The human hyaluronidase Hyal-1, one of six human hyaluronidase subtypes, preferentially degrades hyaluronic acid present in the extracellular matrix of somatic tissues. Modulations of Hyal-1 expression have been observed in a number of malignant tumors. However, its role in disease progression is discussed controversially due to limited information on enzyme properties as well as the lack of specific inhibitors. Therefore, we expressed human Hyal-1 in a prokaryotic and in an insect cell system to produce larger amounts of the purified enzyme. In Escherichia coli, Hyal-1 formed inclusion bodies and was refolded in vitro after purification by metal ion affinity chromatography. However, the enzyme was produced with extremely low folding yields (0.5%) and exhibited a low specific activity (0.1 U/mg). Alternatively, Hyal-1 was secreted into the medium of stably transfected Drosophila Schneider-2 (DS-2) cells. After several purification steps, highly pure enzyme with a specific activity of 8.6 U/mg (consistent with the reported activity of human Hyal-1 from plasma) was obtained. Both Hyal-1 enzymes showed pH profiles similar to the hyaluronidase of human plasma with an activity maximum at pH 3.5–4.0. Deglycosylation of Hyal-1, expressed in DS-2 cells, resulted in a decrease in the enzymatic activity determined by a colorimetric hyaluronidase activity assay. Purified Hyal-1 from DS-2 cells was used for the investigation of the inhibitory activity of new ascorbic acid derivatives. Within this series, l-ascorbic acid tridecanoate was identified as the most potent inhibitor with an IC50 of 50 ± 4 µM comparable with glycyrrhizic acid.
Mammalian hyaluronidases (E.C. 126.96.36.199, glycoside hydrolase family 56) are β-1,4-endoglycosaminidases that degrade hyaluronic acid (HA), a high molecular weight glycosaminoglycan of the extracellular matrix. The major amount of hyaluronan in the somatic tissue is assumed to be degraded by Hyal-1 in combination with Hyal-2 and, to a minor extent, by PH-20 (Stern 2005). While Hyal-2 and PH-20 are both linked to the cell membrane by a glycosylphosphatidylinositol anchor, Hyal-1 is a soluble, acid active lysosomal enzyme (Gold 1982). Acid active hyaluronidase has been isolated and purified from plasma, serum, and liver, but characteristics as well as molecular sizes of the proteins isolated varied: 57 kDa for plasma hyaluronidase (Frost et al. 1997), 72 kDa for serum hyaluronidase (Afify et al. 1993), and 76 kDa for human liver hyaluronidase (Gold 1982). Although only plasma hyaluronidase has been identified as a gene product of hyal-1, the hyaluronidases isolated from serum and liver also showed characteristic properties of Hyal-1. In urine, an additional protein binding to anti-Hyal-1 antibodies was described by Csoka et al. (1997) with a molecular mass of 45 kDa. This second, enzymatically active Hyal-1 isoform was supposed to be a truncated form of the 57 kDa isoform with an interior part of the protein missing. However, function and mechanism of this truncation are unknown.
An extraordinarily high turn over of hyaluronan is observed during the processes of tissue remodeling, i.e., embryogenesis, wound healing, cell proliferation, cell migration, and angiogenesis (for review see Toole 2004). Therefore, the influence of hyaluronan, hyaluronan synthases, and hyaluronidases on cancer growth, angiogenesis, and metastasis is currently an area of intensive research (Neudecker et al. 2004; Fiszer-Szafarz et al. 2005; Paiva et al 2005; Boregowda et al. 2006).
Modulation of Hyal-1 expression has been observed in a genetic disorder called mucopolysaccharidosis IX (Triggs-Raine et al. 1999) and in a number of malignant tumors. On the basis of the identification of the hyal-1 gene in a region of major deletions in tobacco-related lung carcinomas, the gene was originally called lung cancer protein-1 prior to correlation of the protein with its hyaluronidase function (Hosoe et al. 1994; Wei et al. 1996). In recent years, it became obvious that hyal-1 functions not only as a tumor suppressor gene as observed in lung cancer, but can also promote tumor growth and aggressiveness in prostate and bladder cancer (Lokeshwar et al. 2001; Franzmann et al. 2003). Hyal-1 in human urine has therefore been investigated for its potential as a marker protein for these malignant urogenital cancers (Posey et al. 2003; Lokeshwar et al. 2005; Aboughalia 2006). However, information on the role of Hyal-1 in cancer progression remains controversial possibly because several different mechanisms control the activity of Hyal-1 in vivo, e.g., control of Hyal-1 expression on the DNA as well as on the RNA level (Lokeshwar et al. 2002; Stern 2005) and the existence of physiological inhibitors of hyaluronidase activity (Mio and Stern 2002).
In-depth characterization of the enzyme has been hampered by the extremely low amounts of Hyal-1 present in the plasma and by the instability of the enzyme during the purification process. Furthermore, structural information with respect to human hyaluronidases is available only from the X-ray data of bee venom hyaluronidase (BVH; Botzki et al. 2004; Jedrzejas and Stern 2005).
Highly pure and homogeneous Hyal-1 was not available till now in quantities sufficient for intensive enzymological studies, crystallization and the development of specific inhibitors, which can be used, for example, as pharmacological tools. Therefore, we tried a prokaryotic as well as a eukaryotic expression system for the recombinant production and purification of human Hyal-1. The recombinant enzymes were compared with respect to their catalytic activity and the effects of glycosylation, pH, and NaCl concentration on the enzymatic activity. In case of human Hyal-1, expressed in Drosophila Schneider-2 (DS-2) cells, we identified ascorbic acid derivatives as a new class of Hyal-1 inhibitors.
Synthesis of Hyal-1 expression vectors of E. coli and DS-2 cells
The primary structure analyses of Hyal-1 reveal the presence of three domains: an N-terminal signal peptide directing its expression into the endoplasmic reticulum (ER), a hyaluronidase catalytic domain, and a C-terminal domain with similarity to epidermal growth factor (EGF) domains (Figure 1). The residues of the active site in the catalytic domain are highly conserved compared with those in BVH (D111, E113, Y184, Y227, and W301 in BVH; Markovic-Housley et al. 2000; Stern and Jedrzejas 2006) suggesting a similar degradation mechanism. The function of the C-terminal domain of Hyal-1, however, is unknown.
Four cysteine (Cys) residues linked by disulfide bridges in BVH are strictly conserved in all human hyaluronidases, indicating the presence of disulfide bridges in the human enzymes as well. The C-terminal domain contains another six Cys residues, which may also be linked by disulfide bridges. Therefore, human Hyal-1, expressed in Escherichia coli (Hyal-1–E. coli), was expected to aggregate as insoluble inclusion bodies (IBs) in the reducing cytoplasm of E. coli (Baneyx and Mujacic 2004). Expression with an N-terminal His-tag allowed for the purification under denaturing conditions. The signal peptide targeting the protein to the ER was excluded from the coding sequence of Hyal-1–E. coli (Figure 1).
Human Hyal-1, expressed by DS-2 cells (Hyal-1–DS-2), was planned to be secreted in a catalytically active conformation into the medium of the insect cells due to an exchange of the mammalian signal sequence for an insect cell signal peptide targeting the protein to the extracellular compartment. The hyal-1 cDNA was fused C-terminally to a His-tag and a V5-epitope (Figure 1).
Expression of Hyal-1– E. coli and purification
Hyal-1–E. coli was expressed in E. coli BL21(DE3)RIL with maximum expression levels reached 4–5 h after induction of expression. A 48-kDa protein, comprising up to 31% of the total cell protein, was overexpressed as determined by densitometric analysis of the complete cell lysate after sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE). The molecular mass is consistent with the calculated molecular mass of 48.3 kDa for Hyal-1–E. coli, and the identity of the protein was confirmed by western blot analysis using anti-his-antibodies as primary antibodies (Figure 2). As expected, Hyal-1–E. coli was exclusively found in the insoluble cell fraction. The IBs were isolated, completely solubilized in 8 M urea or 6 M guanidinium chloride (GdmCl), respectively, and reduced by the addition of 50 mM dithiothreitol (DTT). Purification by Ni-ion metal affinity chromatography (Ni-IMAC) was achieved in the presence of denaturing agents, and Hyal-1–E. coli was eluted by reducing the pH from 7.8 to 4.5 (Figure 2).
In vitro folding of Hyal-1–E. coli
A variety of folding conditions was tested by direct dilution of denatured Hyal-1/E. coli in different folding buffers. During initial folding experiments, a strong dependency of aggregation in the folding mixture on the type of denaturing agent (urea or GdmCl) was observed. Measurements of the turbidity in the folding mixture, immediately after the addition of increasing amounts of protein, revealed that folding solutions containing Hyal-1–E. coli, denatured in urea, remained almost clear after the addition of protein, whereas GdmCl-denatured Hyal-1–E. coli showed a strong tendency for aggregation even at low protein concentrations (Figure 3). Addition of l-arginine·HCl (Arg), which was found to have advantageous effects on the enzymatic activity during the folding reaction did not influence the turbidity caused by protein aggregation in the folding mixture (Figure 3).
Although different folding additives (sugars, detergents, glycerol, ammonium sulfate, urea) and conditions (pH, temperature, protein concentration) were tested (data not shown), only variations of the redox-shuffling system containing reduced (GSH) and oxidized glutathione (GSSG) and addition of Arg resulted in the formation of enzymatically active Hyal-1–E. coli (Figure 4) (Lilie et al. 1998; Mayer and Buchner 2004). Hyaluronidase activity was determined in the colorimetric activity assay (Morgan-Elson assay) after the removal of folding additives by dialysis. During dialysis extensive formation of aggregates was observed, which were removed by centrifugation after dialysis.
For comparison of enzymatic activities, independent of the protein concentration used in the folding reaction, the enzymatic activity was expressed as nanomoles N-acetylglucosamine (GlcNAc) ends formed per minute and per milligram IB material present in the folding mixture. Proper folding of Hyal-1–E. coli was found to depend decisively on the correct formation of disulfide bridges with the highest enzymatic activity found at a GSSG to GSH ratio of 1:3 mM in the folding reaction (Figure 4A). Additionally, the efficiency of folding was strongly dependent on the presence of Arg: at a GSSG to GSH ratio of 1:3 mM and in the absence of Arg, only about 8% of the enzymatic activity observed in the presence of 1 M l-arginine–HCl was obtained (Figure 4A, bar D and E). When the gain of enzymatic activity was observed over 120 h in the presence of various disulfide shuffling systems, maximum hyaluronidase activity was reached in a folding system containing a GSSG to GSH ratio of 1:1 mM and 1 M Arg after 24 h of incubation (Figure 4B).
Arg is a common folding additive (De Bernadez Clark et al. 1999), although its exact mechanism of folding support has not been elucidated till now. Presumably, it facilitates the solubilization of folding intermediates without the denaturing effect of chaotropic salts (Tsumoto et al. 2004; Baynes et al. 2005; Reddy et al. 2005). However, as shown previously, Arg could not inhibit the aggregation during folding of GdmCl-denatured Hyal-1–E. coli. Folding experiments performed with GdmCl-denatured Hyal-1–E. coli yielded only extremely low quantities of active enzyme.
Expression of Hyal-1–DS-2 in insect cells
Transfection of DS-2 cells with the vector pMT/Hygro/hyal-1 allowed transient expression of Hyal-1–DS-2 as well as direct selection of transfected clones for hygromycin resistance. Ten days after transient transfection and 9 days after induction of expression, respectively, a hyaluronidase activity of 0.6 nmol GlcNAc/min mL cell suspension was measured in the medium, while the complete cell lysate exhibited an activity of 0.03 nmol GlcNAc/min mL cell suspension. The medium and complete cell lysate of wild-type DS-2 cells did not exhibit any hyaluronidase activity under identical assay conditions.
The presence of hyaluronidase activity in the cell lysate of transfected cells might be due to the fact that Hyal-1 is a lysosomal enzyme (Gold 1982). Although no simple consensus sequence exists for the transport of proteins to lysosomes, a structural motif in the three-dimensional structure of lysosomal proteins is supposed to be responsible for the sorting of the majority of proteins to lysosomes via the mannose-6-phosphate receptor (Ni et al. 2006). Possibly, the human sequence motif for protein transport to lysosomes is, at least partially, recognized by insect cells, resulting in an accumulation of recombinant Hyal-1–DS-2 not only in the extracellular compartment, but also in lysosomes.
The low hyaluronidase activity in the medium indicates that the transfection efficiency was rather low. After selection of hygromycin resistant DS-2 cells and optimization of growth and expression conditions ca. 150 nmol GlcNAc/min mL of cell suspension could be achieved in the medium 10 days after induction.
Purification of Hyal-1–DS-2
Firstly, Hyal-1–DS-2 was concentrated from the medium of DS-2 cells by ammonium sulfate precipitation. At 30% saturation of the medium with ammonium sulfate, Hyal-1–DS-2 was found in the precipitate almost completely. This precipitation step reduced the sample volume by a factor of 10. Purification of Hyal-1–DS-2 was achieved by cation exchange chromatography and Ni-IMAC (Figure 5). The isoelectric point of Hyal-1–DS-2 was calculated to be 7.3; thus, the net charge of Hyal-1–DS-2 was positive at pH 6.0, the working pH of the cation exchange chromatography. Hyal-1–DS-2 was eluted at 0.1–0.3 M NaCl using a linear NaCl gradient. Therefore, Ni-IMAC could be performed directly after the addition of 0.2 M NaCl and adjustment of the pH to 7.8. Cu2+ ions that had been present in the cell medium after induction and could block the His-tag from binding to the Ni-IMAC were removed by the addition of ethylenediaminetetraacetic acid (EDTA) to the medium directly after separation from the cells. During the whole purification procedure, Triton X-100 had to be added to the buffer to avoid adhesion of Hyal-1–DS-2 to column materials as well as to centrifugal devices in the absence of detergent.
The identity of the purified protein with a molecular mass of 54 kDa was confirmed by western blot analysis using anti-V5-antibody as primary antibody. The experimentally determined molecular mass was ca. 4 kDa higher than the molecular mass calculated from the amino acid sequence of Hyal-1–DS-2 (50.5 kDa), indicating the presence of posttranslational modifications, e.g., glycosylations. A short overview of the characteristics of Hyal-1–E. coli and Hyal-1–DS-2 is given in Table I.
|Expression system||E. coli||DS-2 cells|
|Molecular mass||48 kDa||54 kDa|
|Glycosylation||None||N-glycosylation (3–4 kDa)|
|Active protein yield||up to 1.0 mU/mL of folding buffer||up to 150 mU/mL of medium|
|Specific activity||0.11 U/mg||8.6 U/mg|
|Expression system||E. coli||DS-2 cells|
|Molecular mass||48 kDa||54 kDa|
|Glycosylation||None||N-glycosylation (3–4 kDa)|
|Active protein yield||up to 1.0 mU/mL of folding buffer||up to 150 mU/mL of medium|
|Specific activity||0.11 U/mg||8.6 U/mg|
Analysis of protein glycosylation
To confirm the presence of glycosidic modifications, purified Hyal-1–DS-2 was stained after SDS–PAGE using the periodic acid–Schiff (PAS) technique. As expected, Hyal-1–E. coli did not react with Schiff's reagent, whereas Hyal-1–DS-2 formed a pink band on the gel, indicating the presence of glycosidic residues (Figure 6).
The amino acid sequence of Hyal-1–DS-2 contains three potential N-glycosylation sites with the consensus sequence N-X-T/S and no O-glycosylation sites as determined by the ExPASy NetOGlyc 3.1 tool (Julenius et al. 2005). When native Hyal-1–DS-2 was deglycosylated by incubation with increasing amounts of N-glycosidase F (PNGase F), a second band was detected in the western blot with a molecular mass reduced by ca. 3–4 kDa compared with the untreated control sample (0 units N-glycosidase F) (Figure 7). Although deglycosylation could not be followed till the completion of the process, a 40% reduction in enzymatic activity could be observed in the colorimetric activity assay when about half of the protein present on the western blot was deglycosylated.
pH dependent degradation of HA
Human plasma hyaluronidase was reported to be an acid active protein with an optimum pH of 3.0–4.0 for the degradation of HA (Gold 1982; Afify et al. 1993; Frost et al. 1997; Muckenschnabel et al. 1998). The pH profiles of Hyal-1–E. coli and Hyal-1–DS-2 resemble closely to the profile of human plasma hyaluronidase reported by Muckenschnabel et al. (1998), which was measured under comparable assay conditions (Figure 8). The pH profile of Hyal-1–DS-2 was shifted slightly to higher pH values with an optimum pH 4.0 instead of 3.5, possibly due to effects of the buffer present in the enzyme samples. Hyaluronidase activity profiles were also measured in the turbidimetric activity assay to determine whether the pH optima of Hyal-1 were dependent on the type of assay as previously described for the pH profiles of bovine testicular hyaluronidase (BTH; Hoechstetter et al. 2001). However, the turbidimetric and the colorimetric activity assay resulted in nearly identical pH profiles for both recombinant enzymes, Hyal-1–E. coli and Hyal-1–DS-2.
Effect of NaCl on Hyal-1 activity
For human hyaluronidase isolated from liver and plasma, an inhibitory effect of NaCl at concentrations exceeding 100 mM has been reported (Gold 1982; Afify et al. 1993). Indeed, colorimetric determination of Hyal-1 activity at pH 3.5 with increasing NaCl concentration (0–1 M) revealed an inhibition of hyaluronidase activity by NaCl. However, Hyal-1–E. coli was already inhibited at concentrations >50 mM, while Hyal-1–DS-2 showed an optimum hyaluronidase activity at 100 mM NaCl with a fast drop in activity at concentrations >100 mM. At 1 M NaCl, the activity of both enzymes was completely inhibited. Without NaCl, the activity of Hyal-1–DS-2 decreased by ca. 40% compared with the optimum concentration of 100 mM, whereas the activity of Hyal-1–E. coli remained at its maximum even without any NaCl present.
l-ascorbic acid derivatives as inhibitors of Hyal-1–DS-2
Recently, the weak inhibition of BTH by ascorbic acid (Menzel and Farr 1998) was increased via addition of hydrophobic residues, and the crystal structure of bacterial hyaluronidases from Streptococcus pneumoniae in complex with ascorbic acid 6-O-palmitate was elucidated (Botzki et al. 2004). On basis of these results, we investigated ascorbic acid derivatives bearing lipophilic alkanoyl residues. These compounds, recently synthesized and identified as inhibitors of bacterial and bovine hyaluronidases (Spickenreither et al. 2006), were investigated with respect to inhibition of Hyal-1–DS-2 (Table II). The concentration-dependent inhibition of the enzyme by compounds 2–5 and 8 determined in a turbidimetric hyaluronidase activity assay is depicted in Figure 9. Within the series of new ascorbic acid derivatives, l-ascorbic acid tridecanoate (5) proved to be the most potent inhibitor with an IC50 of 50 ± 4 µM comparable to glycyrrhizic acid (8) for which the IC50 was determined to be 26 ± 1 µM.
|No.||R||IC50 (µM)a (pH 3.5)|
|2||C9H19||378 ± 12|
|3||C10H21||143 ± 4|
|4||C11H23||76 ± 3|
|5||C12H25||50 ± 4|
|8||–||26 ± 1|
|No.||R||IC50 (µM)a (pH 3.5)|
|2||C9H19||378 ± 12|
|3||C10H21||143 ± 4|
|4||C11H23||76 ± 3|
|5||C12H25||50 ± 4|
|8||–||26 ± 1|
aInhibition of enzyme expressed as IC50 (µM), mean values ± SEM (N = 2, experiments performed in duplicate), maximal concentrations of the test compounds were selected with respect to individual terminal solubilities in the incubation mixture.
The choice of a suitable expression system for a protein cannot always be made a priori. Moreover, an empirical approach is often necessary to find the best system for a particular protein. Comparison of Hyal-1 expression and purification in an E. coli and a DS-2 cell system clearly revealed the advantages and problems of each system. While the bacterial system is definitely the simplest and fastest one to achieve high amounts of protein, insect cells grow slower with much lower yields, but produce protein with a glycosylation pattern similar to mammalian proteins.
Hyal-1 expressed in bacteria formed IBs due to the absence of disulfide bonds in the reducing cytoplasm of E. coli (Schoemaker et al. 1985; Bowden et al. 1991). The enzyme was purified under denaturing conditions and then refolded in several folding buffer systems (Misawa and Kumagai 1999). A strong tendency for aggregation was observed when folding was performed with GdmCl denatured Hyal-1–E. coli. Analysis of the primary sequence revealed a high amount of charged residues present in Hyal-1–E. coli, while other parameters like hydrophobicity (Kyte and Doolittle 1982) did not indicate any extraordinary behavior of the enzyme. Uversky (2002) proposed a systematic explanation for the tendency of proteins to form stable intermediates during the unfolding process dependent on the ratio of the mean net charge (number of charged amino acids at pH 7 per total number of amino acids) to the mean hydrophobicity (sum of hydrophobicity values per total number of amino acids). In this model, proteins with low to medium net charge and low hydrophobicity tend to form stable intermediates, whereas proteins with a high overall hydrophobicity tend to unfold without any intermediates rather independent on the net charge. The mean net charge of Hyal-1–E. coli with a value of 0.15 is extraordinarily high compared with the proteins investigated by Uversky. Hyal-1–E. coli is positioned in a range with medium hydrophobicity (0.45) and a very high net charge, i.e., in a region of intermediate-forming proteins if the model is being extrapolated to higher mean net charge values. In contrast to urea, GdmCl is a charged molecule, hence electrostatic interaction of GdmCl to Hyal-1–E. coli with this extraordinarily high overall charge could explain the effect of strong aggregation when folding was performed with GdmCl denatured protein (Bhuyan 2002). Another possible explanation could be the weaker denaturing ability of urea compared with GdmCl. However, this would imply the presence of partially folded structures in the IB aggregates.
Folding of Hyal-1–E. coli was to a great extent dependent on the presence of a suitable disulfide shuffling system indicating the necessity of the correct formation of disulfide bonds in the native protein (Schoemaker et al. 1985). Furthermore, Arg was essential to gain enzymatic activity, probably stabilizing intermediate structures (Baynes et al. 2005). In summary, folding of Hyal-1–E. coli yielded enzymatically active protein, but the specific hyaluronidase activity as well as the yield of active protein could not be improved.
In DS-2 cells, human Hyal-1 was expressed in the native conformation into the medium. Hyaluronidases have been described in the venom of arthropods and insects, but in the extracellular matrix of Drosophila melanogaster, no hyaluronan has been found (Toyoda 2002). Furthermore, no sequences with any significant similarity to hyaluronidase genes were found in the genome of D. melanogaster (communication by one of the reviewers). Drosophila Schneider-2 cells are undifferentiated, embryonic cells growing at high densities (up to 20 × 106 cells/mL) in serum-free medium (Schneider 1972; Deml and Wagner 1998). Both the absence of endogenous hyaluronidases and the continuous growth at high densities were well suited for the expression of human Hyal-1.
Purification of Hyal-1–DS-2 had to be optimized using various column materials and purification steps, mainly due to unspecific binding of insect cell proteins to the Ni-IMAC and the tendency of Hyal-1–DS-2 to adhere to surfaces, e.g., column materials and filter membranes. Adhesion could be minimized by adding a nonionic detergent in the buffer systems. Earlier studies also described the necessity of detergents during the purification of human Hyal-1 (Frost et al. 1997).
Although both expression systems yielded enzymatically active Hyal-1, a comparison of yields as well as specific activities clearly shows the superiority of the insect cell system (Table I). HA was catabolized by both enzymes with similar pH activity profiles, but the absolute activity seems to be influenced by the glycosylation, which is only present in Hyal-1–DS-2. The decrease in activity in the course of deglycosylation can be explained as a consequence of either increased aggregation or surface adhesion of the enzyme or by a direct effect on the enzymatic activity.
The sensitivity of Hyal-1 to NaCl concentrations >100 mM has been reported (Gold 1982) and was observed for the recombinant enzymes as well. However, it remains unclear whether NaCl influences the structure of the substrate, HA, or exerts a direct effect on the enzyme. As the sensitivity of Hyal-1–E. coli and Hyal-1–DS-2 against NaCl differs, it can be assumed that in the relevant concentration range, NaCl does not significantly affect the structure of HA.
Up to now only very few low molecular weight inhibitors of Hyal-1 with IC50 values in the (sub)millimolar range have been described, glycyrrhizic acid and tetradecane-1-sulfonic acid are among them (Isoyama et al. 2005). l-ascorbic acid derivatives esterified with fatty acids have already proved as potent inhibitors of bacterial and BTHs (Spickenreither et al. 2006). Whereas the inhibitory activity of those compounds on bacterial hyaluronidases and BTHs increased with the length of the aliphatic chain (highest potency was found for l-ascorbic acid 6-octadecanoate with IC50 values of 39 µM (BTH) and 0.8 µM (hylB4755), l-ascorbic acid tridecanoate showed highest inhibition of Hyal-1–DS-2 (IC50 = 50 µM ± 4). Together with glycyrrhizic acid, the investigated vitamin C derivatives are the most potent low molecular weight inhibitors of Hyal-1–DS-2 known to date. The IC50 value of glycyrrhizic acid determined in the turbidimetric assay is in good agreement with the value published by Isoyama et al. (2005), who determined an IC50 of 39.4 µM, in an enzyme-linked immunosorbent assay-like assay using partially purified Hyal-1 at pH 4.2.
In summary, Hyal-1 expressed in DS-2 cells was found to be superior to Hyal-1 expressed in E. coli and can be used for enzymological studies, crystallization, and the development of further specific inhibitors.
Materials and methods
Materials and standard techniques
All sequencing experiments were done by Entelechon, Regensburg, Germany. Primers were synthesized by MWG Biotech AG, Ebersberg, Germany. dNTPs, restriction enzymes, ligase, RNase A, polymerases, and respective buffers were purchased from Fermentas, St. Leon-Rot, Germany. N-glycosidase F was purchased from Roche, Mannheim, Germany. Antibiotics and Arg were purchased from Sigma-Aldrich, Munich, Germany. Substances for microbiological experiments (tryptone, yeast extract, agar) were purchased from Roth, Karlsruhe, Germany. Ascorbic acid 6-palmitate and glycyrrhizic acid ammonium salt were purchased from Sigma-Aldrich. If not indicated otherwise all chemicals were of analytical grade or high-performance liquid chromatography grade and obtained from Merck, Darmstadt, Germany.
E. coli Top 10 was obtained from Invitrogen, Karlsruhe, Germany, E. coli BL21(DE3)RIL from Stratagene, La Jolla, CA, and pET15b from Merck. DS-2 cells, pMT/Bip/V5-HisA, and pCoHygro were kindly provided by V. Runza (Institute of Immunology, Faculty of Medicine, University of Regensburg, Regensburg, Germany).
Subcloning of hyal-1 cDNA into pET15b
Human hyal-1 cDNA was purchased as I.M.A.G.E. clone IRAKp961O1779Q2 from the RZPD (Resource Centre/Primary Database, Berlin, Germany) and the sequence was confirmed by sequencing. The hyal-1 cDNA was polymerase chain reaction (PCR)-amplified using the I.M.A.G.E. clone as template, a forward primer generating an additional NdeI restriction site (5′-CTTAATTC ATATGTTTAGGGGCCCCTTGC-3′) and a reverse primer generating an additional BamHI restriction site (5′-TAAGGATCCTCACCAC ATGCTCTTCCG-3′). After cleavage with NdeI and BamHI, the PCR product was ligated into the multiple cloning site of pET15b. The recombinant vector was amplified in E. coli Top 10, isolated, and sequenced.
Expression of Hyal-1–E. coli
For expression of Hyal-1–E. coli, pET15b/hyal-1 was transformed into CaCl2-competent E. coli BL21(DE3)RIL. An overnight culture of freshly transformed bacteria was used to inoculate LB (Luria–Bertani)-medium supplemented with 100 µg/mL ampicillin to an OD600 (optical density at 600 nm) of 0.15. Bacteria were grown for 2–3 h at 200 rpm and 37 °C until an OD600 of 0.6 ± 0.5 was reached. Protein expression was induced by the addition of 1 mM isopropylthiogalactoside and bacteria were cultivated 4 h before they were harvested by centrifugation (30 min, 5000g, 4 °C).
Isolation and purification of IBs
Inclusion bodies were isolated as described by De Bernadez Clark et al. (1999) and solubilized after washing in denaturing solution containing 20 mM Tris–HCl, 1 mM EDTA, and 50 mM DTT, pH 8.5 containing either 8 M urea or 6 M GdmCl. After stirring overnight, the suspension was centrifuged and the supernatant dialyzed extensively against denaturing solution without EDTA and DTT (MWCO 14 000, Millipore, Schwalbach, Germany) (Jaenicke and Rudolph 1989). Hyal-1–E. coli was bound under denaturing conditions (20 mM Tris–HCl, 8 M urea, pH 8.5) on a Ni-IMAC column (Quiagen, Hilden, Germany) equilibrated in dialysis buffer and eluted by decreasing the pH from 8.5 to 4.5.
In vitro protein folding
Completely denatured and reduced Hyal-1–E. coli was concentrated using centrifugal filter devices (MWCO 14 000, Millipore) to a concentration of 5–10 mg/mL. This protein solution was added drop wise under intensive stirring to a degassed folding buffer [100 mM Tris–HCl, 1 mM EDTA, 0.2 mg/mL bovine serum albumin (BSA), pH 8.5] supplemented with additives as indicated for each experiment.
Aggregation was observed by measuring the turbidity at 600 nm immediately after the addition of the protein to the folding buffer. For the determination of hyaluronidase activity, samples of the folding mixture were dialyzed against 20 mM Tris–HCl, pH 7.5, and centrifuged to remove protein aggregates.
Subcloning of hyal-1 cDNA into the insect cell expression vector pMT/Hygro
pMT/BiP/V5-HisA and pCoHygro (both developed by Invitrogen) were cleaved with AccI and SapI and ligated to gain the expression plasmid pMT/Hygro, containing a gene for hygromycin resistance as well as the cloning/expression site for a foreign gene.
Human hyal-1 cDNA was amplified by PCR from the I.M.A.G.E. clone using 5′-ATCTA TTCCCGGGTTTAGGGGCCCCTTGC-3′ as forward primer generating a SmaI cleavage site and 5′-ATTATATCGCGGCCGCCACATGCTCTTCC-3′ as reverse primer with an additional NotI cleavage site. The PCR product was ligated into pMT/Hygro using the cloning procedures described in Subcloning of hyal-1 cDNA into pET15b.
Expression of Hyal-1/DS-2 in insect cells
DS-2 cells were grown in serum-free Insect-XPRESSTM medium (Cambrex Bio Science, Copenhagen, Denmark) at 27 ± 1 °C as adherent culture for transfection and continuous growth. For large-scale protein expression, the cells were cultivated in a shaker at 135 rpm. DS-2 wild-type cells were transfected with pMT/Hygro/hyal-1 using lipofectamin (Invitrogen) as transfection reagent. For the selection of stably transfected DS-2/pMT/Hygro/hyal-1 cells, the medium was supplemented with 300 µg/mL hygromycin (A.G. Scientific Inc., San Diego, CA). Protein expression was induced by the addition of 0.5 mM CuSO4 to a suspension culture with a cell density of 1.5 − 2.0 × 106 cells/mL. Ten days after induction, the cells were harvested by centrifugation (1500g, 15 min, 4 °C) and the medium was used for isolation of Hyal-1–DS-2.
Purification of Hyal-1–DS-2 from insect cell medium
After addition of 1 mM EDTA and 0.1% Triton X-100 to the medium, Hyal-1–DS-2 was precipitated by saturation of the medium with 30% ammonium sulfate. The suspension was incubated on ice for 1 h before the precipitate was pelleted by centrifugation (20 000g, 60 min, 10 °C). The pellet was solubilized in one-tenth of the original medium volume of 50 mM sodium phosphate, 0.1% Triton X-100, pH 6.0. After stirring overnight at 4 °C, insoluble protein aggregates were removed by centrifugation (20 000g, 60 min, 4 °C) and the clear supernatant was applied to CM (Carboxymethyl)-Sephadex C-50 (Amersham Biosciences, Uppsala, Sweden) packed in a 26/20 column [40 mL column volume (CV), Amersham Biosciences] and equilibrated in 50 mM sodium phosphate, 0.1% Triton X-100, pH 6.0. The column was washed with 2 CV of equilibration buffer before Hyal-1–DS-2 was eluted by a linear gradient of NaCl concentrations from 0 to 0.5 M increasing over 10 CV. After elution, the column was washed with 2 CV of equilibration buffer supplemented with 1 M NaCl. Cation exchange chromatography was performed at a flow rate of 1 mL/min and a detector wavelength of 280 nm. Fractions of 8 mL were collected automatically. Protein purification was performed with ÄKTA FPLC device with a Frac-950 fraction collector (Amersham Biosciences) using UNICORNTM v5.10 software (Amersham Biosciences) for data analysis.
Fractions exhibiting hyaluronidase activity were pooled and the pH of the eluate was carefully adjusted to pH 7.8 with 1 M NaOH. The sample was then packed in a HisTrapTM FF column (15 mL CV, Amersham Biosciences) equilibrated in 50 mM sodium phosphate, 0.5 M NaCl, 0.1% Triton X-100, pH 7.8. The column was washed with 2 CV of equilibration buffer and elution was achieved by decreasing the buffer pH to 4.5. Ni-IMAC was performed at a flow rate of 0.5 mL/min and a detector wavelength of 280 nm. Fractions of 1 mL were collected and pooled after determination of their specific enzymatic activity. Enzyme purity was confirmed by SDS–PAGE.
Determination of protein concentrations
During expression, purification, and folding of Hyal-1–E. coli, protein concentrations were determined by Bradford (1976) method using the Bio-Rad protein assay (Bio-Rad, Munich, Germany) with BSA as standard. The assay was performed according to the supplier's instructions.
As detergents disturb the Bradford assay, protein concentrations during the purification procedure of Hyal-1–DS-2 were determined by the bicinchoninic acid (BCA) assay (Smith et al. 1985) using BSA as standard. The assay was performed by mixing 70 µL of the BCA working solution, described by Smith et al. (1985), with 70 µL of sample. Absorption was quantified after 1 h incubation at 60 °C at 580 nm in a microplate reader (Tecan Deutschland GmbH, Crailsheim, Germany) applying 10 flashes/well.
SDS–PAGE and western blotting
Proteins were analyzed by SDS–PAGE under reducing conditions. Electrophoresis was performed in a PerfectBlue gel electrophoresis system Twin S (Peqlab, Erlangen, Germany) using 12% polyacrylamide gels according to the method described by Laemmli (1970).
For blotting, the proteins were transferred to a nitrocellulose membrane (0.2 µm, Peqlab) in a PerfectBlue “Semi-Dry” electro blot apparatus (Peqlab, Erlangen, Germany) for 1.5 h at 10 mA/10 cm2 of membrane. The membrane was subsequently blocked by washing with 5% (w/v) milk powder in phosphate-buffered saline (PBS). After 3 × 5 min washing with washing buffer [0.05% (v/v) Tween 20 in PBS], the primary antibody diluted to 2000-fold in 10 mL washing buffer was incubated overnight with the membrane. After washing, 5 µL biotinylated anti-mouse/rabbit IgG (Vector Laboratories, Burlingame, CA) in 10 mL washing buffer was added and shaken with the membrane for 1 h. A biotin–avidin–horseradish peroxidase mixture was prepared as described by the supplier (VECTASTAIN® ABC-kit Standard, Vector Laboratories) and incubated with the membrane after washing for 30 min. After washing, the membrane was stained for 2–15 min with diaminobenzaldehyde (DAB) staining solution (DAB kit, Vector Laboratories), rinsed with water, and dried. Blots and gels were analyzed in a Bio-Rad gel detection system (GS-710 Imaging Densitometer) using Quantity One quantification software, version 4.0.3. (Biorad, Munich, Germany).
Identification of glycoproteins
For the detection of glycoproteins, SDS-PAGE gels were stained by PAS technique with Schiff's reagent, purchased from Sigma-Aldrich, following the procedure described by Zacharius et al. (1969).
An amount of 4 µg of partially purified Hyal-1–DS-2 (0.6 U/mg) was incubated with 0–2 U N-glycosidase F for 4 h at 37 °C in 100 µL 50 mM sodium phosphate, 10 mM EDTA, pH 6.0. A volume of 20 µL of each sample was used for western blotting and 50 µL was analyzed by the colorimetric hyaluronidase activity assay.
Colorimetric activity assay
Hyaluronidase activity was determined by a colorimetric activity assay (Morgan-Elson assay) according to the method of Reissig et al. (1955). A detailed description of the assay conditions is given elsewhere (Muckenschnabel et al. 1998). If not indicated otherwise McIlvaine's buffer was used as incubation buffer (solution A: 0.2 M Na2HPO4, 0.1 M NaCl, solution B: 0.1 M citric acid, 0.1 M NaCl; solutions A and B were mixed in the appropriate proportions to reach the desired pH). For the measurement of NaCl sensitivity, McIlvaine's buffer without NaCl was used and NaCl was added separately to obtain the concentrations of 0–1 M NaCl in the incubation mixture. The enzymatic activities (in U; 1 U of hyaluronidase catalyzes the liberation of 1 µmol NAcGlc at the reducing ends of sugars per min) were determined using 0.2 M sodium formate, 0.1 M NaCl, pH 3.68, as incubation buffer according to the definition of the International Union of Biochemistry.
Turbidimetric activity assay
The turbidimetric hyaluronidase activity assay described by Di Ferrante (1956) was modified to perform turbidimetric activity studies in 96-well plates. Incubation mixtures contained the following compounds: 31 µL incubation buffer, 8 µL BSA (0.2 mg/mL), 8 µL HA (2 mg/mL), 13 µL H2O. For inhibitor tests, 3 µL inhibitor solution in dimethylsulfoxide was additionally added to the incubation mixture. If not indicated otherwise McIlvaine's buffer was used as incubation buffer (see Colorimetric activity assay). The enzymatic reaction was started by the addition of 10 µL enzyme solution. After incubation at 37 °C, the enzymatic reaction was stopped by the addition of 200 µL of alkaline cetyltrimethylammonium bromide (CTAB) solution [2.5% (w/v) CTAB in 0.5 M NaOH], and the plates were incubated for 20 min at room temperature. The turbidity was quantified by the measurement of the OD at 355 nm in a microplate reader (Tecan Deutschland GmbH, Crailsheim, Germany). The plate was shaken in the reader for 10 s and then the OD was measured after 2 s of settling time by five flashes at the center of each well.
For the investigation of potential inhibitors, samples without inhibitor and without both inhibitor and enzyme were taken as references. The activities were plotted against the logarithm of the inhibitor concentration, IC50 ± SEM values were calculated by curve fitting of the experimental data with Sigma Plot 8.0 (SPSS Inc., Chicago, MI), and the means of at least two independent experiments were performed in duplicate.
Synthesis of ascorbic acid derivatives
Compounds were synthesized as described elsewhere (Spickenreither et al. 2006). The results of elemental analysis (C, H) agreed with calculated values within ± 0.4%.
This work was supported by the Graduate Training Program (Graduiertenkolleg) GRK 760 “Medicinal Chemistry: Molecular Recognition–Ligand-Receptor Interactions” of the Deutsche Forschungsgemeinschaft (DFG).
Conflict of interest statement
bovine serum albumin
bovine testicular hyaluronidase
bee venom hyaluronidase
- DS-2 cells
Drosophila Schneider-2 cells
- E. coli
Escherichia coli; EDTA, ethylenediaminetetraacetic acid
epidermal growth factor
recombinant human Hyal-1 expressed in DS-2 cells
- Hyal-1–E. coli
recombinant human Hyal-1 expressed in E. coli
ion metal affinity chromatography
sodium dodecyl sulfate-polyacrylamide gel electrophoresis.