You are currently viewing [Paper Review #34] Facile, high-throughput protein purification enabled by high-efficiency protease elution

[Paper Review #34] Facile, high-throughput protein purification enabled by high-efficiency protease elution

  • Post category:Knowledge
  • Reading time:8 mins read
Paper headline: Facile, high-throughput protein purification enabled by high-efficiency protease elution

https://onlinelibrary.wiley.com/doi/10.1002/pro.70746

Facile, high-throughput protein purification enabled by high-efficiency protease elution

Key figures

  • Figure 1 (p. 3). On Ni–NTA beads loaded with 10×His–SUMO–sfGFP, a directly His-tagged SUMO protease released about 10% of the reporter by 30 minutes and 25% by 100 minutes. Inserting the disordered tardigrade protein CAHS_107838 between the His tag and protease restored near-complete release by 30 minutes. This triplicate time course used 150 nM enzyme at room temperature and 500 rpm, with fluorescence measured at 488/510 nm. The dashed reference is fluorescence from intact fusion eluted with 400 mM imidazole.
  • Figure 2 (p. 5). The three-day workflow starts with linear DNA. Golden Gate assembly and direct transformation feed overnight starter cultures, followed by overnight autoinduction, parallel sonication and magnetic-bead purification. Four 24-well plates supply one 96-well collection plate. The supplement maps the transfers for both a manual multichannel pipette and a liquid handler (Figs. S6, S7 and S11).
  • Figure 3a–d (p. 6). Pooled nanopore sequencing of the sfGFP assembly in E. coli DH10B yielded 1,434 mapped reads. The authors report >95% correct insertion and no detectable empty vector. A separate inoculum titration showed that approximately 188 colony-forming units (CFUs) could seed a 1 mL starter. The authors selected conditions yielding about 900 CFUs for production in Lemo21(DE3). Sequencing one reporter pool leaves the sequence composition of individual production wells unresolved (§1.4, pp. 5–6).
  • Figure 4a,b (p. 7). In triplicate 50 mL flask cultures, TBM-5052 produced roughly three times the sfGFP fluorescence of ZYM-5052 after 16 hours. In deep wells, 0.5 mL gave the highest yield per milliliter, whereas 2 mL gave the most protein per well. Across 96 independent runs with the same reporter, whole-culture fluorescence had a coefficient of variation below 5% (Fig. S3).
  • Figure 5a–d (p. 8). Sonication recovered more soluble sfGFP than 0.5–2% nonyl glucoside. Without active cooling, extraction plateaued after about four minutes, by which point the lysate exceeded 50 °C. An ice-water bath allowed efficient extraction with shorter pulses. The capture experiment then tracked disappearance of reporter fluorescence from clarified lysate as bead volume and binding time varied (Figs. S4–S5).
  • Figure 6a–f (p. 10). The benchmark comprised 95 disulfide-free, single-domain proteins from mesophiles, each with a known structure and prior expression in E. coli, plus sfGFP. The mean pairwise TM-align score was 0.28, and 64 CATH superfamilies were represented among 79 assigned proteins. Yield showed no obvious trend with molecular weight. Panels d–f separately assess thermal unfolding, dispersity and secondary structure (§1.9, pp. 9–10, Table S3).

1) Thesis (one sentence)

To overcome inefficient cleavage or protease carryover during on-bead tag removal, Cárcamo Noriega and colleagues insert a disordered spacer between a His tag and SUMO protease, restoring rapid sfGFP release while retaining the protease on nickel and supporting a three-day, 96-sample workflow evaluated by purification yield, gels and biophysical assays (Figs. 1–6, §§1.2,1.9).

2) Evidence card (three bullets only)

  • Strongest result: The authors ran the panel in biological triplicate and report that 87 of 96 proteins yielded >100 µg from a 2 mL culture at >90% purity by SDS–PAGE. This is the authors’ reported yield-and-purity success count; the volume/concentration discrepancy is noted below (§1.9, p. 9, §3.6, p. 12, Fig. S8a,b).
  • Method enabler: AIProtease combines a 10×His anchor, the 26.4 kDa CAHS_107838 spacer and the Chaetomium thermophilum SUMO protease. The untagged enzyme also released sfGFP rapidly but appeared in the eluate. In a separate carryover test using 0.1 mg/mL of each protease for 30 minutes, the AIProtease eluate had a strong sfGFP band and no visible protease band. Together, the time course and gel support improved access to bead-bound substrate while retaining the enzyme on nickel (§§1.1–1.2, pp. 2–4, Table S4 and Fig. S1).
  • Critical limitation: The benchmark is restricted to single-domain, disulfide-free mesophilic proteins with prior E. coli expression. Among the 87 reported production successes, 68 gave a cooperative thermal transition by differential scanning fluorimetry and 64 gave circular dichroism spectra consistent with the known secondary structure. Thirteen samples were too dilute for reliable CD at ≤0.1 mg/mL. Size-exclusion chromatography gave 53 single symmetric peaks and 12 broad peaks or altered retention times. These assays support folding for subsets of the panel. A single SEC peak can represent an oligomer, and the study includes no panel-wide activity assay (Fig. 6d–f and §1.9, p. 10).

Optional

Key comparisons (1–3 lines)

Imidazole releases the His-tagged fusion, leaving tag removal for a later step. AIProtease releases the cleaved target directly into PBS. Figure 1 compares sfGFP recovery across elution methods, but the paper provides no matched imidazole workflow for all 96 proteins (Fig. 1, p. 3, §3.6, p. 12).

Methods I might copy (protocol hooks)

Assemble and express. Combine 150 ng pAIP4 with insert at a 1:3 vector-to-insert molar ratio in 20 µL. Incubate at 37 °C for 60 minutes, then 60 °C for 10 minutes. The assembly screen used electrocompetent TOP10. For production, mix 5 µL assembly with 50 µL chemically competent Lemo21(DE3), chill for 30 minutes, heat-shock at 42 °C for 30 seconds and select in broth after recovery. Inoculate 2 mL TBM-5052 containing 1 mM L-rhamnose with 20 µL saturated starter. Use square 24-deep-well plates at 25 °C for 16–24 hours, 800 rpm on a 3 mm orbit (§§3.2–3.5, pp. 11–12, Table S2).

Check the product termini. The pAIP4 diagrams encode Ser–Gly–target–Gly–Ser, although §3.1 calls the C-terminal pair Ser–Gly. Figure S9 shows higher eluate concentrations with the N-terminal Ser–Gly for ten cleavage-sensitive targets tested in triplicate. For proteins requiring native termini, the authors offer pAIP53 (Addgene 258140). The scar comparison makes junction-specific cleavage worth testing with that vector (Fig. 3b, p. 6, §3.1, p. 11, Figs. S9–S10).

Lyse and capture. Resuspend each pellet in 2 mL lysis buffer containing 2 mM MgCl₂, 0.25 g/L lysozyme, 25 U/mL nuclease and EDTA-free protease inhibitor. Use the ice-cooled 24-tip horn at 50% amplitude for ten cycles of 30 seconds on and 30 seconds off, then clarify at 3,000 × g for 15 minutes. In the reporter test, 50 µL of 25% bead slurry, equivalent to 12.5 µL packed beads, captured approximately 250 µg in 30 minutes. Production uses 200 µL of 25% slurry, or 50 µL packed beads, per sample, leaving capacity for higher expression and for the added protease (§§1.6–1.7, pp. 7–9, §3.6, p. 12).

Wash and release. Bind for at least 30 minutes at room temperature and 500 rpm. Wash three times with 2 mL of 20 mM sodium phosphate, pH 7.4, 500 mM NaCl and 20 mM imidazole, then once with 2 mL PBS. Elute in 500 µL PBS, pH 7.4, with 150 nM AIProtease at room temperature and 500 rpm for approximately two hours. The authors used 1 mM DTT during protease preparation and added no reductant to the production elution buffer. Figure S6 labels purification as roughly 90 minutes without reconciling that label with the two-hour elution hold; use the explicit §3.6 incubation when planning elapsed time (§§3.6–3.7, pp. 12–13, Fig. S6).

Cautions / Limitations

  • Yield reporting: At the stated 500 µL elution volume, >100 µg requires >0.20 mg/mL. Independently reading the rounded means in Fig. S8b gives 71 values above 0.20 and one exactly at that boundary, which does not reconcile with the reported 87 successes. Rounding, per-replicate data and purity criteria prevent treating 71 as a corrected success rate; the underlying yield/volume accounting needs clarification (§§1.9,3.6; Fig. S8).
  • Buffer compatibility. Table S1 measured eluate fluorescence after two hours with 150 nM AIProtease, with n = 1 per buffer, relative to PBS at pH 7.4. For example, 25 mM BICINE at pH 8.4 gave a ratio of 2.5. Both pH-dependent sfGFP brightness and buffer-dependent retention of cleaved reporter on the beads affect that ratio. It cannot establish a 2.5-fold catalytic acceleration or a pH optimum (Table S1§1.2, p. 4).
  • Protease loading. Raising enzyme concentration to 300 or 600 nM gave little kinetic benefit. At the highest concentration, uncleaved fusion and cleaved His–SUMO appeared in eluates with His-tagged proteases. Competition for nickel sites is the authors’ explanation for the loss of selectivity (§1.2, pp. 3–4, Fig. S2).

Open questions / Theoretical implications (2–5 bullets)

  • How much of the gain depends on spacer length, disorder or the particular CAHS sequence? The authors tested one spacer with Cth SUMO protease. They also report that a CAHS–superTEV construct had no detectable activity, with no corresponding data shown (§1.2, p. 4).
  • Can the local structure around a target’s first two residues predict when a native N-terminus will cleave efficiently? Figure S9 measures recovered protein concentration. A matched cleavage time course could separate an effect on protease access from differences in expression or capture (§1.9, pp. 9–10, Fig. S9).
  • How far can culture volume fall before the workflow loses the material needed for characterization? Lower fill volumes improved sfGFP yield per milliliter, yet some benchmark eluates already fell below the concentration needed for CD. Maintaining enough material for the intended assay sets a practical test for miniaturization (Fig. 4b, p. 7, §§1.9–2, pp. 10–11).