Overcoming challenges in the production of a human-like L-asparaginase: advanced strategies for expression and purification

Highlights

• •Mild expression conditions reduced hASNase1 aggregation in Escherichia coli.
• •Asparagine incubation improved hASNase1 purification by reducing chaperones.
• •hASNase1 exhibits allosteric kinetics with a Hill coefficient of 3.9.
• •In silico analysis identified hydrophobic regions linked to aggregation.

Abstract

Human l-asparaginase 1 (hASNase1) is a promising next-generation candidate for Acute Lymphoblastic Leukemia (ALL) therapy due to its reduced immunogenicity and superior physiological compatibility. However, its recombinant production remains challenging, as expression in Escherichia coli typically results in insoluble aggregates and strong association with host chaperones, limiting biochemical characterization. Here, we established an optimized workflow combining mild induction conditions with tailored solubilization and purification strategies. Incubation with l-asparagine followed by ATP-mediated washing during affinity chromatography significantly improved protein recovery while reducing chaperone co-purification. This approach enabled the isolation of catalytically active hASNase1 displaying allosteric regulation and cooperative substrate binding, consistent with its proposed tetrameric organization. Complementary in silico analyses identified aggregation-prone regions, potentially providing structural insight into the observed expression challenges. Collectively, this study contributes to the development of improved recombinant production strategies for challenging human enzymes in prokaryotic systems.

Graphical abstract

1. Introduction

L-asparaginase is one of the most clinically successful enzyme-based therapies and remains an essential component of chemotherapy regimens for Acute Lymphoblastic Leukemia (ALL) [[1], [2], [3]]. By depleting circulating l-asparagine, this enzyme selectively induces apoptosis in leukemic cells that lack the ability to synthesize this amino acid [[4], [5], [6]]. Despite its decades-long clinical success, currently available l-asparaginase formulations, derived from E. coli and Erwinia chrysanthemi, are associated with serious drawbacks [[7], [8], [9]]. Their bacterial origin leads to high immunogenicity, resulting in hypersensitivity reactions, neutralizing antibodies, and treatment interruptions [4,10,11].

A promising solution lies in human l-asparaginase 1 (hASNase1), a human enzyme with asparaginase activity and reduced risk of immune recognition. This enzyme corresponds to the N-terminal domain (residues 1–369) of a 60-kDa lysophospholipase (573 residues). The full 60-kDa lysophospholipase displays multiple catalytic activities, including transacylase, platelet-activating factor acetylhydrolase (PAF-AH), lysophospholipase, and asparaginase-like functions. It also participates in the regulation of epithelial sodium channels (ENaC) and cell proliferation [6,12,13] (Uniprot code Q86U10, LPP60_HUMAN). The hASNase1 subunit was named for its high sequence similarity to E. coli type I l-asparaginase [14,15]. Unlike its bacterial counterparts, hASNase1 exhibits greater biological compatibility, supporting further biochemical and structural studies, and potential future commercial development. However, despite its potential, only a handful of studies have explored the biochemical properties of hASNase1 and the challenges associated with its recombinant production [6,14,16].

The heterologous production of biopharmaceuticals is well established and utilizes a range of host systems, including bacterial, yeast, and mammalian cells [1,2]. Among these, E. coli has emerged as a particularly efficient and cost-effective platform for recombinant protein biosynthesis, enabling the production of numerous therapeutics such as hormones, enzymes, and monoclonal antibodies [[17], [18], [19], [20]]. Despite recent advancements in resolving common problems using the E. coli expression system, for some proteins, it frequently results in insoluble inclusion bodies, aggregation-prone intermediates, and strong binding to host-derived chaperones, all of which hinder purification and impair enzymatic activity [21]. These bottlenecks have delayed systematic characterization and the advancement of hASNase1 toward preclinical or clinical applications [[22], [23], [24]].

To date, only two studies have reported recombinant expression hASNase1. Karamitros and Konrad, 2014 [14] obtained mostly insoluble aggregates requiring complex ATP buffer and chaperone-assisted purification, whereas Belviso et al., 2017 [6], achieved refolding from inclusion bodies without recovering activity. In contrast, our study introduces an optimized strategy combining low-temperature expression, systematic buffer screening, and mild solubilization, which allows the recovery of soluble, active hASNase1 without refolding. These improvements significantly enhance yield and purity while minimizing chaperone contamination, thereby overcoming the major technical bottlenecks identified in previous works.

In this study, the critical gaps in the production of hASNase1 are addressed by optimizing heterologous expression, refining solubilization conditions, and implementing innovative purification strategies that stabilize oligomeric assembly while minimizing chaperone contamination. Beyond improving the active enzyme recovery, this study also provided structural and biochemical insights that explain its aggregation tendency, offering valuable guidance for future engineering efforts. By establishing a robust framework for producing catalytically active hASNase1, this work supports further investigation of its biochemical properties and functional potential.

2. Methods

2.1. Expression and optimization of culture conditions for recombinant hASNase1

The codon-optimized gene encoding the N-terminal domain (amino acid residues 1 to 369) of human lysophospholipase (hASNase1, 60 kDa) was cloned into the pET-28a vector for E. coli expression, incorporating an N-terminal 6 × His-tag fused to a SUMO tag. The resulting recombinant construct was transformed into E. coli BL21(DE3) Star, BL21(DE3) Rosetta, and ArcticExpress strains for expression trials.

Each transformed E. coli strain was cultured in 50 mL of Terrific Broth (TB – 12 g/L tryptone; 24 g/L yeast extract; 2.2 g/L potassium phosphate monobasic; 9.4 g/L potassium phosphate dibasic; 4 mL/L glycerol) until reaching an optical density at 600 nm (OD₆₀₀) of 0.6–0.8, at which point protein expression was induced with isopropyl-β-d-thiogalactopyranoside (IPTG – Sigma-Aldrich) (Table 1). Samples were collected before and after induction, and cultures were centrifuged at 8000 x g for 5 min. Pellets were resuspended in lysis buffer (50 mM Tris pH 8.0; 0.3 M NaCl; 1 mM Phenylmethylsulfonyl fluoride (PMSF); 0.01% Triton X-100) at 25 µL per 0.1 OD₆₀₀ unit, followed by sonication on a Q700 Sonicator (5 cycles: 5 s ON/15 s OFF, 4% amplitude). Protein expression was analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). To assess solubility, lysates were centrifuged at 10,000 x g for 10 min at 4 °C to separate soluble (supernatant) and insoluble (pellet) fractions.

Table 1.

Expression conditions of hASNase1 in different E. coli strains.

Condition	Strain	Antibiotics	OD600	IPTG	Temperature	Expression time
1	BL21 (DE3) Star	Kanamicin	0.6 – 0.8	1 mM	28/37 °C	4 h
2	BL21 (DE3) Rosetta	Kanamicin and Chloramphenicol	0.6 – 0.8	0.3 mM	15 °C	17 h
3	Arctic Express System	Ampicilin and Gentamicin	0.6 – 0.8	0.1/0.2/0.3 mM	10 °C	4 h/17 h

Expression conditions for E. coli BL21(DE3) Rosetta were optimized for large-scale production of hASNase1 in 500 mL cultures, as described in Table 1. After incubation, cultures were centrifuged at 10,000 x g for 10 min at 4 °C. Expression of hASNase1 was confirmed by Western blot, following the protocol by Gomes, 2023 [25]. Proteins were separated by 12.5% SDS-PAGE and transferred onto a nitrocellulose membrane in a BioRad Trans-Blot Turbo system. Detection was performed using anti-His antibodies conjugated to alkaline phosphatase (Sigma-Aldrich, A5588) and visualized with Western Blue stabilized substrate (Promega).

2.2. Solubility assessment of hASNase1

Biomass from expression condition 2 (Table 1) was resuspended in the lysis buffers listed in Table 2 using a standardized volume of 25 µL per 0.1 OD₆₀₀ unit. Cell disruption was performed by 30 sonication cycles (20 s ON/ 40 s OFF, 30% amplitude), followed by centrifugation to separate soluble (supernatant) and insoluble (pellet) protein fractions. Both fractions were analyzed by 12.5% SDS-PAGE to evaluate the distribution of recombinant protein.

Table 2.

Buffers utilized in the development of a solubilization protocol for the insoluble fraction of hASNase1.

Strains	Lysis buffer	Composition
BL21 DE3 star	Buffer 1	20 mM Tris pH 8.0; 0.5 M NaCl; 5% Glycerol; 1 mM β-mercaptoethanol.
	Buffer 2	20 mM Tris pH 8.0; 1 mM ethylenediaminetetraacetic acid (EDTA), 0.1% Triton X-100, 500 mM NaCl.
	Buffer 3	20 mM Tris pH 8.0; 0.1% Triton X-100, 0.4 M l-arginine.
	Buffer 4	Cell litic (Sigma-Aldrich)
	Buffer 5	20 mM Tris pH 8.0; 2 M, 4 M, 6 M or 8 M Urea.
Rosetta and Arctic	Buffer 6	50 mM Tris pH 8.0; 0.3 M NaCl; 0.01% Triton X-100.
Rosetta	Buffer 7	50 mM Tris pH 8.0; 0.3 M NaCl; 0.01% Triton X-100; 0.5 M, 1 M or 2 M Urea.

2.3. Refolding

hASNase1 was initially solubilized in 5 mL of Buffer 5 (2 M urea). Refolding was performed by buffer exchange using an Amicon Ultra filtration unit (GE Healthcare) with a 10 kDa molecular weight cutoff. The refolding process involved sequential additions of urea-free buffer (20 mM Tris–HCl, 150 mM NaCl, pH 7.4) to gradually reduce the urea concentration. First, 2 mL of urea-free buffer was added to the sample (5 mL), followed by centrifugation at 3000 × g for 10 min at 4 °C. After each centrifugation step, 2 mL of filtrate was discarded, and an equal volume (2 mL) of fresh urea-free buffer was added to the retentate. This process was repeated for a total of 6 cycles, progressively reducing the urea concentration to approximately 0.26 M. Subsequently, 12 additional cycles were performed using 3.5 mL buffer exchanges per step, further decreasing the urea concentration to approximately 0.4 mM. The urea concentration at each step was estimated using the dilution equation (C₁V₁ = C₂V₂), where C₁ is the initial urea concentration, V₁ the initial sample volume before buffer addition, C₂ the resulting urea concentration, and V₂ the total volume after buffer addition.

2.4. Immobilized metal affinity chromatography

Bacterial pellets were resuspended in buffer 6 with 1 mM PMSF, sonicated, and centrifuged (10,000 x g, 30 min, 4 °C) to separate soluble (supernatant) and insoluble fractions (pellet). Soluble extracts (30 mL) were incubated with 1 mL nickel resin (Ni-resin – HisLink, Promega) in Tris buffer (50 mM Tris, 300 mM NaCl, pH 8.0), for 2 h at 4 °C (130 rpm). The mixture was loaded onto a polypropylene column and flow-through was recirculated twice to enhance binding. The resin was washed with 20 column volumes of wash buffer (50 mM Tris, 300 mM NaCl, 40 mM imidazole, pH 8.0) to remove non-specifically bound proteins, and elution was performed with a Tris buffer containing 300 mM imidazole buffer. Purity was monitored by SDS-PAGE, and eluates were dialyzed a Tris buffer (50 mM Tris, 300 mM NaCl, pH 8.0), tested for asparaginase activity, and stored at –80 °C in 16% (w/v) sucrose for future analysis.

2.5. Anion exchange chromatography followed by immobilized metal affinity chromatography

The soluble fraction was loaded onto a 1 mL DEAE FF column (GE Healthcare) equilibrated with Buffer A (20 mM Tris, pH 8.0) on an ÄKTA pure system. Elution was performed with a linear gradient of Buffer B (20 mM Tris, 500 mM NaCl, pH 8.0) from 0 to 100% over 40 mL, collecting 20 fractions (2 mL each). Fractions containing hASNase1 were pooled and purified by immobilized metal affinity chromatography (IMAC) (Section 2.4). Final samples were dialyzed in Tris Buffer and tested for activity.

2.6. Mass spectrometry

Samples were subjected to trypsin digestion. Following the digestion step, peptides were purified and concentrated using ZipTip C18 pipette tips (Merck Millipore), according to the manufacturer's protocol. Purified peptides were resuspended in 0.1% formic acid in LC-MS-grade water (LiChrosolv, Merck Millipore) and quantified by fluorometry using the Qubit 4 system. NanoLC-MS was performed with an EASY-Spray source coupled to a Q Exactive Plus Biopharma (ThermoFisher Scientific) in high-resolution Orbitrap mode. Fragmentation was by HCD, and spectra were analyzed in Xcalibur 2.2 [26].

2.7. Strategies to refine hASNase1 isolation

2.7.1. IMAC in the presence of asparagine

In this assay, the clarified supernatant from hASNase1 expression (assay 2, Table 1) was split into 2 Erlenmeyer flasks and incubated with HisLink Ni-resin (Promega) at 4 °C for 2 h under agitation (100 rpm). A 30 mM l-asparagine substrate was added at 2 different time points: 30 and 60 min after the start of incubation. IMAC was then carried out as described in Section 2.4. Aliquots from each step were collected and analyzed by 12.5% SDS-PAGE.

2.7.2. Incubation and washing in the presence of ATP

Ni-NTA affinity chromatography was performed as described by Karamitros & Konrad, 2014 [14], with modifications. Specifically, 10 mM ATP, 20 mM MgCl₂, and 50 mM KCl were added to the incubation of the clarified supernatant with Ni-resin in Erlenmeyer flask, which was carried out at 4 °C for 2 h under agitation. After binding, the resin was washed with buffer C (50 mM Tris, 300 mM NaCl, 10 mM ATP, 20 mM MgCl₂, 50 mM KCl, pH 8.0), followed by a second wash with 10 column volumes of wash buffer with 50 mM Tris, 300 mM NaCl, and 40 mM imidazole (pH 8.0). Aliquots from each step were collected and analyzed by 12.5% SDS-PAGE. Protein concentration was determined by BCA assay. The relative abundance of hASNase1 (%) was estimated by densitometric analysis of SDS-PAGE bands using ImageJ. The amount of hASNase1 (mg) was calculated based on its relative abundance. Recovery (%) was calculated relative to the total hASNase1 present in the crude lysate.

2.7.3. IMAC in the presence of asparagine and washing with ATP

The assay was conducted by incubating the protein with Ni-resin in the presence of 30 mM l-asparagine for 30 min (Section 2.7.1), followed by IMAC with an additional wash using buffer C (Section 2.7.2). Purified proteins were dialyzed against a Tris buffer and subsequently assessed for asparaginase activity. Protein concentration was determined by BCA assay. The relative abundance of hASNase1 (%) was estimated by densitometric analysis of SDS-PAGE bands using ImageJ. The amount of hASNase1 (mg) was calculated based on its relative abundance. Recovery (%) was calculated relative to the total hASNase1 present in the crude lysate.

2.7.4. x his-sumo cleavage

To remove the N-terminal His₆-SUMO tag, purified hASNase1 was incubated with SUMO protease (ULP) at a 1:10 (protease:protein) molar ratio for 16 h at 4 °C. The reaction mixture was then applied to a nickel affinity column: the cleaved protein was recovered in the flow-through, while the His₆-SUMO tag and SUMO protease were eluted with a Tris buffer containing 300 mM imidazole.

2.8. Enzyme activity assay

Enzymatic activity was assessed using the Nessler assay, adapted from previously described methods [27,28]. This colorimetric assay detects ammonia released during the hydrolysis of l-asparagine into l-aspartate and ammonia [29]. For each reaction, 500 µL of 100 mM l-asparagine was mixed with 500 µL of enzyme solution and incubated at 37 °C for 30 min. The reaction was stopped by adding 100 µL of 1.5 M trichloroacetic acid (TCA), after which 125 µL of the mixture was added to 1.75 mL of distilled water and 125 µL of Nessler’s reagent. Absorbance was measured at 436 nm using an EnVision 2105 multimode plate reader (PerkinElmer). All reactions were performed in triplicate.

Ammonia concentrations were calculated using a standard curve prepared with ammonium sulfate ((NH₄)₂SO₄; 0.375–12 mM). 1 unit of enzyme activity was defined as the amount of enzyme that releases 1 µmol of ammonia per minute at 37 °C.

For kinetic analyses, specific activity (U/mg) was measured at substrate concentrations from 10 to 100 mM using the same assay. Kinetic parameters were determined by fitting the data to the Hill equation using OriginPro 8.5, assuming an enzyme concentration of 0.5 µM (based on quantitative concentration of densitometry analysis). The kinetic parameters presented correspond to the optimized purification batch and represent technical triplicate measurements. Enzymatic activity measurements were normalized to the total protein concentration, as determined by BCA assay. Activity values were interpreted considering the relative abundance of hASNase1 in the preparation, as estimated by densitometric analysis of SDS-PAGE bands.

2.9. Native PAGE and Western blot analysis

Native polyacrylamide gel electrophoresis (native PAGE) was performed to evaluate the oligomeric state of hASNase1 under non-denaturing conditions. Protein samples were prepared in the absence of SDS and reducing agents and loaded onto a 7% polyacrylamide gel.

For Western blot analysis, proteins separated by native PAGE were transferred onto a nitrocellulose membrane. The membrane was blocked with 5% non-fat milk in PBS-T and incubated with a primary antibody against hASNase1 (1:10.000) followed by a suitable HRP-conjugated secondary antibody (1:15.000). Detection was performed using Pierce™ DAB Substrate (ThermoFisher) method. Molecular weight estimation was performed using HiMark™ Pre-stained Protein Standard (Invitrogen).

2.10. In silico analyses of hASNase1 structure

To better understand the structural features of hASNase1, in silico analyses were performed. As no experimental structure is currently available, a structural model previously generated by Guimarães et al., 2021 [15], using the Robetta server program [30], was employed. Aggregation-prone regions were predicted using AGGRESCAN3D [31], based on the final molecular dynamics (MD) structure obtained after 300 ns of simulation in an aqueous environment. This MD simulation, which provides a more realistic representation of residue exposure in solution, was carried out as described by Guimarães, et al., 2021 [15].

3. Results

3.1. Expression and solubility evaluation

hASNase1 exhibits asparaginase activity and corresponds to the N-terminal domain of the 60 kDa human lysophospholipase (Fig. 1), from the amino acid residue 1 to 369. The C-terminal ankyrin repeat domain, which does not contribute to asparaginase activity, was excluded from the construct, following the strategy described by Karamitros and Konrad, 2014 [14].

Fig. 1.

Schematic representation of 60-kDa human lysophospholipase. A) The N-terminal domain contains asparaginase activity, while the C-terminal domain comprises a series of ankyrin repeats. B) The gene that codifies hASNase1 cloned into pET-28a+SUMOtag vector.

Expression of hASNase1 in E. coli BL21(DE3)-Star was first evaluated at 28 °C and 37 °C after IPTG induction. The E. coli BL21(DE3)-Star strain was initially selected for its high expression capacity and for decreasing RNA and protein degradation due to rne131 mutation and Ion and OmpT protease deficiency, respectively [32,33]. As shown in Figure S1, both temperatures yielded high levels of recombinant protein (∼52 kDa). The red arrow in Figure S1 indicates the band corresponding to His₆–SUMO–hASNase1 (approximately 52 kDa), comprising the hASNase1 protein (∼42 kDa) fused to the 10 kDa His₆–SUMO tag. However, solubility assessment (Figure S2A–B) showed that hASNase1 was entirely present in the insoluble fraction, indicating that the protein accumulated as inclusion bodies.

To improve solubility, various lysis buffer compositions were tested under milder expression conditions (28 °C), using buffers 3, 4, and 5 (Table 2). Mild concentrations of non-denaturing detergents, such as Triton X-100, can enhance the solubilization of protein aggregates by reducing hydrophobic interactions between partially folded polypeptides without disrupting their native structure [34,35]. Buffer 3 also contained l-arginine, an additive known to facilitate the solubilization of loosely packed inclusion bodies through direct interaction with charged and hydrophobic residues [36,37]. Similarly, the commercial Cell Lytic reagent (buffer 4) includes mild detergents designed to prevent aggregation [38]. As determined by SDS-PAGE (Fig. 3), neither buffer 3 (L-arginine) nor buffer 4 (commercial Cell Lytic) solubilized aggregates. In contrast, complete solubilization was achieved in 4–8 M urea (buffer 5), whereas 2 M urea was insufficient.

Fig. 3.

Expression and solubility analysis using BL21 (DE3) Rosetta strain at 15 °C for 17 h. A) SDS-PAGE 12.5%. B) His-Tagged Recombinant Protein Detection via Western Blot with Alkaline Phosphatase-Conjugated Anti-His Antibody. Legend: MM. Molecular Marker (TrueColor High Range); 1. Non-induced pellet; 2. Cell pellet induced by 1 mM IPTG; 3. Soluble fraction; 4. Insoluble fraction. The red arrow indicates the presence of hASNase1 in the soluble fraction.

3.2. Expression and solubility of hASNase1 in E. coli DE3 Rosetta

To further improve solubility, expression was assessed in E. coli BL21(DE3)-Rosetta under the conditions described by Karamitros and Konrad, 2014 [14], with minor modifications (15 °C, 0.3 mM IPTG, 17 h; condition 2, Table 1). The Rosetta strain was employed because it supplies tRNAs for rare codons present in the human gene sequence, improving translation efficiency for heterologous protein expression [39]. Under these conditions, a weak band corresponding to His₆-SUMO-hASNase1 (∼52 kDa) appeared in the soluble fraction and western blotting confirmed its identity using anti-His antibody (Figs. 3).

Despite this improvement, a large proportion of the protein remained insoluble. Nevertheless, this expression strategy yielded sufficient soluble material for activity assays. To optimize recovery, additional solubilization experiments were performed using lower urea concentrations (0.5–2 M). SDS-PAGE analysis (Fig. 4) showed increased solubility in 1 M urea and complete solubilization in 2 M urea. These conditions were superior to the earlier protocol (28 °C, 1 mM IPTG), where hASNase1 remained insoluble even in 2 M urea (Fig. 2B). Importantly, low urea concentrations (< 4 M) have been shown not to disrupt protein structure and allow refolding of enzymatically active species [40,41].

Fig. 4.

Solubility assay of hASNase1 at varying urea concentrations. Legend: MM. Molecular Marker (TrueColor High Range); 1. Cell pellet induced by 0.3 mM IPTG; 2. Soluble fraction - 50 mM Tris pH 8.0; 0.3 M NaCl (buffer without urea); 3. Insoluble fraction - 50 mM Tris pH 8.0; 0.3 M NaCl (buffer without urea); 4. Soluble fraction - Buffer 7, 0.5 mM urea; 5. Insoluble fraction - Buffer 7, 0.5 mM urea; 6. Soluble fraction - Buffer 7, 1 M urea; 7. Insoluble fraction - Buffer 7, 1 M urea; 8. Soluble fraction - Buffer 7, 2 M urea; 9. Insoluble fraction - Buffer 7, 2 M urea.

Fig. 2.

Solubility assessment from induced clones (1 mM IPTG) at 28 °C. A) Buffers 3 and 4. B) Urea 2 M and 4 M (buffer 5). C) 6 M and 8 M (buffer 5). Legend: MM. Molecular Marker (Benchmark protein ladder, Thermo Fisher Scientific); 1. Non-induced pellet; 2. Cell pellet induced by 1 mM IPTG; 3. Soluble fraction; 4. Insoluble fraction; 5. Cell pellet induced by 1 mM IPTG; 6. Soluble fraction; 7. Insoluble fraction. Red arrows indicating hASNase1 band.

After IMAC purification (Figure S3), urea was gradually removed by buffer exchange for refolding. This buffer exchange step was aimed at restoring the three-dimensional structure of the protein and obtaining it as individual active units. However, Nessler assay measurements indicated no detectable enzymatic activity, suggesting that hASNase1 failed to correctly refold into its active conformation.

3.3. Expression in the ArcticExpress system

Considering the challenges in obtaining soluble hASNase1 in E. coli, expression was subsequently evaluated using the ArcticExpress strain (Agilent Technologies), which is engineered for low-temperature protein production through the co-expression of cold-adapted chaperonins (Cpn60/Cpn10) that assist protein folding at 10 °C [42]. Induction was performed with 0.1 mM IPTG (Fig. 5) and 0.2/0.3 mM IPTG (Figure S4) for either 4 h or 17 h. As shown in Fig. 5, this system did not improve solubility and overall expression levels were lower than those obtained using BL21(DE3)-Star and Rosetta. Consequently, ArcticExpress was not pursued further.

Fig. 5.

Expression and solubility assay using the Arctic Express System (Agilent Technologies) - Protein expression induction with 0.1 mM IPTG. Legend: MM. Molecular Marker (TrueColor High Range); 1. Non-induced pellet; 2. Soluble fraction - 4 h; 3. Insoluble fraction 4 h; 4. Soluble fraction −17 h. 5. Insoluble fraction - 17 h. Red arrow indicating hASNase1 band.

3.4. Strategies for hASNase1 isolation

3.4.1. IMAC purification

Expression condition 3 (Table 1) was selected for large-scale production and purification by IMAC. SDS-PAGE analysis of the elution fraction (Fig. 6A) revealed a prominent band corresponding to hASNase1, accompanied by minor contaminants (45–60 kDa) after elution with 300 mM imidazole. To assess protein stability and integrity after purification, the hASNase1 sample was stored at −80 °C for six months and subsequently analyzed by SDS-PAGE and mass spectrometry (Fig. 6B). The His₆–SUMO–hASNase1 fusion (∼52 kDa), which was initially detected immediately after IMAC (Fig. 6A), was no longer observed after storage. Instead, a new band corresponding to the cleaved hASNase1 (∼40 kDa; Fig. 6B) appeared. Mass spectrometry confirmed that self-cleavage of the fusion tag occurred during storage, indicating that the enzyme remained structurally stable but underwent spontaneous hydrolysis at the SUMO–hASNase1 junction.

Fig. 6.

A) 12.5% SDS-PAGE analysis of samples from hASNase1 affinity chromatography steps under expression condition 2. B) 12.5% SDS-PAGE - Replicates of hASNase1 stored at −80 °C for 6 months following affinity chromatography. Legend: MM. Molecular Marker (TrueColor High Range); 1. Flow-through; 2. 40 mM imidazole wash buffer; 3. 300 mM Imidazole elution buffer; 4. hASNase1 stored at −80 °C; 5. Positive control: Bovine Serum Albumin (BSA); Negative control (gel excision with no protein). Red arrows indicating hASNase1 band.

Proteomic analysis of the soluble fraction (Table S1) detected 15,722 spectra corresponding to 409 proteins: approximately 26% SUMO-tag peptides, 9% GroEL, 7% hASNase1 (63% coverage), and 5% DnaK. While removal of these contaminants is desirable, previous studies have shown that GroEL does not interfere with hASNase1 activity [14]. Therefore, enzyme activity assays were performed using unfractionated IMAC eluates.

3.4.2. Ion exchange chromatography followed by IMAC

Sequential purification combining Ion-Exchange Chromatography (IEX) and IMAC was evaluated to enhance purity. Despite partition across a 0–500 mM NaCl gradient (Figure S5), SDS-PAGE showed no significant enrichment of hASNase1. Although fractions 5–7 displayed asparaginase activity, subsequent IMAC did not improve contaminant removal (Fig. 7). Nearly all hASNase1 was lost, and the remaining band lacked activity. Mass spectrometry identified a 42 kDa contaminant (Fig. 7, marked in red) as an E. coli elongation factor. This protein was absent when purification was performed solely by IMAC, confirming IMAC as a more effective primary step.

Fig. 7.

12.5% SDS-PAGE of hASNase1 Affinity Chromatography: second purification strategy. Legend: MM. Molecular Marker (TrueColor High Range); 1. Flow-through; 2. 40 mM imidazole wash buffer; 3. 300 mM Imidazole elution buffer.

3.4.3. IMAC in the presence of l-Asparagine

Molecular simulations predicted that l-asparagine binding may induce conformational changes in hASNase1, including narrowing of the central cleft, which could be compatible with a more compact and potentially active structural arrangement 15. Guided by this rationale, the soluble fraction from bacterial lysate was incubated with Ni-resin in the presence of 30 mM l-asparagine for either 30 min or 1 h prior to IMAC purification. This approach was explored as a means to favor a more stable conformational state of the enzyme and potentially reduce chaperone association. SDS-PAGE analysis (Fig. 8A) indicated that neither condition effectively removed GroEL, although protein recovery was improved compared to previous protocols. Pooled samples were subsequently treated with SUMO protease (Ulp1) for tag removal and subjected to reverse IMAC (Fig. 8B). Although most contaminants were removed, two high-molecular-weight bands persisted, and the enzyme remained inactive following cleavage, suggesting instability of the mature protein.

Fig. 8.

Strategies to separate chaperones from hASNase1 during chromatography steps. A) Affinity Chromatography with 30 mM asparagine (Asn) at two incubation times: 30 min and 1 hour. B) Affinity Chromatography following His6-SUMO tag cleavage by Ulp1 enzyme. Legend: MM. Molecular Marker (TrueColor High Range); 1. Flow-through - 1 h Asn incubation. 2. 40 mM Imidazole wash buffer - 1 h Asn incubation; 3. 300 mM Imidazole elution buffer - 1 h Asn incubation; 4. Flow-Through - 30 min Asn incubation. 5. 40 mM imidazole wash buffer – 30 min Asn incubation. 6. 300 mM Imidazole elution buffer – 30 min Asn incubation. 7. 300 mM Imidazole elution buffer (30 min + 1 h); 8. Ulp1 enzyme reaction; 9. Flow-Through - Ulp1 enzyme reaction; 10. 300 mM Imidazole elution buffer from Ulp1 enzyme reaction; 11. Ulp1. Red arrow indicating hASNase1 band.

3.4.4. ATP-mediated chaperone removal and optimized IMAC

To displace tightly bound E. coli chaperones, ATP–Mg incubation was initially evaluated as described by Karamitros and Konrad, 2014 [14]. Under these conditions, hASNase1 showed reduced binding to the Ni-resin and was predominantly detected in the flow-through (Figs. 9A–B). Densitometric analysis of SDS-PAGE bands using ImageJ (Figure S6) indicated that hASNase1 accounted for approximately 14% of the total protein in the recovered fraction, while chaperone levels were reduced.

Fig. 9.

Optimization of hASNase1 purification. A) 12.5% SDS-PAGE analysis of Affinity Cromatography with ATP-Mg incubation. B) His-Tagged Recombinant Protein Detection via Western Blot with Alkaline Phosphatase-Conjugated Anti-His Antibody. C) 12.5% SDS-PAGE analysis of Affinity Chromatography. D) His-Tagged Recombinant Protein Detection via Western Blot with Alkaline Phosphatase-Conjugated Anti-His Antibody.

Legend: MM. Molecular Marker (TrueColor High Range); 1. 300 mM Imidazole elution buffer; 2. 40 mM imidazole wash buffer; 3. Flow-Through; 4. Flow-Through; 5. 40 mM imidazole wash buffer; 6. ATP-Mg wash buffer; 7. 300 mM Imidazole elution buffer, 8. Flow-Through. 9. 40 mM imidazole wash buffer; 10. ATP-Mg wash buffer; 11. 300 mM Imidazole elution buffer.

To improve both recovery and purity, a second approach combining l-asparagine incubation with an ATP washing step was evaluated. As shown in Figs. 9A–B/Figure S6, this strategy increased the relative abundance of hASNase1 to approximately 40% of the final preparation, while substantially reducing the presence of GroEL and DnaK. In contrast, ATP washing alone effectively reduced contaminant levels but resulted in significant loss of hASNase1 (Figs. 9C–D), indicating that prior incubation with l-asparagine is critical to preserve enzyme recovery. Together, these results demonstrate that the combination of l-asparagine incubation followed by ATP washing provides the best balance between recovery and reduction of chaperone co-purification under the conditions tested.

The relative abundance of hASNase1 in the lysate and soluble fraction (Figure S7), as well as in samples obtained after ATP incubation alone and after l-asparagine incubation followed by ATP washing (Figure S6), was determined by densitometric analysis. To quantitatively evaluate purification performance, a purification summary is presented in Table 3. Purification using ATP incubation alone - IMAC (ATP) - resulted in a low recovery of 0.09%, despite increasing the relative abundance of hASNase1 to 14.09%. In contrast, the combined strategy of l-asparagine incubation followed by ATP washing - IMAC (Asn + ATP) - improved both purity and recovery, yielding a preparation containing 40.53% hASNase1 with a recovery of 0.29%.

Table 3.

Purification summary of hASNase1. The amount of hASNase1 (mg) was calculated based on its relative abundance. Recovery (%) was calculated relative to the total hASNase1 present in the crude lysate. IMAC (ATP) corresponds to purification with ATP-Mg washing alone, whereas IMAC (Asn + ATP) includes l-asparagine incubation prior to ATP-Mg washing.

Step	Volume (mL)	Prot (mg/mL)	Total (mg)	hASNase1 (%)	hASNase1 (mg)	Recovery (%)
Lysate	25	7.5	188	37.85	71.16	100
Soluble fraction	25	5.1	127.5	3.92	4.99	7.01
IMAC (ATP)	2	0.27	0.54	14.09	0.07	0.09
IMAC (Asn + ATP)	2	0.26	0.52	40.53	0.21	0.29

Steady-state kinetic analysis of hASNase1 using the optimized protocol, including l-asparagine incubation followed by ATP-Mg washing, demonstrated enzymatic activity between 10 mM and 100 mM l-asparagine. The enzyme exhibited typical allosteric behavior, with a Hill coefficient of 3.9, S₀.₅ = 52.1 ± 2.0 mM, and Vmax = 101.7 ± 4.4 U/mg (Fig. 10). Therefore, the reported activity values should be interpreted as apparent specific activity under partially purified conditions. Kinetic parameters were obtained by fitting the experimental data to the Hill equation using nonlinear regression. The resulting fit showed good agreement with the experimental data, supporting the cooperative behavior previously described for this enzyme.

Fig. 10.

Kinetic assay of hASNase1. Specific enzymatic velocity (U/mg) was evaluated as a function of substrate concentration (10–100 mM). All experiments were performed at 37 °C with 0.5 µM of enzyme. Data points represent the means ± SD from triplicate measurements. Steady-state kinetic curves were fitted using the Hill equation with non-linear regression in OriginPro 8.5 software.

3.4.5. Native PAGE and immunodetection of hASNase1

To obtain experimental insight into the oligomeric state of hASNase1 under non-denaturing conditions, native PAGE analysis was performed followed by immunodetection. As shown in Fig. 11, the protein migrated as high-molecular-weight species, with bands observed between ∼171 and ∼460 kDa, suggesting the presence of oligomeric assemblies. Western blot analysis confirmed that these bands correspond to hASNase1-containing species, indicating that the enzyme is present in higher-order complexes under the conditions tested.

Fig. 11.

Native PAGE and Western blot analysis of hASNase1 under non-denaturing conditions. A) Native PAGE of purified hASNase1 showing migration as high-molecular-weight species. B) Immunoblot analysis of the native gel using an anti-His antibody. Legend: 1. Elution obtained from IMAC with ATP incubation alone; 2. Elution obtained from IMAC following l-asparagine incubation combined with ATP washing. MM. Molecular weight marker (HiMark™ Pre-stained Protein Standard, Invitrogen).

Although this approach does not allow precise determination of oligomer stoichiometry, the results provide qualitative evidence consistent with oligomeric organization. These observations are in agreement with previous studies by Karamitros and Konrad, [14], which demonstrated tetrameric assembly using size-exclusion chromatography. Taken together, these data support the presence of higher-order assemblies of hASNase1, while mechanistic interpretations regarding oligomer stabilization remain tentative.

3.5. In silico analysis of hASNase1 structure

The representation of hASNase1 in Fig. 12 reveals that hASNase1 is a homotetramer, consisting of dimers formed by pairs A-D and B-C. Furthermore, the catalytic site of one monomer (Fig. 13) is solely constituted by the interaction of the monomers forming the dimer. The structure of hASNase1 was modeled and thermally equilibrated as described in Guimarães et al., 2021 [15]. The catalytic sites are positioned in the central region of each monomer. It is noteworthy that the allosteric and catalytic sites are in proximity within pairs of monomers, specifically in the A-D and B-C pairs (Fig. 12) [15].

Fig. 12.

Representation of the tetrameric structure of hASNase1. A) A ribbon model is used to represent the tetrameric structure. B) The tetrameric structure is depicted in a sphere format. C) Illustration highlighting the catalytic sites (yellow) and allosteric sites (light blue) represented as spheres. D) Schematic depiction of the tetramer of hASNase1 illustrating the formation of the A-D and B-C dimers. Adapted with permission from Guimarães, A. V. F. Simulações de Dinâmica Molecular de l-Asparaginases: Estudo Comparativo Entre l-Asparaginase Humana, Bacteriana e do Porquinho da Índia. Universidade Federal do Ceará, 2019. Legend: Monomer A: black; Monomer B: red; Monomer C: green; Monomer D: blue; spheres representing carbon atoms in yellow indicate the catalytic sites in each monomer; spheres representing carbon atoms in light blue denote the allosteric sites in each monomer; CS: catalytic site; AS: allosteric site.

Fig. 13.

Representation of the catalytic site of hASNase1. Adapted with permission from Guimarães, A. V. F. Simulações de Dinâmica Molecular de l-Asparaginases: Estudo Comparativo Entre l-Asparaginase Humana, Bacteriana e do Porquinho da Índia. Universidade Federal do Ceará, 2019. Legend: Monomer A: black; Monomer D: blue. Residues of the catalytic site in monomer D are depicted in stick format with carbon atoms colored yellow. Allosteric site residues can be visualized in the surroundings colored in cyan.

The AGGRESCAN program revealed many regions with high aggregation propensity in hASNase1, which may contribute to the formation of inclusion bodies. In the AGGRESCAN results, the regions of the sequence with the highest predicted tendency for aggregation are highlighted in red (Fig. 14) [31]. According to the analysis, amino acid residues with a score above 0 (indicated in red – Fig. 14) exhibit aggregation propensity, whereas negative scores indicate solubility (Table S2 and Figure S8).

Fig. 14.

Prediction of regions with high aggregation propensity in hASNase1 using the AGGRESCAN program. A) front view. B) rotated 90 degrees. C) rotated 180 degrees. D) rotated 270 degrees. Legend: Regions highlighted in red have a positive score, indicating the highest predicted tendency for aggregation.

The association between predicted aggregation hotspots and the observed experimental behavior is primarily correlative, since aggregation prediction alone does not establish a direct causal relationship between specific residues and aggregation in the absence of experimental validation. Given this limitation, the effects of these residues are currently being evaluated in a separate study. Based on the aggregation-prone regions identified, selected residues were substituted with counterparts from Cavia porcellus l-asparaginase, previously associated with increased solubility, as part of a related project within our group. The resulting soluble variants are currently under evaluation, and their detailed characterization will be presented in a forthcoming publication. These findings illustrate how identifying aggregation-prone regions can guide protein engineering strategies to improve solubility and stability.

4. Discussion

Interest in human-derived l-asparaginases has grown over the past decade, particularly in the context of their potential to overcome limitations associated with bacterial enzymes [6,14,16,43,44]. However, beyond their well-known clinical application, l-asparaginases also play broader roles in metabolism and have been explored in diverse biotechnological contexts. For instance, these enzymes catalyze the hydrolysis of l-asparagine into aspartate and ammonia, a reaction that is relevant not only for cellular nitrogen metabolism but also for industrial applications, such as the reduction of acrylamide formation in food processing [6,45].

In the present study, it was aimed to establish a scalable method for producing active wild-type hASNase1 in E. coli while overcoming known challenges related to aggregation and chaperone co-purification, limitations that remain under-explored in the literature. Initial expression attempts resulted in extensive inclusion body formation, consistent with other reports of heterologous human protein expression in prokaryotic hosts [40,46]. To better understand the influence of the host expression system on hASNase1 solubility, three E. coli strains with distinct characteristics were evaluated. BL21(DE3)-Star, optimized for high-level expression, produced large amounts of protein that accumulated in inclusion bodies. The Rosetta strain, which provides tRNAs for rare codons corresponding to eukaryotic genes, yielded a small soluble fraction when expression was induced at low temperature (15 °C). In contrast, Arctic Express, engineered to co-express cold-adapted chaperonins, did not enhance solubility and showed reduced total expression.

Expression under milder conditions (lower temperature and IPTG concentrations) led to detectable levels of soluble protein in E. coli Rosetta, though a large fraction remained insoluble (Fig. 4). Strategies to extract active enzyme from inclusion bodies using low urea concentrations (≤ 2 M) allowed solubilization without denaturation; however, refolding attempts yielded catalytically inactive hASNase1, in line with observations by Karamitros and Konrad, 2014 [14] and Upadhyay et al., 2014 [40]. The lack of successful refolding likely reflects the structural complexity of hASNase1. Proper tetramer assembly may not have been achieved under the conditions tested, which would compromise enzymatic activity. It is also possible that non-native disulfide bonds [47] or kinetically trapped intermediates formed during the process, preventing correct folding [48]. The limited success of detergent and arginine-based buffers compared with urea-containing solutions suggests that hASNase1 aggregates display strong hydrophobic interactions that require chaotropic agents to be disrupted.

Mass spectrometry analysis revealed substantial co-purification of hASNase1 with chaperonins, particularly GroEL and DnaK. This interaction is commonly observed during overexpression of aggregation-prone eukaryotic proteins in E. coli [[22], [23], [24]]. While these chaperones facilitate folding, their strong affinity can hinder dissociation, complicating downstream purification and posing potential immunogenic risks if not removed [49]. Our results confirmed the frequent GroEL association with hASNase1, which is likely promoted by solvent-exposed hydrophobic regions identified in silico (Fig. 13). This may also hinder the formation of the active oligomeric structure, as dimerization and tetramerization of hASNase1 are critical for catalytic activity [14,15], stabilized in part by conserved Tyr residues previously reported in other asparaginases [[50], [51], [52], [53], [54]].

Native PAGE analysis further supported the presence of higher-order assemblies of hASNase1 under non-denaturing conditions. Although this approach does not allow precise determination of oligomer stoichiometry, these observations are consistent with previous SEC-based studies [[14], [15]] and support the presence of oligomeric forms in solution.

A key novelty of this work is the evaluation of small-molecule additives to dissociate hASNase1–chaperone complexes while favoring a more stable and potentially active conformational state. Incubation with l-asparagine during IMAC has previously been associated with stabilization of hASNase1 oligomeric forms [14]; here, we demonstrate that its combination with an ATP-wash significantly enhances recovery of active hASNase1 and reduces contaminant chaperones (Fig. 9). Importantly, this lower-cost method is less laborious than multi-step IEX/IMAC protocols, which caused substantial target protein loss (Figure S5).

Although the present study focuses on hASNase1, the principle underlying the ATP-mediated washing step is based on the well-established ATP-dependent dissociation mechanism of molecular chaperones such as GroEL and DnaK. These chaperone systems interact with a broad range of aggregation-prone or partially folded polypeptides during heterologous expression in E. coli, particularly in the case of complex or multidomain eukaryotic proteins. Therefore, similar strategies may potentially be applicable to other aggregation-prone recombinant proteins expressed in E. coli [24,49,55]. However, the efficiency of this approach is likely to depend on the specific folding pathway and oligomerization requirements of each target protein.

In line with this optimization strategy, the purification workflow enabled the recovery of enzymatically active hASNase1, allowing its subsequent kinetic characterization. The enzyme obtained from optimized purification protocol showed typical positive cooperative behavior, with a Hill coefficient of 3.9 and S₀.₅ of 52.1 mM, reinforcing existing evidence that hASNase1 is allosterically regulated and exhibits maximal activity only at high l-asparagine concentrations [[14], [15], [16],44,54]. The relatively high S0.5 observed here is consistent with previous reports describing hASNase1 as an allosteric enzyme with low substrate affinity compared with bacterial l-asparaginases [6,14]. It is worth noting that residual contaminants were still present in the final preparation; however, previous studies have shown that the association of chaperones such as GroEL does not significantly affect the enzymatic activity of hASNase1, as comparable kinetic parameters were observed in their presence or absence [14].

This optimized purification protocol allows the acceleration of screening studies of hASNase mutants aiming to enhance some enzymatic properties, such as catalytic activity and solubility. This way, new promising molecules can be directed to expression platforms more suitable for in vivo studies and clinical formulation development.

In conclusion, the intrinsic tendency of hASNase1 to aggregate and associate with E. coli chaperones presents a major barrier to its heterologous production. Nonetheless, by applying optimized expression conditions and a novel l-asparagine/ATP-based IMAC protocol, we succeeded in isolating catalytically active hASNase1 with reduced chaperone contamination and without excessive use of costly or harsh reagents. This strategy provides a practical platform for future functional studies and protein engineering efforts aimed at improving the catalytic properties and solubility of hASNase1. More broadly, this work establishes a framework for enhancing the recovery of aggregation-prone human enzymes expressed in E. coli, particularly in the context of chaperone-associated purification challenges.

CRediT authorship contribution statement

Maísa Pessoa Pinheiro: Writing – review & editing, Writing – original draft, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Ana Virgínia Frota Guimarães: Writing – review & editing, Writing – original draft, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Lucas Almeida de Freitas: Writing – review & editing, Methodology, Formal analysis, Data curation. Gilvan Pessoa Furtado: Writing – review & editing, Supervision, Project administration, Data curation. Marcos Roberto Lourenzoni: Writing – review & editing, Supervision, Project administration, Funding acquisition, Data curation, Conceptualization. Marco Alberto Medeiros: Writing – review & editing, Supervision, Project administration, Funding acquisition, Data curation, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

Data will be made available on request.