Rapid large-scale purification of myofilament proteins using a cleavable His₆-tag

Large-scale production of high-quality myofilament proteins is a standard method for experiments exploring cardiac mechanics. Current approaches to produce and purify proteins are costly and labor intensive. Here, we describe an extensive improvements to the current method that reduce time and increase yield.

Abstract

With the advent of high-throughput DNA sequencing, the number of identified cardiomyopathy-causing mutations has increased tremendously. As the majority of these mutations affect myofilament proteins, there is a need to understand their functional consequence on contraction. Permeabilized myofilament preparations coupled with protein exchange protocols are a common method for examining into contractile mechanics. However, producing large quantities of myofilament proteins can be time consuming and requires different approaches for each protein of interest. In the present study, we describe a unified automated method to produce troponin C, troponin T, and troponin I as well as myosin light chain 2 fused to a His₆-tag followed by a tobacco etch virus (TEV) protease site. TEV protease has the advantage of a relaxed P1′ cleavage site specificity, allowing for no residues left after proteolysis and preservation of the native sequence of the protein of interest. After expression in Esherichia coli, cells were lysed by sonication in imidazole-containing buffer. The His₆-tagged protein was then purified using a HisTrap nickel metal affinity column, and the His₆-tag was removed by His₆-TEV protease digestion for 4 h at 30°C. The protease was then removed using a HisTrap column, and complex assembly was performed via column-assisted sequential desalting. This mostly automated method allows for the purification of protein in 1 day and can be adapted to most soluble proteins. It has the advantage of greatly increasing yield while reducing the time and cost of purification. Therefore, production and purification of mutant proteins can be accelerated and functional data collected in a faster, less expensive manner.

NEW & NOTEWORTHY

Large-scale production of high-quality myofilament proteins is a standard method for experiments exploring cardiac mechanics. Current approaches to produce and purify proteins are costly and labor intensive. Here, we describe an extensive improvements to the current method that reduce time and increase yield.

recent advances in gene screening have identified dozens of novel cardiomyopathy-causing mutations in sarcomeric proteins such as the troponin complex (8). An effort to identify putative consequences of such mutations is currently underway (19, 24). However, functional studies measuring the contractile properties of myofilaments usually require large amount of recombinant protein (1, 4, 7, 18, 21). Traditionally, recombinant proteins are produced by expression in Esherichia coli competent cells followed by purification using fast protein liquid chromatography (FPLC) based on the biophysical properties of the protein of interest (1, 2, 11, 12, 16, 17, 20). For instance, to assemble a functional troponin complex, troponin T is purified by anion exchange on a DEAE fast flow column. Troponin C is purified on a DE52 column and phenyl sepharose based on anion exchange and Ca²⁺ affinity, respectively, whereas troponin I is purified first using cation exchange on a CM Sepharose column and then on a custom troponin C capture column. These methods are very time consuming (4+ days/protein) and have low efficiency (1, 2, 11, 12, 16, 17, 20). To scale up experiments, a rapid method for sarcomeric protein purification is necessary.

One method to streamline production is to use a tag to help in the purification process (25). Unfortunately, for most sarcomeric proteins, any tag or leftover amino acids on either the NH₂-terminal or COOH-terminal can potentially, but not necessarily, affect function. One exception is a myc tag placed on the NH₂-terminus of cardiac troponin T, which has been shown to be benign (3). As a workaround, protease sites can be engineered to cleave off the tag, although until relatively recently, all proteases left one or more amino acids behind (25). Additionally, large quantities of highly active purified protease necessary are cost prohibitive. Recently, a solution to both problems has been found. Novel point mutations to tobacco etch virus (TEV) protease have greatly improved its activity while making it resistant to self-proteolysis. These advances have made the protease much more suitable and reproducible for protein purification. Additionally, TEV protease can be itself His₆-tagged to aid in its purification as well as allowing it to be removed after digestion (23). His₆-TEV protease cleaves at the amino acid sequence of ENLYFQ/G (23). Furthermore, the P1′ recognition site of TEV protease is relatively flexible so that glycine can be substituted by methionine, the universal start codon of proteins, or by almost any other residue (except proline), thereby resulting in a “native” protein sequence after cleavage (9). Indeed, most mammalian and bacterial proteins have their NH₂-terminal processed by methionine amino peptidases that cleave the initial methionine (5). Therefore, care can be taken to have the sequence match the native protein. Conversely, charged residues can replaced the NH₂-terminal amino acid to mimic N-acetylation, such as in the case of tropomyosin (13, 14), without having to alter the purification approach.

In the present study, we describe a rapid method for the purification of recombinant protein to make an active troponin complex for exchange into cardiac myofilament preparations. This new approach allows for a simplified workflow to purify all three troponin proteins as well as myosin light chain 2v (MLC2v), PKA, and others. After expression in a bacterial system, the recombinant protein is initially purified using a nickel-sepharose affinity column followed by a sephadex desalting column to perform buffer exchange into the TEV digestion buffer. After incubation with TEV protease to cleave the His₆ tag off, the protein is run again through the affinity column to remove any undigested protein as well as the protease itself and then through the desalting column again to replace into the desired buffer (22). We show that time and monetary cost are significantly decreased while the yield of the target protein is increased at the same time.

MATERIALS AND METHODS

Transformation and expression of cleavable His₆-tag troponin proteins in E. coli.

DNA fragments of His₆-human cardiac troponin C, troponin I, and troponin T-myc with a TEV protease cleavage site (ENLYFQ/X, where X is the desired first amino acid of the native protein) at the NH₂-terminal (Integrated DNA Technologies) were ligated into pET28a vector (Novagen), which contains a T7 promoter and a His₆-tag coding sequence (Fig. 1A). Given that both mammalian cells and E. coli have methionine aminopeptidases that can cleave NH₂-terminal methionines (5), care was taken to have the appropriate first amino acid (methionine for troponin C, glutamate for c-myc-troponin T, and alanine for troponin I). His₆-tagged human cardiac troponin C and His₆-tagged human cardiac troponin I were transformed into BL21 (DE3) competent cells (Novagen) and plated on Luria agar plates with 40 μg/ml kanamycin, whereas His₆-tagged human cardiac troponin T-myc was transformed into Rosetta (DE3) competent cells (Novagen) and grown on plates with 40 μg/ml kanamycin and 34 μg/ml chloramphenicol. Rosetta cells were chosen as they coexpress rare codons for mammalian proteins and can improve protein translation. Colonies were allowed to grow in a 37°C incubator overnight. Up to 5 colonies/construct were selected and grown in suspension in 4 ml Luria broth (LB) with selection antibiotics at 37°C overnight. Colonies were grown overnight in 4 ml LB and then split into two aliquots; one aliquot was kept uninduced and the other was induced with 1 mM isopropyl β-d-1-thiogalactopyranoside. After 3 h, both bacterial growths were centrifuged, lysed, and run on SDS-PAGE side by side to look at the production of protein after induction (Fig. 1B). The colony that resulted in the highest expression level combined with the fastest growth was determined by looking at the relative amount of protein produced and then selected for large-scale protein production.

Fig. 1.

Construct and expression of mammalian protein in Esherichia coli. A: schematic of the pET 28a vector with tobacco etch virus (TEV) insertion for expression of myofilament proteins. Note that the COOH-terminal His₆-tag occurs after the stop codon and is not incorporated into the protein. B: transformation and induction of His₆-tagged recombinant human cardiac troponin T (HcTnT)-myc (50 kDa) expression using isopropyl β-d-1-thiogalactopyranoside (IPTG) in E. coli. After an overnight growth phase, protein expression was induced for 3 h, bacteria were lysed, and expression was verified by SDS-PAGE followed by Coomassie staining. U, uninduced; I, induced using IPTG; M, marker lane. Uninduced and induced pairs were grown from a single colony. The arrow indicates the colony that exhibited the highest expression of protein and was chosen for further purification.

Large-scale production and purification.

Optimal colonies were picked (Fig. 1B) and grown overnight in 20 ml LB, which was then used to inoculate 1 liter LB with selection antibiotics. Bacteria were allowed to grow for 8 h at 37°C until absorbance at 600 nm, a common surrogate for bacterial growth in suspension, reached 0.4. Induction of protein expression was initiated with 1 mM isopropyl β-d-1-thiogalactopyranoside at 32°C overnight. The bacterial pellet was collected by spinning down the culture at 9,000 g for 20 min at 4°C. Pellets were stored at −80°C until needed.

The bacterial pellet was resuspended and lysed in combined lysis/equilibration buffer [6 M ultrapure urea, 50 mM NaH₂PO₄, 300 mM NaCl, and 0.05% (vol/vol) Tween 20; pH 8.0] at a 1:10 (wt/vol) ratio and then sonicated for 15 min on ice (30 s on and 30 s off). The lysate was centrifuged at 23,000 g for 45 min at 4°C, and the clarified supernatant was loaded onto an automated FPLC system (AKTÄ FPLC, GE Healthcare).

The target protein was purified in three steps (Fig. 2). Step 1 was purification using the His₆-tag with a HisTrap nickel affinity column (GE Healthcare) equilibrated with lysis/equilibration buffer followed by desalting on a HiPrep 26/10 (GE Healthcare) equilibrated with TEV digestion buffer (1.5 M NaCl, 50 mM Tris·HCl, and 5 mM DTT, pH 8.0). The sample was loaded onto the HisTrap column, and the column was washed with 4% elution buffer [6 M ultrapure urea, 500 mM ultrapure imidazole, 50 mM NaH₂PO₄, 300 mM NaCl, and 0.05% (vol/vol) Tween 20, pH 8.0] until ultraviolet absorbance at 280 nm reached equilibrium, around 100 ml. His₆-tagged proteins were eluted out with 50 ml of 70% elution buffer, and their buffer was then exchanged with 53 ml TEV buffer using the desalting column. In step 2, the His₆-tag was cleaved by TEV protease at 0.3 mg protease/mg target protein at 4°C overnight or at 30°C for 4 h. In step 3, finally, any undigested protein and the protease itself was removed using the HisTrap nickel affinity column. At this step, however, the flowthrough was collected instead and then desalted with 70 ml of the desired buffer. To facilitate troponin complex formation, troponins were desalted into urea buffer (6 M ultrapure urea, 1 M KCl, 3 mM MgCl₂, 0.5 mM DTT, and 10 mM MOPS, pH 7.0) (1, 17), whereas other proteins, such as MLC2v, were desalted into a simplified relax buffer (10 mM EGTA, 100 mM BES, 66.32 mM KOH, 15 mM NaCl, 6.48 mM MgCl₂, and 49.76 mM potassium propionate, pH 7.0).

Fig. 2.

Flowchart of the purification of protein using a cleavable His₆-tag. Starting from bacteria, proteins were purified using two successive fast protein liquid chromatography (FPLC) runs. Proteins can then be frozen or used to form the troponin complex, which can be further purified by FPLC. TnT, troponin T; TnI, troponin I; TnC, troponin C.

Production of His₆-TEV protease.

The TEV protease plasmid, pRK793, was a gift from David Waugh (Addgene plasmid no. 8827) (10). His₆-TEV protease was expressed in Rosetta (DE3) competent cells as described above and lysed in lysis buffer without urea [50 mM NaH₂PO₄, 300 mM NaCl, and 0.05% (vol/vol) Tween 20, pH 8.0]. Purification was performed using a HisTrap nickel affinity column, and the elution buffer was replaced with 2× TEV digestion buffer (3 M NaCl, 100 mM Tris·HCl, and 10 mM DTT, pH 8.0) using the desalting column. His₆-TEV protease was stored at −20°C in 50% glycerol to prevent freezing.

Formation and purification of the troponin protein complex.

The troponin complex was formed by sequential desalting mimicking the sequential dialysis procedure previously described (1, 17). Briefly, equal molar quantities of each subunit were mixed, and the buffer was slowly depleted of urea and salt, which allows denatured proteins to refold back into the native conformation. The starting concentration of all three troponin subunits (T, I, and C) was 60 μM in 1 ml urea buffer, and 50 mM l-arginine and 50 mM l-glutamine were added to help prevent degradation and precipitation that can occur later. Four sequential desalting steps were then performed by FPLC. The first step was with 60 ml of 2 M Urea, 1 M KCl, 3 mM MgCl₂, 0.5 mM DTT, and 10 mM MOPS (pH 7.0) and then with 60 ml of 1 M KCl, 3 mM MgCl₂, 0.5 mM DTT, and 10 mM MOPS (pH 7.0) and finally with 60 ml of 0.15 M KCl, 3 mM MgCl₂, 0.5 mM DTT, and 10 mM MOPS (pH 7.0). The collected peak from the third desalting step, containing the now-assembled troponin complex, was then loaded onto a ResourceQ column (GE Healthcare) equilibrated with 0.15 M KCl, 3 mM MgCl₂, 0.5 mM DTT, and 10 mM MOPS (pH 7.0). A gradient (1%/ml) of troponin elution buffer (1 M KCl, 3 mM MgCl₂, 0.5 mM DTT, and 10 mM MOPS, pH 7.0) was then used to separate the formed complex from troponin monomers. The peak fractions were then pooled, and SDS-PAGE was performed to validate troponin complex purity and concentrated using a centrifugal filter unit, or “spin column,” with a cutoff of 10,000 Da (Millipore) at 3,000 g for 30 min or until a concentration of ∼2 mg/ml was reached, according to the manufacturer's instructions.

Functional comparison of the recombinant troponin complex produced by the traditional versus rapid method exchanged in human permeabilized myofilaments.

The Institutional Review Board of Loyola University (Chicago, IL) approved the protocol for the use of deidentified human donor samples. Permeabilized (skinned) myofilaments were prepared from healthy human myocardium as previously described (1) by an incubation in relaxing solution [containg (in mM) 97.92 KOH, 6.24 ATP, 10 EGTA, 10 Na₂CrP, 47.58 K-proprionate, 100 BES, and 6.54 MgCl₂] with 1% Triton X-100 added for 10–15 min on a slowly rotating shaker. Triton X-100 was then removed, and myofilaments had their troponin compex exchanged with 2 mg/ml recombinant complex overnight at 4°C. Myofilaments were then washed three times in relaxing solution, attached with glue to a force transducer and motor, and exposed to free Ca²⁺ concentrations ranging the activation of contraction at both short (1.9 μm) and long (2.3 μm) sarcomere length. Force data were fit to a modified Hill equation for each individual cell to obtain the Hill coefficient. pCa₅₀ was calculated from the Ca²⁺ concentration required to achieve 50% force activation. Data are expressed as means ± SE. Statistical analysis was performed using two-way ANOVA as appropriate (SigmaPlot).

RESULTS

Purification of cleavable His₆-tag protein produces native myofilament proteins. His₆-tagged human cardiac troponin C, I, and T (with the myc tag at the NH₂-terminal) with a TEV protease cleavage site were expressed in competent cells and purified by FPLC (Figs. 1B and 3). After purification, protein purity was assessed by SDS-PAGE. Here, we show that our novel and rapid method results in purified native troponin C and troponin I (Fig. 4, A and B), whereas troponin T exhibited two degradation products (Fig. 4C) that were <5% of the total protein and did not incorporate into the complex after purification, as they were absent from gels of the assembled complex (Fig. 5C). Additionally, we successfully purified MLC2v (Fig. 4D), which demonstrated <1% of nonintact protein for MLC2v. Yields per liter of bacterial growth were significantly higher using FPLC for the purification, with 14.8 ± 0.5 mg/l of tropoinin C, 13.0 ± 0.7 mg/l of troponin T, and 13.6 ± 1.0 mg/l of troponin I compared with 2–4 mg/l using traditional approaches. Importantly, the new FPLC method can easily accommodate several liters of medium in the same run to produce >50 mg protein per purification.

Fig. 3.

Representative traces of the purification of TnC by FPLC. A: initial purification of TnC using the His₆ tag. B: removal of undigested proteins and protease from purified TnC. In both A and B, the top graph shows protein concentration measured as absorbance at 280 nm was plotted against elution volume through the FPLC columns, whereas salt concentration measured as conductivity was plotted concurrently in dashed lines. The bottom graph shows eluent concentration as a percentage of the total buffer.

Fig. 4.

Purification of myofilament proteins TnC, TnI, and TnT (A–C) and the human myosin light chain 2 ventricular isoform (MLC2v; D). For TnT, there were very few degradation products when purified properly compared with traditional methods (C, right). All results were screened by 15% SDS-PAGE and visualized by Coomassie staining. M, molecular weight standard (in kDa). Lane 1, clarified E. coli lysate; lane 2, pooled peak fraction after the first run through HisTap and HiPrep desalting columns; lane 3, after overnight TEV protease digestion; lane 4, pooled peak fraction after the second run through HisTrap and HiPrep desalting columns.

Fig. 5.

On-column forming via desalting and purification of the troponin complex by FPLC. A: serial desalting by FPLC to refold the troponins and allow them to form a complex. B: purification of the recombinant human troponin complex on a ResourceQ column with an increasing concentration of KCl. Samples were run using 12% SDS-PAGE and visualized by Coomassie staining. C: consecutive 5-ml fractions from the third FPLC run shown in B were then assessed via SDS-PAGE. Supporting the labels in B, fraction 1 displayed a band with a molecular weight consistent with TnI and fractions 3 and 4 with TnC. The complex eluted out in fractions 5–11, and a TnT-TnI dimer eluted out in fraction 12. M, molecular weight standards (in kDa). D: the complex purity after concentration by spin column was verified using size-exclusion chromatography by FPLC. Only a single peak of protein eluted out of the size-exclusion column at a volume consistent with the size of the troponin complex.

Recombinant human troponin complex (T, I, and C) assembly and purification by FPLC.

Troponin T, I, and C were purified as described above. Equal amounts of troponin T, I, and C were mixed and then desalted sequentially to decrease urea and salt concentration, thereby allowing the denatured proteins to slowly refold back into their native conformation and form a protein complex without precipitating (Fig. 5A). After purification on a ResourceQ column, we collected all fractions at 1-ml intervals (Fig. 5B). Troponin I eluted out at 5–10% of elution buffer, and troponin C eluted out at ∼20% of elution buffer. Troponin T typically precipitated during dialysis as the salt concentration and its solubility decreased. The complex eluted out at ∼25% elution buffer (Fig. 5C). The eluted complex, then concentrated via spin column, did not contain any monomers or dimers of troponin C, I, or T (Fig. 5D). While the yield of the complex was slightly lower that using the dialysis method (∼23% lower, 3.0 ± 0.1 vs. 3.9 ± 0.2 mg, P < 0.05, modified vs. traditional method), the time savings of sequential desalting greatly made up for the loss of protein (Table 1).

Table 1.

Approximate cost and speed comparison between the traditional method and our modified method

	Traditional Method			Modified Method
	Cost per run, $	Cost per milligram, $	Time per run, h	Cost per run, $	Cost per milligram, $	Time per run, h
Troponin I	145	15	180	76	6	27
Troponin T	59	15	136	74	5	25
Troponin C	85	7	181	53	3	24
Troponin complex	42	10	50	9	4	8.6

The recombinant troponin complex behaves similarly using both purification methods.

To determine whether the modifications to the production and purification of the troponin complex affect its function compared with traditional methods, we performed exchange experiments in a frozen human left ventricular myocardium sample and measured force-pCa responses (Fig. 6, A–F). Both complexes exchanged to a similar extent into permeabilized myofilament preparations (<90%; data not shown). We found that there was no difference in any parameter measured between complexes produced by the new method compared with complexes purified using traditional methods.

Fig. 6.

Troponin complex produced by sequential desalting behaves identically to the troponin complex produced by traditional methods. A and B: force (F)-pCa curves at short (A) and long (B) sarcomere lengths (SLs) showed no differences between methods (n = 5 traditional method and n = 8 modified method from one human donor sample, P = not significant). Similarly, normalized curves at short (C) and long (D) SLs, pCa₅₀ (E), and Hill coefficient (n_H; F) were similar.

DISCUSSION

In the present study, we describe a rapid large-scale purification of myofilament proteins by adapting common techniques used in other fields to simplify the current standard methods. We demonstrated that, using cleavable tags to aid purification, we can produce large quantities of fully functional troponins that can assemble into a complex and are no different from troponins produced in traditional ways. Furthermore, we show that assembling the complex on a desalting column speeds the process up even further and does not substantially affect yield.

The modifications that we have made to the traditional method have tremendously reduced the time needed to produce fully functional troponin complex (Table 1). While traditionally it would require up to 2 weeks to go from a plasmid to a fully assembled complex using traditional methods, our improvements allowed us to cut the time by half, with a fully purified and assembled complex available within 4–7 days, with most of the time used for bacterial growth and protein expression (3–5 days) while the protein purification and complex assembly can be done in 1–2 days.

One major obstacle to increasing protein yield is precipitation. As the amount of protein per milliliter of solution is increased, we also increased the protein lost to precipitate. As a workaround, we adopted two strategies. The first strategy consisted in keeping all volumes large and the proteins diluted until the very last step of purification and complex assembly. Second, we added two charged amino acids, l-glutamate and l-arginine, at a relatively high concentration (50 mM each) and dramatically increased the maximum amount of soluble protein (6). These amino acids were then removed at the same time as the concentration step, using a spin column with a cutoff of 10,000 Da. These two strategies allowed us to increase our yield per run by three to five times our traditional purification procedures while maintaining the time savings.

Further improvements to the technique can be achieved, such as using a gel filtration column to increase purity or using an on-column digest with immobilized TEV protease. A more advanced FPLC system could also theoretically combine all three steps of the purification into a single method, with several holding loops and column valves. However, the time gained from such techniques would not outweigh the additional costs of the purification, as buffer switching in FPLC uses large volumes of buffer and a significant amount of time. We believe that we have identified a good compromise that increases efficiency and yield while also decreasing cost.

In conclusion, the method described above borrows from different protein purification methods to result in a simple, rapid, high-yield procedure to produce large amounts of troponin complex. This greatly improves current techniques and should save investigators time and allow experiments with several different mutations in troponins to be carried simultaneously with little extra effort.

GRANTS

This work was partially supported by National Heart, Lung, and Blood Institute Grants HL-62426 and HL-75494 (to P. P. de Tombe) and F32-HL-120643 (to R. J. Khairallah).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: M.Z., R.J.K., and P.P.d.T. conception and design of research; M.Z., J.L.M., M.K., and R.J.K. performed experiments; M.Z., J.L.M., M.K., R.J.K., and P.P.d.T. analyzed data; M.Z., J.L.M., R.J.K., and P.P.d.T. interpreted results of experiments; M.Z., J.L.M., and R.J.K. prepared figures; M.Z. and R.J.K. drafted manuscript; M.Z., J.L.M., R.J.K., and P.P.d.T. edited and revised manuscript; M.Z., J.L.M., M.K., R.J.K., and P.P.d.T. approved final version of manuscript.