Method for Preparing Recombinant Galectin-2 Protein without Escherichia coli-Specific Post-translational Modifications

Abstract

Galectin-2 (Gal-2) is an animal lectin with specificity for β-galactosides. It is predominantly expressed and suggested to play a protective function in the gastrointestinal tract; therefore, it can be used as a protein drug. Recombinant proteins have been expressed using Escherichia coli and used to study the function of Gal-2. The recombinant human Gal-2 (hGal-2) protein purified via affinity chromatography after being expressed in E. coli was not completely homogeneous. Mass spectrometry confirmed that some recombinant Gal-2 were phosphogluconoylated. In contrast, the recombinant mouse Gal-2 (mGal-2) protein purified using affinity chromatography after being expressed in E. coli contained a different form of Gal-2 with a larger molecular weight. This was due to mistranslating the original mGal-2 stop codon TGA to tryptophan (TGG). In this report, to obtain a homogeneous Gal-2 protein for further studies, we attempted the following methods: for hGal-2, 1) replacement of the lysine (Lys) residues, which was easily phosphogluconoylated with arginine (Arg) residues, and 2) addition of histidine (His)-tag on the N-terminus of the recombinant protein and cleavage with protease after expression; for mGal-2, 3) changing the stop codon from TGA to TAA, which is commonly used in E. coli. We obtained an almost homogeneous recombinant Gal-2 protein (human and mouse). These results have important implications for using Gal-2 as a protein drug.

INTRODUCTION

Galectins are animal lectins with an affinity for β-galactosides.^1–6) Galectin-2 (Gal-2) is primarily expressed in mucosal epithelial cells of the stomach^7–9) and is thought to be involved in gastric protection.¹⁰⁾ Gal-2 was a major substrate in a mouse stomach screen analysis of S-nitrosylated substrates.¹¹⁾ Gal-2 has two cysteine (Cys) residues highly conserved across mammals (Cys⁵⁷ and Cys⁷⁵ for human and mouse Gal-2). In our previous studies with recombinant Gal-2, we revealed that Gal-2 lost its sugar-binding activity when treated with hydrogen peroxide, the sugar-binding activity was maintained after hydrogen peroxide treatment if Gal-2 was S-nitrosylated beforehand, and S-nitrosylation did not affect the biochemical properties of Gal-2, such as the sugar-binding properties and dimer formation.¹²⁾ LC-tandem mass spectrometry (LC/MS/MS) analysis revealed that Cys⁵⁷ and Cys⁷⁵ were S-nitrosylated, which was maintained after hydrogen peroxide treatment. Nitric oxide (NO) binding to Cys residues suggests protection from oxidative inactivation by hydrogen peroxide. In the same study, we revealed that using a higher-order structural approach, S-nitrosylation stabilizes the structure of Gal-2.¹³⁾ We further examined how oxidation of both Cys residues is involved in the sugar-binding activity loss of Gal-2 using recombinant proteins with a site-directed mutation in both Cys residues (C57M and C75S). The C75S mutant (Cys⁵⁷ remains the only Cys residue) lost its sugar-binding activity via hydrogen peroxide treatment as with the wild type (WT), while pre-S-nitrosylation retained its carbohydrate-binding activity after hydrogen peroxide treatment. In contrast, C57M (Cys⁷⁵ remains the only Cys residue) retained its carbohydrate-binding activity after hydrogen peroxide treatment. These results indicate that Cys⁵⁷ is involved in the oxidative inactivation of Gal-2 and that S-nitrosylation of this residue is essential for protection against the loss of sugar-binding activity via oxidation.¹⁴⁾

Gal-2 is abundantly expressed in gastric mucosal epithelial cells. Therefore, we hypothesized that it interacts with mucin, a major mucus component, and investigated the interaction between Gal-2 and mucin from porcine stomach. Recombinant Gal-2 and gastric mucin were suggested to form an insoluble complex, possibly a matrix complex, inhibited by the competing sugar lactose. This suggests that Gal-2 cross-links mucin via its sugar chains.¹⁵⁾ In addition, potential Gal-2 ligands were explored from mouse gastric mucus fractions. Gastric mucus fractions collected from mice were solubilized in the presence of a reducing agent and added to a Gal-2-immobilized column. The fraction bound to the column and eluted using lactose was analyzed via LC/MS/MS, and MUC5AC, a major mucin of gastric mucus, was identified. Gal-2 was detected in the mouse gastric mucous fraction and was eluted via lactose addition. In contrast, MUC5AC remained in the insoluble fraction despite lactose treatment. Recombinant Gal-2 was added to the insoluble fraction after incubating the mouse gastric mucus fraction with lactose and was bound to the insoluble gastric mucus fraction. This binding was lost due to lactose. Thus, the results suggested that Gal-2 is bound to MUC5AC in the insoluble mucus fraction in a carbohydrate-dependent manner, resulting in a stronger barrier structure protecting the mucosal surface.¹⁶⁾

In this study, we observed that human Gal-2 (hGal-2) protein expressed in Escherichia coli contained some modified hGal-2. This possible modification was predicted to be phosphogluconoylated based on the mass difference of the two forms of hGal-2 observed after protein purification using anion exchange column chromatography and MS analysis. Obtaining a more purified recombinant hGal-2 expressed in E. coli with no different forms should be useful for further analysis of the Gal-2 protein and its potential as a protein drug. Considering the use of Gal-2 as a protein drug, studies have already reported that adding polyethylene glycol (PEG) to Gal-2 to improve solubilization and stabilization extends the half-life of Gal-2.¹⁷⁾ In this study, we replaced the lysine (Lys) residues, which were easily phosphogluconoylated with arginine (Arg) residues. Furthermore, since the N-terminus of the recombinant protein with histidine (His)-tag expressed in E. coli tend to be phosphogluconoylated,^18,19) we added His-tag to the N-terminus of hGal-2, aiming to reduce phosphogluconoylation.

In mouse Gal-2 (mGal-2), when the recombinant protein was expressed in E. coli and purified using affinity chromatography, a different Gal-2 protein form with a higher molecular weight was expressed and eluted from the affinity chromatography. We investigated the nature of this different form of Gal-2 using MS analysis of the purified recombinant proteins.

Homogenous Gal-2 is necessary for a more detailed analysis of Gal-2 and for future plans of using Gal-2 as a pharmaceutical product, such as providing a stronger barrier structure protecting the mucosal surface in the stomach. In order to develop Gal-2 as a protein drug for humans, human galectin-2 is needed. However, animal experiments should be conducted first. As the mouse is the most appropriate animal model for a detailed analysis of the protective effect of Gal-2 in the stomach, both human and mouse Gal-2 were prepared and used for experiments. The amino acid identity between human and mouse Gal-2 is approximately 66%, and the two cysteine residues, Cys⁵⁷ and Cys⁷⁵, are conserved. Since there is no significant difference in sugar-binding properties and dimer formation, the functions of mouse Gal-2 seem to be mostly similar to those of human Gal-2.

Uniform recombinant Gal-2 protein could not be obtained via the general recombinant protein preparation using E. coli, including purification using affinity chromatography after expressing E. coli. Therefore, we explored methods for obtaining homogenous recombinant Gal-2 proteins using E. coli without E. coli-specific post-translational modifications.

MATERIALS AND METHODS

Construction of Wild-Type Human and Mouse Gal-2 Expression Plasmids

First-strand cDNA was synthesized from human small intestine total RNA (Ambion Inc., Austin, TX, U.S.A.). The DNA fragment encoding hGal-2 was amplified via PCR using first-strand cDNA as a template. Forward and reverse primer sequences containing NdeI and BamHI restriction sites (underlined) were 5′-ATTACATATGACGGGGGAACTTGAGGTT-3′,5′-CTAGGGATCCGCTGGAAGTCTTTTATTC-3′, respectively. The DNA fragment encoding mGal-2 was amplified via PCR using first-strand cDNA (GenoStaff, Co., Ltd., Tokyo, Japan) as a template. Forward and reverse primer sequences were 5′-ATTACATATGTCGGAGAAATTTGAGGTC-3,′5′-CTAGGGATCCGGCTAAGGTCTTCTGAGG-3′, (NdeI and BamHI sites were underlined), respectively. The PCR products were digested with NdeI and BamHI and inserted into the corresponding restriction site of the pET21a vector (Novagen, Merck KGaA, Darmstadt, Germany).

Construction of Mutant Gal-2 Expression Plasmids

The mutation was introduced via PCR using hGal-2 WT-pET21a (for K129R and K131R mutants), hGal-2 K131R-pET21a (for the K129/131R mutant), or mGal-2 WT-pET21a (for the mGal-2 *TAA mutant) as a template using the following primers according to PrimeSTAR max manual (TaKaRa Bio Inc., Shiga, Japan). hGal-2 K129R: 5′-TCTTTCCGCTTAAAAGAATAAAAGACTGG-3′,5′-TTTTAAGCGGAAAGAGGACATGTTGAACC-3′, hGal-2 K131R: 5′-AAGTTACGTGAATAAAAGACTGGATCC-3′,5′-TTATTCACGTAACTTGAAAGAGGACATG-3′, hGal-2 K129/131R: 5′-TCTTTCCGCTTACGTGAATAAAAGACTGGATCC-3′, 5′-TTATTCACGTAAGCGGAAAGAGGACATGTTGAAC-3′, mGal-2 TAA: 5′-ACTGGAATAAGCGGCACCTCAGAAGACC-3′, 5′-TGCCGCTTATTCCAGTTTGAAGGAGGAG-3′.

His-tag-hGal-2 DNA was first prepared via PCR using hGal-2 WT-pET as a template and primers (5′-TATCATATGGGCCATCATCATCATCATCATAGCAGCGGCATGACGGGG-3′,5′-CTAGGGATCCGCTGGAAGTCTTTTATTC-3′). The PCR product was then cleaved with NdeI and BamHI and ligated to pET21a, which was treated using the same restriction enzymes. The human rhinovirus 3C (HRV 3C) protease recognition sequence was inserted between His-tag and hGal-2 using this plasmid as a template via PCR using the following primers: 5′-TTCCAGGGTCCGATGACGGGTGAAC-3′,5′-CAGAACTTCCAGGCCGCTGCTATGATG-3′.

Extraction and Purification of Recombinant Human and Mouse Gal-2

Recombinant human and mouse Gal-2 were expressed and purified as previously described.^20,21) Briefly, E. coli BL21(DE3) were transformed with Gal-2 expression plasmids created above. The E. coli were pre-cultured and grown at 37 °C in an LB medium containing 125 µg/mL ampicillin until the optical density at 600 nm was 0.6–0.8. Protein expression was induced by adding isopropyl-β-D-thiogalactopyranoside at a final concentration of 0.4 mM. After overnight incubation at 20 °C, E. coli cells were collected and disrupted via sonication. Gal-2 expressed in the soluble fraction of the cell lysate was purified using lactose-immobilized resin (Sigma-Aldrich, St. Louis, MO, U.S.A). The proteins eluted in fractions from the column were separated using sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) and stained with Coomassie brilliant blue (CBB).

Ion Exchange Column Chromatography

The purified recombinant hGal-2 WT buffer was exchanged via ultrafiltration using Amicon Ultra-15 (Merck KGaA) and 50 mM Tris–HCl, pH 8.0, containing 1 mM dithiothreitol (DTT). The hGal-2 WT protein solution was then applied to a RESOURCE™ Q column (1 mL) (Cytiva, Tokyo, Japan), equilibrated with 50 mM Tris–HCl, pH 8.0, containing 1 mM DTT, and eluted with a liner gradient of 0–1.0 M NaCl in 20 column volumes. Fraction volumes were 1 mL. Ion exchange column chromatography was performed using the AKTA™ purifier system.

Trypsin Digestion

Approximately 105 µg of hGal-2 WT was digested with trypsin using trypsin/Lys-C mix (Promega Corporation, Madison, WI, U.S.A.). The samples were denatured using 8 M Urea and 5 mM DTT for 1 h at 37 °C, alkylated by adding 14 mM iodoacetamide for 1 h at room temperature in the dark. The iodoacetamide was inactivated by adding 9 mM DTT for 1 h at room temperature in the dark. After diluting the samples with 50 mM ammonium bicarbonate to make the concentration of Urea to <1 M, 4 µg trypsin/Lys-C mix was added, and the solution was incubated overnight at 37 °C.

LC/MS/MS

LC/MS/MS analyses were performed using an Ultimate 3000 HPLC system coupled online to a Q Exactive Orbitrap mass spectrometer (Thermo Fisher Scientific, Waltham, MA, U.S.A). The proteins were separated via reversed phased HPLC (RPHPLC) using a C4 column (Proteonavi, 2.0 × 10 mm, Osaka Soda, Osaka, Japan) with a 10-min linear gradient of 2–90% acetonitrile in the presence of 0.1% formic acid at 0.2 mL/min. The peptides were separated via RPHPLC using a C18 column (CAPCELL PAK C18 ACR, 1.0 × 150 mm, Shiseido, Tokyo, Japan) with a 60-min linear gradient of 2–45% acetonitrile in the presence of 0.1% formic acid at 0.1 mL/min.

Phos-Tag SDS-PAGE

Phos-tag^® SDS-PAGE, which can separate phosphorylated and non-phosphorylated proteins, separated modified from unmodified Gal-2. Phos-tag SDS-PAGE was performed on 15% polyacrylamide gels containing 100 µM Phos-tag acrylamide (FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan) according to the manufacturer’s instruction. The gel was stained with CBB.

HRV 3C Protease Digestion

His-tag-hGal-2 was dissolved in phosphate buffered saline (PBS) for a concentration of 1 mg/mL, HRV 3C protease (TaKaRa Bio Inc.) was added to reach a concentration of 0.01 U/µL, and His-tag-hGal-2 was digested at 4 °C overnight.

RESULTS

The recombinant Gal-2 protein expressed in E. coli and purified using affinity chromatography has resulted in single bands when subjected to SDS-PAGE. However, during the structural analysis of Gal-2, the possibility of modifying recombinant Gal-2 emerged. Therefore, we analyzed the modification in detail in this study.

Recombinant proteins in both peaks eluted using the ion exchange column chromatography—peak 1 (elution volume 4–5 mL) and peak 2 (elution volume 7–8 mL) in Fig. 1—were analyzed using LC/MS/MS. Mass spectra of hGal-2 and hGal-2 + 258 were observed for each peak. The ratios of the mass spectrum intensity for both recombinant proteins (hGal-2 : hGal-2 + 258) were 1 : 1 for peak 1 and 30 : 1 for peak 2 (30 : 1) (data not shown). The protein post-translational modifications Delta Mass database (Association of Biomolecular Resources Facilities, https://www.abrf.org/delta-mass) analysis suggested the possibility of phosphogluconoylation (+258.014) of hGal-2 + 258. Therefore, recombinant hGal-2 WT prepared using E. coli was trypsin-digested and subjected to LC/MS/MS to find the modified peptide fragment. A peak with a +258 mass for the trypsin-digested hGal-2 peptides was searched, and a signal for a peptide with [M + H]⁺ = 1602.67 was observed, possibly from ¹²¹GGFNMSSFKLKE¹³² phosphogluconoylation (Fig. 2A). This peak was further analyzed using MS/MS. The results of the b and y ions detected suggest that Lys¹²⁹ or Lys¹³¹ could be modified (mass +258) (Figs. 2B, C).

Fig. 1. The Ion Exchange Column Chromatogram of Recombinant Wild-Type hGal-2 Protein Purified via Affinity Chromatography

The hGal-2 protein—purified using affinity chromatography following E. coli expression—was subjected to anion exchange column chromatography using RESOURCE™ Q column. The solid and dotted lines indicate 280 nm absorbance and NaCl concentration, respectively. The volume for each fraction was 1 mL. Elution volumes of 4–5 and 7–8 mL were collected as Peaks 1 and 2, respectively.

Fig. 2. Identification of Modification Sites for Phosphogluconoylation (+258)

Mass chromatogram of (A) possible phosphogluconoylated peptide—¹²¹GGFNMSSFKLKE¹³² + 258 ([M + H]⁺ = 1602.67)—obtained via trypsin treatment of recombinant hGal-2. (B)(C) MS/MS spectrum of the phosphogluconoylated peptide (¹²¹GGFNMSSFKLKE¹³² + 258). (B) Lys¹²⁹ was identified as the phosphogluconoylation site using the theoretical b and y ions. (C) Lys¹³¹ was identified as the phosphogluconoylation site with theoretical values of b and y ions.

To determine whether Lys¹²⁹ or Lys¹³¹ is more likely to be phosphogluconoylated, we generated the protein mutants—hGal-2 K129R, K131R, and K129/131R. One or both Lys (Lys¹²⁹ and Lys¹³¹) near the C-terminus of hGal-2 were replaced with Arg and compared with the post-translational modifications of the hGal-2 WT expressed in E. coli. First, DNA for each mutant was prepared, inserted into an expression plasmid, and E. coli was transformed with the expression plasmids. Using this E. coli, each Lys-substituted mutant protein was expressed as described previously and purified using affinity chromatography. We specifically eluted each mutant bound to the lactose-immobilized column using lactose. SDS-PAGE revealed that each purified hGal-2 mutant was a single band of approximately 14 kDa (Fig. 3A). In addition, the results of Arg-mutated recombinant hGal-2 hemagglutination were similar to those of wild-type hGal-2 (data not shown). These results suggest that the effect of these Arg-mutations was minimal on sugar-binding activity.

Fig. 3. hGal-2 Mutants Were Separated Using SDS-PAGE and Phos-Tag SDS-PAGE; Their Molecular Weights Were Measured via Mass Spectrometry

The recombinant hGal-2 mutants were separated using (A) SDS-PAGE and (B) Phos-tag SDS-PAGE. The black arrowheads indicate unmodified recombinant hGal-2. White arrowheads indicate modified hGal-2 bands. (C) MS spectra of recombinant hGal-2 proteins WT, K129R, K131R, and K129/131R are presented. All illustrate the spectra for the 10⁺ charge state. Peaks labeled “ + PGlc” indicate each recombinant hGal-2 with modified phosphogluconoylation. The single asterisk (*) indicates the unmodified form of each recombinant hGal-2 in which one Met (presumably the N-terminal Met) is lost. The double asterisk (**) indicates each recombinant hGal-2 protein in which the one Met is lost and phosphogluconoylated (+PGlc).

Next, we attempted to separate post-translationally modified and unmodified hGal-2 proteins via Phos-tag SDS-PAGE, which can separate phosphorylated proteins. Consequently, two bands with smaller mobility were detected in addition to the main band for hGal-2 WT and K131R, and one band for K129R. However, these additional bands were substantially less for the K129/131R mutant (Fig. 3B).

The phosphogluconoyl modification was analyzed using mass spectrometry. The peaks of hGal-2 + 258 (phosphogluconoylation) were detected in each mutant (Fig. 3C).

In addition, each mutant was trypsin-digested and subjected to LC/MS/MS analysis. The results revealed that for hGal-2 WT, phosphogluconoylation was detected at one location in the peptide fragment—¹²¹GGFNMSSFKLKE¹³². For the K129R mutant, the corresponding peptide fragment—¹²¹GGFNMSSFRLKE¹³²—and the peptide fragment—¹²¹GGFNMSSFKLRE¹³²—were detected with single sites subjected to phosphogluconoylation. In K129/131R, no phosphogluconoylation peak was detected in the corresponding peptide fragment—¹²¹GGFNMSSFRLRE¹³². Furthermore, the K129/131R mutant was mainly unmodified, with minimal phosphogluconoylation detected at positions other than K129 and K131.

Phosphogluconoylation of the His-tag on recombinant proteins was reported.^18,19) Therefore, we added His-tag to the N-terminus of hGal-2, aiming to reduce phosphogluconoylation. Furthermore, hGal-2 WT, with or without His-tag, was expressed and purified as previously described and subjected to SDS-PAGE and Phos-tag SDS-PAGE, followed by CBB staining (Fig. 4). The normal SDS-PAGE revealed that purified hGal-2, with or without His-tag mutant, was a single band (Fig. 4A). Phos-tag SDS-PAGE revealed that although multiple bands with low mobility were detected in hGal-2 without His-tag (Fig. 4B, lane 1), low-mobility proteins were substantially smaller for hGal-2 with His-tag (Fig. 4B, lane 2).

Fig. 4. Comparison of hGal-2 WT and His-Tag-hGal-2 Modifications

HRV 3C protease-treated hGal-2 WT, His-tag-hGal-2, and HRV 3C protease-treated His-tag-hGal-2 were separated using (A) SDS-PAGE and (B) Phos-tag SDS-PAGE. The black arrowheads and white arrowheads and arrows indicate bands of unmodified hGal-2, modified hGal-2, and HRV 3C protease (22 kDa) used to cleave His-tag, respectively.

hGal-2 with His-tag on the N-terminus was treated with HRV 3C protease to cleave the His-tag. The additional band with possible phosphogluconoylation was virtually undetectable for the hGal-2 WT with the added His-tag cleaved with HRV 3C treatment. (Fig. 4B, lane 3). Therefore, His-tag on the N-terminus seemed to have absorbed most phosphogluconoylation modification.

Since hGal-2 was subjected to modification when expressed in E. coli, we assessed mGal-2 for similar modifications. mGal-2 WT was expressed in E. coli similarly to hGal-2, and the results of SDS-PAGE and CBB staining of fractions purified using lactose-immobilized columns are presented in Fig. 5. A major and minor band of approximately 14  and 18 kDa, respectively, were detected in the lactose-eluted fractions. The protein with a minor band of approximately 18 kDa was subjected to trypsin-digestion and LC/MS/MS analysis. The results suggested that this protein resulted from translation not stopped at the original but rather at the next stop codon TGA. A peptide sequence with a tryptophan residue (codon: TGG) inserted at the supposed termination site during translation and with extended translation toward the next possible stop codon was detected. According to the Kazusa DNA research institute codon usage database,²²⁾ the stop codon usage rates of the E. coli B strain are 0.64 for TAA, 0.27 for TGA, and 0.09 for TAG. Therefore, a mutant—mGal-2 TAA cDNA—was generated by changing the stop codon of mGal-2 from TGA to TAA, which is frequently used in E. coli, and expressed in E. coli similarly to the hGal-2 mutant. SDS-PAGE of the E. coli extract fraction purified on a lactose-immobilized column had a single band of approximately 14 kDa in the lactose-eluting fraction.

Fig. 5. SDS-PAGE Analysis of Recombinant mGal-2 WT and mGal-2 TAA Proteins Specifically Eluted from the Lactose-Immobilized Column

Recombinant proteins were expressed in E. coli and purified using affinity chromatography on lactose-immobilized columns. Proteins were separated by reducing SDS-PAGE, and the gels were stained using CBB. The molecular masses of marker proteins are indicated to the left of the panel. The black and white arrowheads indicate the recombinant mGal-2 WT and mGal-2 positions with a larger molecular weight, presumably resulting from extended translation toward the next possible stop codon, respectively.

DISCUSSION

Gal-2 is expressed mainly in the gastrointestinal tract and is thought to be associated with protecting the gastrointestinal tract. Gal-2 has two Cys residues that are highly conserved among mammals. Our previous studies with recombinant Gal-2 revealed that Cys⁵⁷ is responsible for oxidative inactivation of the sugar-binding activity of Gal-2, and S-nitrosylation of the Cys⁵⁷ is crucial to protection from oxidative loss of sugar-binding activity.^12,14) Gal-2 is abundantly expressed in gastric mucosal epithelial cells. Thus, it may interact with mucin, a major mucus component. We identified mucin MUC5AC, a major gastric mucus component, as a potential Gal-2 ligand in gastric mucus.^15,16)

The recombinant hGal-2 protein separated using the Mono Q™ column was used for X-ray crystallographic analysis. Since the Mono Q™ column was no longer available to us, we purified recombinant hGal-2 WT using a similar anion-exchange column, Resource™ Q. The protein was also separated into few peaks (Fig. 1). When the mass of each peak protein was determined using LC/MS/MS, the eluted peak 1 contained more hGal-2 + 258 (data not shown) as was the case for the recombinant protein expressed previously and separated using the Mono Q™ column. The Protein post-translational modifications Delta Mass database (Association of Biomolecular Resources Facilities, https://www.abrf.org/delta-mass) results suggested the possibility of phosphogluconoylation (+258.014). Phos-tag SDS-PAGE also separated the modified and unmodified forms, suggesting that at least a phosphate group was added.

A homogeneous Gal-2 protein is needed for more detailed biochemical studies of Gal-2. Hence, we attempted to prepare un-phosphogluconoylated hGal-2. Little is known about post-translational modifications to recombinant proteins expressed in E. coli. Nonetheless, when His-tag was added to the N-terminus of recombinant proteins expressed in E. coli, the N-terminal amino group was phosphogluconoylated. Phosphogluconoylation to protein body (not tag) has also been reported.²³⁾ E. coli BL21(DE3)—used for expressing recombinant proteins—lacks phophogluconolactonase of the pentose phosphate circuit; therefore, 6-phosphogluconolactone is thought to accumulate in the bacteria. In addition, excess accumulated 6-phosphogluconolactone results in phosphogluconoylation of the expressed recombinant proteins in a non-enzymatic manner. Phosphogluconoylation of the His-tag reportedly affects protein crystallization²⁴⁾; thus, recombinant protein preparation using E. coli expressing 6-phosphogluconolactonase to prevent phosphogluconoylation has been reported.²³⁾ In this study, we revealed that replacing the residues highly susceptible to phosphogluconoylation, such as Lys¹²⁹ and Lys¹³¹ residues, with Arg, adding His-tag to the N-terminus, and cleaving the tag with a protease after expression considerably reduced the phosphogluconoylation. This approach may be useful for preparing uniform Gal-2 protein in E. coli.

Furthermore, phosphogluconoyl His-tag-Gal-2 bands (marked with a white arrowhead in the second lane from the right in Fig. 4B) became fainter when the HRV 3C protease recognition sequence was inserted between His-tag and Gal-2 than those of hGal-2 WT (leftmost lane in Fig. 4B). When only His-tag was added to the Gal-2 protein, the same amount of phosphogluconoylated modification was detected as in hGal-2 WT (data not shown). Although the exact reason for this is unknown, adding the HRV 3C protease recognition sequence possibly resulted in a weaker affinity towards the phos-tag in some way, preventing the separation of the phosphogluconoyl-enriched product. Alternatively, the insertion of the HRV 3C protease recognition sequence may have prevented phosphogluconoylation itself.

When recombinant mGal-2 was expressed in E. coli, substantial recombinant proteins with a higher molecular weight were specifically bound to the affinity column and eluted specifically with lactose addition. We uncovered that this was due to the mistranslation of the original mGal-2 stop codon as tryptophan using MS analysis. Lu et al. reported a similar phenomenon when recombinant human platelet-derived growth factor was expressed in E. coli.²⁵⁾ Via site-directed mutagenesis of the stop codon from TGA to TAA, commonly used in E. coli, mGal-2 with a longer C-terminus was no longer observed. Contrary, such a high molecular weight protein was not observed in hGal-2 expression. We assume this is likely because the stop codon for hGal-2 was TAA, the same one frequently used in E. coli.

Recombinant Gal-2 with or without possible phophogluconoylation (+258) or Gal-2 with extended translation was retained via lactose-immobilized resin and specifically eluted using lactose. This suggests that these modifications do not result in a substantial difference in galectin activity. The expression levels in E. coli of the recombinant Gal-2 with Lys to Arg substitution mutation and His-tag addition (using 500 mL E. coli culture) were nearly 10 mg for K129R and K131R mutants and more than 14 mg for wild-type Gal-2 with added His-tag, while they were slightly lower for the K129/131R mutant at 4 mg. Since the K129/131R double mutated protein seemed to be slightly less stable than the other recombinant proteins, further improvement may be necessary.

Phosphogluconoyl modifications on mGal-2 were also detected by LC/MS and Phos-tag SDS-PAGE (data not shown). Since the Lys residue corresponding to Lys¹²⁹ in mice and humans is also conserved in other animals such as rats, wolves, boars, dogs, and bovines, and the amino acid sequence near this Lys residue are homologous to each other, Gal-2 in these animals (other than human and mouse) may also be phosphogluconoylated, and the methods reported in this paper to prevent phophogluconoylation could be applied to the Gal-2′s from those animals as well.

Thus, the methods used in this study will allow the expression of purer and homogenous Gal-2 with substantially less chemical modification for a more detailed Gal-2 analysis. Furthermore, the resulting Gal-2 could be more useful and safer as a biopharmaceutical and for other applications.

Acknowledgments

The authors thank Ms. Tomoko Yamashita for her technical assistance.

This work was partially supported by BINDS from AMED (JP21am0101083 and JP22ama121001). Research Grants from Teikyo University and Josai University partially supported this work.

Conflict of Interest

The authors declare no conflict of interest.

REFERENCES

• 1) Kasai K, Hirabayashi J. Galectins: a family of animal lectins that decipher glycocodes. J. Biochem., 119, 1–8 (1996).
• 2) Leffler H, Carlsson S, Hedlund M, Qian Y, Poirier F. Introduction to galectins. Glycoconj. J., 19, 433–440 (2002).
• 3) Di Lella S, Sundblad V, Cerliani JP, Guardia CM, Estrin DA, Vasta GR, Rabinovich GA. When galectins recognize glycans: from biochemistry to physiology and back again. Biochemistry, 50, 7842–7857 (2011).
• 4) Johannes L, Jacob R, Leffler H. Galectins at a glance. J. Cell Sci., 131, jcs208884 (2018).
• 5) Bertuzzi S, Quintana JI, Ardá A, Gimeno A, Jiménez-Barbero J. Targeting galectins with glycomimetics. Front. Chem., 8, 593 (2020).
• 6) Liu F-T, Stowell SR. The role of galectins in immunity and infection. Nat. Rev. Immunol., 23, 479–494 (2023).
• 7) Nio-Kobayashi J, Takahashi-Iwanaga H, Iwanaga T. Immunohistochemical localization of six galectin subtypes in the mouse digestive tract. J. Histochem. Cytochem., 57, 41–50 (2009).
• 8) Saal I, Nagy N, Lensch M, Lohr M, Manning JC, Decaestecker C, André S, Kiss R, Salmon I, Gabius H-J. Human galectin-2: expression profiling by RT-PCR/immunohistochemistry and its introduction as a histochemical tool for ligand localization. Histol. Histopathol., 20, 1191–1208 (2005).
• 9) Oka T, Murakami S, Arata Y, Hirabayashi J, Kasai K-II, Wada Y, Futai M. Identification and cloning of rat galectin-2: expression is predominantly in epithelial cells of the stomach. Arch. Biochem. Biophys., 361, 195–201 (1999).
• 10) Negedu MN, Duckworth CA, Yu L-G. Galectin-2 in health and diseases. Int. J. Mol. Sci., 24, 341 (2022).
• 11) Ohtake K, Shimada N, Uchida H, Kobayashi J. Proteomic approach for identification of protein S-nitrosation in mouse gastric mucosa treated with S-nitrosoglutathione. J. Proteomics, 72, 750–760 (2009).
• 12) Tamura M, Saito M, Yamamoto K, Takeuchi T, Ohtake K, Tateno H, Hirabayashi J, Kobayashi J, Arata Y. S-nitrosylation of mouse galectin-2 prevents oxidative inactivation by hydrogen peroxide. Biochem. Biophys. Res. Commun., 457, 712–717 (2015).
• 13) Sakakura M, Tamura M, Fujii N, Takeuchi T, Hatanaka T, Kishimoto S, Arata Y, Takahashi H. Structural mechanisms for the S-nitrosylation-derived protection of mouse galectin-2 from oxidation-induced inactivation revealed by NMR. FEBS J., 285, 1129–1145 (2018).
• 14) Tamura M, Sasai A, Ozawa R, Saito M, Yamamoto K, Takeuchi T, Ohtake K, Tateno H, Hirabayashi J, Kobayashi J, Arata Y. Identification of the cysteine residue responsible for oxidative inactivation of mouse galectin-2. J. Biochem., 160, 233–241 (2016).
• 15) Tamura M, Sato D, Nakajima M, Saito M, Sasaki T, Tanaka T, Hatanaka T, Takeuchi T, Arata Y. Identification of galectin-2–mucin interaction and possible formation of a high molecular weight lattice. Biol. Pharm. Bull., 40, 1789–1795 (2017).
• 16) Tamura M, Tanaka T, Fujii N, Tanikawa T, Oka S, Takeuchi T, Hatanaka T, Kishimoto S, Arata Y. Potential interaction between galectin-2 and MUC5AC in mouse gastric mucus. Biol. Pharm. Bull., 43, 356–360 (2020).
• 17) He L, Wang H, Garamus VM, Hanley T, Lensch M, Gabius H-J, Fee CJ, Middelberg A. Analysis of MonoPEGylated human galectin-2 by small-angle X-ray and neutron scattering: concentration dependence of PEG conformation in the conjugate. Biomacromolecules, 11, 3504–3510 (2010).
• 18) Geoghegan KF, Dixon HBF, Rosner PJ, Hoth LR, Lanzetti AJ, Borzilleri KA, Marr ES, Pezzullo LH, Martin LB, Lemotte PK, McColl AS, Kamath AV, Stroh JG. Spontaneous α-N-6-phosphogluconoylation of a “His tag” in Escherichia coli: the cause of extra mass of 258 or 178 Da in fusion proteins. Anal. Biochem., 267, 169–184 (1999).
• 19) Yan Z, Caldwell GW, McDonell PA. Identification of a gluconic acid derivative attached to the N-terminus of histidine-tagged proteins expressed in bacteria. Biochem. Biophys. Res. Commun., 262, 793–800 (1999).
• 20) Arata Y, Hirabayashi J, Kasai K. The two lectin domains of the tandem-repeat 32-kDa galectin of the nematode Caernorhabditis elegans have different binding properties. Studies with recombinant protein. J. Biochem., 121, 1002–1009 (1997).
• 21) Tamura M, Arata Y. Expression, S-Nitrosylation, and Measurement of S-Nitrosylation Ratio of Recombinant Galectin-2. Lectin Purification and Analysis. (Hirabayashi J. ed.) Vol. 2132, Springer US, New York, NY, pp. 55–63 (2020).
• 22) Nakamura Y, Gojobori T, Ikemura T. Codon usage tabulated from international DNA sequence databases: status for the year 2000. Nucleic Acids Res., 28, 292 (2000).
• 23) Aon JC, Caimi RJ, Taylor AH, Lu Q, Oluboyede F, Dally J, Kessler MD, Kerrigan JJ, Lewis TS, Wysocki LA, Patel PS. Suppressing posttranslational gluconoylation of heterologous proteins by metabolic engineering of Escherichia coli. Appl. Environ. Microbiol., 74, 950–958 (2008).
• 24) Kim KM, Yi EC, Baker D, Zhang KYJ. Post-translational modification of the N-terminal His tag interferes with the crystallization of the wild-type and mutant SH3 domains from chicken src tyrosine kinase. Acta Crystallogr. D Biol. Crystallogr., 57, 759–762 (2001).
• 25) Lu KV, Rohde MF, Thomason AR, Kenney WC, Lu HS. Mistranslation of a TGA termination codon as tryptophan in recombinant platelet-derived growth factor expressed in Escherichia coli. Biochem. J., 309, 411–417 (1995).