8 x 25 mL of buffer for a total of at least 200 mL (~5x the original volume) using a serological pipette.
      NOTE: Slowly add the buffer down the column's side rather than directly down the middle to prevent disturbing the resin bed, which can lead to inefficient or uneven washing.
3. Protein elution
   1. Elute the His-tagged Nsp15 in three elutions with elution buffer. Elutions 1 and 2 are 2 mL elutions, and elution 3 is a 1 mL elution.
   2. Perform a "quick and dirty" Bradford assay by making a 1:1 dilution of Bradford reagent in water. Transfer 10 µL of each elution into tubes and invert to mix. The solution should turn blue in the presence of protein. Assess the color change to determine how to combine elutions.
      NOTE: If elutions 1 and 2 are strong and elution 3 is weak, combine elutions 1 and 2 and proceed. If all are weak, combine all three tubes.
4. Protein concentration and tag cleavage
   1. Transfer the combined elution sample into a 30 kD MWCO concentrator (or appropriate size for protein of interest). Concentrate the protein to ≤2 mL by centrifuging at 3,000 x g for 10 min. Transfer the concentrated sample into a 2 mL microcentrifuge tube. Bring up the final sample volume to 2 mL with cleavage buffer if necessary.
   2. Prepare a disposable desalting column (e.g., PD-10 column) according to the manufacturer's directions. Add the 2 mL sample to the column. Place the column into a clean 50 mL conical tube and elute by centrifugation at 1,000 x g for 2 min.
      NOTE: This step quickly removes the imidazole, which interferes with thrombin cleavage while avoiding concentration-induced aggregation when using MWCO concentrators to buffer exchange.
   3. To the eluate, add β-Me and CaCl₂, to a final concentration of 2 mM each. Add 50 µL of a 1 U/µL thrombin stock. After all components are added, place the tube on a shaker upright and allow the reaction to incubate at room temperature for 4 h with gentle shaking.
      NOTE: Overnight cleavage at 4 ˚C may work for some proteins; improved yields of Nsp15 were observed with a 4 h cleavage at RT compared to an overnight 4 ˚C cleavage.
   4. Re-equilibrate the gravity filtration column in cleavage buffer (wash 3x with 25 mL cleavage buffer). Ensure the stop cock is closed and add 5 mL of cleavage buffer to prevent the resin from drying out. Store the column at 4 ˚C during the cleavage period.
5. Separating cleaved protein from cleaved His-tag and uncleaved protein
   1. Using a serological pipette, carefully pipette the cleavage reaction directly onto the resin and collect the eluate into a 15 mL conical tube.
   2. Wash the resin with 2 mL of cleavage buffer and collect in the same 15 mL tube. Repeat for a total of two washes. When finished, add AEBSF to the tube to quench the thrombin to obtain a final concentration of 10 mM.
   3. Transfer the repass sample into a new 30 kD MWCO concentrator. Concentrate the protein down to ≤500 µL by centrifugation at 3,000 x g for 10 min. Transfer the concentrated sample into a 0.5 mL microcentrifuge tube.
6. Size exclusion chromatography
   1. Run an S200 increase analytical column (or other appropriate sizing column for the protein of interest) according to the manufacturer's specifications. Equilibrate in SEC buffer and set the FPLC fraction size to 500 µL.
   2. Transfer selected fraction tubes from the peaks into new 1.5 mL microcentrifuge tubes and add 2.5 µL of 1 M DTT (final concentration 5 mM) to each fraction to keep the purified Nsp15 reduced.
   3. Measure the A₂₈₀ values of each fraction using a nano-spectrophotometer. Blank the instrument with SEC buffer prior to collecting measurements. Store the purified Nsp15 samples at 4 ˚C until further use.
7. Optional: Purification check
   1. Heat the gel sample tubes on a 95 ˚C heat block for 10 min. Briefly centrifuge the tubes after heating to spin down any condensation. Load the samples onto 4-20% Stain-free Mini-Protean TGX gel/s and run at 300 V for 17 min. Image the gel/s to check purity.

Due to Nsp15's toxicity as a nuclease, a total of ~7 h was usually required for the starter cultures to reach an OD600 of 0.8-1.0. Once inoculated, a total of ~3-5 h was typically required for the same optimal OD600 of 0.8-1.0 for induction. In this case, the doubling time appeared to be ~1 h. When a catalytic-dead mutant of Nsp15 was expressed, the doubling for the E. coli cells appeared normal (~20 min), further evidence that the nuclease activity is responsible for toxicity (Figure 1A).

Following overnight induction, 6-9 g of wet weight cell pellet per 2 L WT growth was typically harvested. Across many different growths with various Nsp15 mutants, we observed that smaller pellets correlate with higher nuclease activity. For the catalytic-dead Nsp15, a much larger pellet, ~18-30 g, was obtained per 2 L growth. Together, this further supports that the nuclease activity leads to toxicity during expression, though we lack direct evidence in the form of a western blot due to resource limitations.

Representative gels of samples taken throughout the procedure showed successful overexpression and purification of WT Nsp15 and catalytic-dead Nsp15 from a gamma-coronavirus, Infectious Bronchitis Virus (Figure 1 to Figure 3). In Figure 1B, there is a clear overexpressed post-induction band at the correct molecular weight for catalytic-dead Nsp15, but not WT Nsp15 (Figure 1B, lanes 3 and 5). However, during purification, the WT band became visible during affinity chromatography (Figure 2, lane 6). A similar pattern was observed with turkey coronavirus Nsp15, where no clear post-induction band was present, but a distinct band appeared after affinity chromatography22. Utilizing affinity chromatography and size exclusion chromatography allowed for the isolation of the active enzyme, first by the 6x His-tag and second by oligomeric state. The resin used for isolation was a cobalt-charged IMAC resin with higher specificity for 6x His-tags than nickel-charged resins. Evidence of this is shown in Figure 1B, lane 5, by the appearance of a predominant band at ~42 kDa (His-Nsp15) and a lack of higher MW bands (ArnA, a common nickel resin binder, has a MW of 75 kDa)23. The first purification gel demonstrates elution of His-Nsp15 by the appearance of the single isolated band at the same molecular weight described earlier (Figure 2A and Figure 3A, lane 7). Most of the protein eluted off with the first imidazole wash, with residual amounts in the second and third washes and trace amounts remaining bound to resin (Figure 2A and Figure 3A, lanes 7-10).

Thrombin cleavage of the His-tag occurs at RT for 4 h. Successful cleavage was observed in the second purification gel, with a ~2 kDa decrease in molecular weight between the pre-cleavage and post-cleavage lanes (Figure 2B and Figure 3B, lanes 2-3). Repassing the cleavage reaction through the resin cleans up the reaction, removing the His-tag, residual uncleaved Nsp15, and non-specific metal binding proteins, with very little cleaved Nsp15 remaining on the resin itself (Figure 2B and Figure 3B, lanes 4-5).

Since oligomeric Nsp15 was in the active state, SEC was used to separate and identify the primary oligomeric states. In the representative chromatogram obtained from a purification, two peaks are observed, with the peak at ~11 mL of elution volume corresponding to the active hexamer and the peak at ~15 mL corresponding to the inactive monomer (Figure 4). The hexamer and monomer peak fractions show purity >95% by SDS-PAGE (Figure 2B and Figure 3B, lanes 6-10). Fractions 1-10 and 14-18 displayed on the trace were not included in SDS-PAGE as no detectable protein was observed in those fractions. Using a dual-labeled (5'-fluoroscein, 3'-Cy5) 51-mer RNA, we then showed the hexameric WT Nsp15 was active, and the monomeric WT Nsp15 and both states of catalytic-dead Nsp15 were inactive (Figure 5).

A nano-spectrophotometer was used to measure the absorbance at 280 nm (A280) of Nsp15 fractions. The A280 values obtained were converted into concentrations in units of µM using Beer's Law and the appropriate extinction coefficient (using Expasy ProtParam; Table 1). Again, when comparing WT and catalytic-dead Nsp15, there was a significant difference in yield (Table 1). Overall, this expression and purification method yields highly pure, active Nsp15 at sufficient concentrations for biochemical and cryo-EM structural studies.

Figure 1: Representative data of a typical H223A catalytic-dead and WT Nsp15 overexpression in C41(DE3) cells. (A) Growth curve of H223A catalytic-dead (triangles) and WT Nsp15 (circles) in OD600 vs. time (h). OD600 measurements taken from the 10 mL starter cultures are represented by SC. OD600 measurements taken from 2 L flasks post-inoculation are represented by PI. (B) SDS-PAGE gel (4-20% TGX) of pre- and post-induction samples of WT and H223A catalytic-dead Nsp15 overexpression. Pre-induction samples were taken prior to the addition of 1 M IPTG to the flasks. Post-induction samples were taken following overnight induction at 16 ˚C. Please click here to view a larger version of this figure.

Figure 2: Representative SDS-PAGE gels (4-20% TGX) of a typical WT Nsp15 purification, overexpressed in C41(DE3) cells. (A) Gel with samples collected from cell lysis to Nsp15 elution from cobalt resin. Lanes 6 and 10 represent samples taken directly from the resin bed after the resin washes in step 4.2.4 and after Nsp15 elution after step 4.3.1, respectively. (B) Gel with samples collected from pre-cleavage through SEC. Pre-cleavage samples were taken prior to the addition of thrombin, after desalting the sample in step 4.4.2. Post-cleavage samples were taken following the 4 h tag cleavage period before performing step 4.5.1. A repass sample was acquired from the eluate collected after passing the cleavage sample over the cobalt resin to capture cleaved His-tag and uncleaved protein after step 4.5.1. Please click here to view a larger version of this figure.

Figure 3: Representative SDS-PAGE gels (4-20% TGX) of a typical H223A catalytic-dead Nsp15 purification, overexpressed in C41(DE3) cells. (A) Gel with samples collected from cell lysis to H223A Nsp15 elution from cobalt resin. Lanes 6 and 10 represent samples taken directly from the resin bed after the resin washes and after H223A Nsp15 elution, respectively. (B) Gel with samples collected from pre-cleavage through SEC. Pre-cleavage samples were taken prior to the addition of thrombin. Post-cleavage samples were taken following the 4 h tag cleavage period. A repass sample was acquired from the eluate collected after passing the cleavage sample over the cobalt resin to capture cleaved His-tag and uncleaved protein. Please click here to view a larger version of this figure.

Figure 4: Representative SEC traces of recombinant Nsp15 overexpressed in C41(DE3) cells. Nsp15 (A) WT and (B) H223A catalytic-dead variants were resolved over a gel filtration column using SEC buffer. Please click here to view a larger version of this figure.

Figure 5: Cleavage gel of a nuclease assay performed with a select RNA substrate. The sequence is shown below the gels with labels colored to match the overlays. IBV Nsp15 (100 nM) was incubated with 5'-FI-RNA-Cy5-3' (500 nM) for 60 min at room temperature. Specific samples were taken at labeled times (min) and quenched with loading buffer. The images for the 5'-FI (blue) and 3'-Cy5 (red) products were overlaid. The purple band is representative of uncleaved RNA. An alkaline hydrolysis of the RNA substrate was performed to generate a ladder. RNA only controls at 0 and 60 min were used to ensure no degradation of the substrate over the time course. (A) Hexameric WT and H223A catalytic-dead Nsp15 time course. (B) Monomeric WT and H223A catalytic-dead Nsp15 time course. Please click here to view a larger version of this figure.

Table 1: Data from three representative WT (4 L) and catalytic-dead (2 L) Nsp15 purifications. Yields were calculated for both total (hexameric + monomeric) and hexameric Nsp15 in mg of purified Nsp15 per L of cell growth and per g of wet weight cell pellet. The averages were calculated for each unit of measure. For less culture volume, significantly more catalytic-dead Nsp15 is purified, which highlights the toxic effects of Nsp15 nuclease activity.

We have presented a detailed protocol that results in the successful overexpression in E. coli of a toxic nuclease and purification using affinity and size exclusion chromatography. Implementation of same-day 10 mL single colony starter cultures instead of overnight starter cultures increased consistency in growth times and protein yields across growths. We chose to use C41(DE3) cells, which contain mutations that decrease T7 polymerase activity, to help manage recombinant protein toxicity^1,4,8,9,10. When optimizing or adapting this protocol, it may be worthwhile to test expression of the protein of interest in Rosetta2, C43(DE3), or pLysS strains. While preparing this article, we discovered the commercial source of the C41(DE3) cells we used in this protocol had been discontinued. We found two additional companies that sell C41(DE3) equivalent cells, but we have not tested them yet. Another cell type, not explored in this method, is BL21-AI cells, which control T7 polymerase expression using the pBAD promoter, which is a tight regulation of expression²⁴. This cell line was used to express and purify the arteriviral version of Nsp15, Nsp11, although the authors note the WT protein still had a very low yield¹².

Nsp15 is conserved across vertebrate nidoviruses, though most studies have focused on the coronavirus family. When comparing the methods sections of these papers, the purifications are largely similar, with the main differences being resin type (nickel vs. cobalt), whether the affinity tag is cleaved, and whether a sizing column is used^{11,12,13,14,15,16,17,20,21,24,25,26,27,28,30}. If the cleavage tag is small (6x-His, FLAG), cleavage is likely not critical. However, if the tag is large (GST, MBP, SUMO), leaving it intact may influence the association and function of Nsp15. Inclusion of an SEC column is especially important for separating the active and inactive forms of the protein. The bacterial overexpression methods are the most different; ours is the only method that uses the same-day starter cultures. Cell types also vary, though cells with increased T7 regulation are used in some instances, such as C41(DE3) or Rosetta(DE3) pLysS^{11,26,27,32,33}. Many of these papers do not report yields. When calculated, reported yields vary dramatically between Nsp15 constructs from different viruses; likely, both the plasmid backbone and source virus affect yields too. It is often not explicitly stated if total soluble Nsp15 or active, hexameric Nsp15 yields are being reported, further limiting direct comparison^{11,12,13,14,15,16,17,20,21,24,25,26,27,28,30}. There are three papers to our knowledge describing the purification of gamma-coronavirus Nsp15. Bhardwaj et al. induced for 36 h at 16 ˚C and used nickel resin and SEC to purify IBV Nsp15, but did not report the purification yield³⁴. Cao et al. used BL21(DE3) cells and a 3 h, 37 ˚C induction to express the turkey coronavirus Nsp15²². They purified it using nickel resin, but they did not use SEC to separate the hexameric and monomeric forms of the protein, and did not report the purification yield. Zhao et al. used BL21(DE3) cells to express the protein, but did not mention the induction conditions¹³. They used a GST-tagged construct and cleaved it during purification, but did not use SEC and did not report the purification yield.

Common molecular biology techniques were optimized for the overexpression and purification of IBV Nsp15, and this protocol can further be adapted to purify other endoribonucleases. The buffers used within the purification included Tris and HEPES at pH 8.0 and 7.5, respectively. The isoelectric point (pI) of the protein of interest must be considered to ensure maximal solubility and can affect the type of Good's buffer used. Depending on the affinity tag and protein sequence, the pI may also change significantly post-cleavage and necessitate a buffer pH change. For IBV Nsp15, lowering the salt concentration in the purification buffers from 500 mM NaCl, used in earlier versions of this method, to 150 mM NaCl increased yields¹⁹. Thus, the salt concentration is another point of optimization for applying this protocol to other proteins. Since this protocol relies on E. coli expression, it would not be suitable for studying nucleases where post-translational modifications are important for function. A similar limitation exists for nucleases that rely on disulfide bonds for proper folding and function, as the E. coli environment is highly reductive, though there are ways to address this³⁵. Additionally, if the protein is extremely toxic, it may be necessary to move to a cell-free expression system.

Our Nsp15 construct contained a 6x-His-tag and was easily purified using IMAC with cobalt resin. The bed volume of the resin (1 mL) was optimized for the typical protein yield (<10 mg). While impurities were present following elution, both pre- and post-cleavage, the subsequent SEC purification resulted in a highly purified protein (>95%). The number of resin washes was optimized for Nsp15, but additional washes may result in higher purity prior to SEC. The elution steps have also been optimized for Nsp15, with trace amounts of protein remaining on the resin. For SEC, the type of gel filtration column used is highly dependent on the overall size and oligomerization of the protein. Here, an S200 column was utilized as Nsp15 primarily oligomerizes into a homohexamer with MW of ~240 kDa. For smaller proteins, an S75 column may be more relevant to the procedure. Finally, following SEC, final protein concentrations can often be increased by concentrating or pooling fractions obtained.

Expressing and purifying nucleases for biochemical assays and structural studies can be challenging due to their toxicity to E. coli, a result of their enzymatic activity against DNA or RNA. In this method, we have consistently improved yields of active Nsp15 protein by using 1) C41(DE3) cells, 2) same-day single colony starter cultures, 3) cobalt IMAC resin, and 4) SEC to isolate the active hexamer. Parts of this method may be applied to achieve successful overexpression and purification of other nucleases or oligomeric enzymes.