INTRODUCTION

Sialidases (EC) hydrolyze -ketosidically linked sialic acids from sialo-glycoconjugates. They are widely distributed in nature and have been isolated from microorganisms (i.e. bacteria, viruses, and protozoa) as well as mammalian tissues (1). Bacterial sialidases have been suggested to initiate microbial infections of animals (2, 3), while mammalian sialidases have been shown to be involved in the catabolism of sialoglycoconjugates (1).

We have reported that the North American leech, Macrobdella decora, contains two sialidases: an ordinary sialidase and an unusual sialidase, sialidase L, which produces 2,7-anhydro-NeuAc instead of NeuAc from various sialoglycoconjugates (4). In addition, this unusual sialidase was found to exhibit a strict linkage specificity toward the hydrolysis of only the NeuAc23Gal linkage (5). This enzyme has been shown to hydrolyze only the NeuAc23Gal linkage in a mixture of sialoglycoconjugates without destroying other sialosyl linkages, such as NeuAc26Gal, NeuAc26GalNAc, NeuAc26GlcNAc, NeuAc28NeuAc and NeuAc29NeuAc linkages. Due to its strict linkage specificity, sialidase L is very useful for studying the biological significance of the NeuAc23Gal linkage and the structural determination of sialoglycoconjugates.

We have purified sialidase L to homogeneity from the Macrobdella leeches and determined the amino acid sequences of nine peptides released by CNBr cleavage and tryptic digestion (5). Two of the tryptic peptides were found to contain the sequence, Ser-X-Asp-X-Gly-X-Thr-Trp, ``Asp box,'' which is conserved and repeated four to five times in many bacterial sialidases (6, 7, 8, 9, 10, 11, 12, 13, 14, 15). This indicates that sialidase L is related to other sialidases. The yield of the pure sialidase L from the original source was only 130 µg/20 kg of leeches (5), which has hampered further biochemical characterization of this novel sialidase. We report here on the isolation of the entire coding region of sialidase L cDNA, using PCR¹-based cloning methods: the rapid amplification of cDNA ends (RACE) (16) and RLM-RACE (17), and its overexpression of functional sialidase L in Escherichia coli. The biologically active recombinant enzyme was purified to an electrophoretically homogeneous form and its specificity toward various sialosyl linkages was studied. We have also compared the primary structure of sialidase L with the previously reported sialidases.

EXPERIMENTAL PROCEDURES

Live M. decora leeches were obtained from St. Croix Biological Supply, Stillwater, MN. For the extraction of the total RNA from M. decora, live leeches were dropped into liquid nitrogen and stored at 70 °C. The following reagents were from commercial sources: CHROMA SPIN-400 gel filtration spin columns, Clontech Laboratories; restriction enzymes, T4 polynucleotide kinase, Superscript II reverse transcriptase, RNase inhibitor, and E. coli DH10B, Life Technologies, Inc.; terminal deoxynucleotidyltransferase and dNTP, Promega; T4 RNA ligase, New England Biolabs; Bluescript II KS(+) vector, Stratagene; Hybond N nylon filters, Amersham Corp.; AmpliTaq DNA polymerase, Perkin-Elmer Corp.; pET-15b expression vector and E. coli BL21 (DE3), Novagen; Fractogel DEAE-650 M and precoated Silica Gel 60 TLC plates, E. Merck (Darmstadt, Germany); octyl-Sepharose CL-4B and protein standards, Pharmacia Biotech Inc.; MU-NeuAc, Tween-20, 3-sialyllactose, 6-sialyllactose, G_D3, and Coomassie Brilliant Blue R-250, Sigma; bicinchoninic acid protein assay reagent, Pierce; and dideoxyadenosine triphosphate, Boehringer Mannheim. Oligodeoxynucleotide primers were synthesized by the Tulane Shared Instrumentation Facility.

Fluorometric assay of NeuAc-cleaving activity using MU-NeuAc as substrate was carried out as described previously (4). One unit of sialidase L is defined as the amount of enzyme which liberates 1 nmol of MU per min at 37 °C. For analyzing the products released from sialoglycoconjugates by TLC, the reaction was carried out in a 1.7 ml Eppendorf tube. The reaction mixture contained 5 to 10 nmol of substrate and 10 units of sialidase in 30 µl of 50 mM sodium acetate buffer, pH 5.5. After incubation at 37 °C for an appropriate time, the reaction was stopped by the addition of 15 µl of 100% ethanol followed by centrifugation to remove the insoluble material. Subsequently, 10 µl-aliquot of this mixture was analyzed by TLC using 1-butanol/acetic acid/H₂O (2:1:1, v/v/v) as the developing solvent. To reveal the glycoconjugates, the plate was sprayed with diphenylamine reagent (18) and heated at 115 °C for 15 min.

Protein concentration was determined by the bicinchoninic acid assay (19) using bovine serum albumin as the standard. SDS-PAGE was carried out by using the discontinuous buffer system described by Laemmli (20).

All reactions were performed in 100-µl reaction volume with 50 pmol of individual primers and 5 units of AmpliTaq DNA polymerase using a GeneAmp PCR system 9600 (Perkin-Elmer). Generally, 35-40 PCR cycles were performed and each cycle consisted of 45-s denaturation at 94 °C, 30-s annealing at 48-53 °C, and 1-min extension at 72 °C. A 5-min predenaturation at 94 °C and an additional 5-min extension at 72 °C were applied before and after the cycle reactions, respectively. Ten microliters of the amplified products were analyzed by agarose-gel electrophoresis. To perform the second round nested PCR, the first round PCR products (5 µl) were used as the templates.

Plasmids were prepared using a Magic (Wazard) miniprep system (Promega). DNA sequencing was carried out by using either Sequenase version 2.0 in standard Sanger-sequencing reactions (U. S. Biochemical Corp.) or by the ``Taq cycling method'' (Epicentre Technologies) according to the manufacturer's instructions.

Unless otherwise indicated, the nucleotide sequences of all the primers used for cloning and expression of sialidase L cDNA are listed in Table I. M1, M2, and M3 were three antisense degenerate primers, corresponding to the T23 internal peptide sequences of PAGIGSSN, MPAGIG, and FADLMPA, respectively (5). P1 was a sense degenerate primer, corresponding to the T18 internal peptide sequence of TDGGNTW (5). Other primers were the sequence specific oligonucleotides made according to the specific cDNA fragments.

Table I. Primers used for the cloning and expression of sialidase L (SL) cDNA Primera Oligonucleotide sequence (5 to 3) Strand Locationb Used for M1 TTNCTICTNCCDATICCNGCIGG   1183-1205 Fragment A RT P1 ACNGAYGGIGGIAAYACNTGG + 991-1012 Fragment A PCR M2 CCDATACCTGCAGGCAT   1180-1196 Fragment A PCR M3 TGCAGGCATYARRTCNGCRAA   1168-1188 Fragment A PCR M4 CTTGTCCTCCAGCAATACCGG   1126-1146 Fragment B RT M5 TTACGCACTTCCCTGGATCTGC   1095-1114 Fragment B PCR Not I M6 AGCTATGTAATCGTCAAAC   1035-1053 Probe M7 ACGCACTGATGCCGGTC   527-543 Fragment C RT M8 TTTATGCCTATGCCGCCAGAG   465-483 Fragment C PCR Not I M9 GTACTGTTTCCTACGCTGAAC   402-422 Probe L1 GATTTTTTTTTTTTTTTTT + Complementary to poly(dA)tail Fragment B and C PCR Xho I T2 GTGGTGTGTGTGCTTAAGAA + Linker tagged to RNA at 3-end T2 TTCTTAAGCACACACACCAC   Complementary to T2 Fragment D RT P2 TTGCCAAAAGCACTGATGGC + 980-999 Fragment D PCR P3 GTTTGACGATTACATAGC + 1035-1052 Fragment D PCR E2 AAGAGAAAATCAGGAACAG + 85-99 SL coding region PCR Nco I M10 ATTGAAGCTTGCCCCGT   2266-2283 SL coding region PCR Bam HI a  Primers P1, M1, M2, and M3 were degenerated and the codes used for the mixed nucleotides were R(A/G), Y(C/T), D(A/G/T) and I (inosine). b  Corresponding nucleotide residues in the cloned sequence shown in Fig. 2.

Leech total RNA was isolated from whole leech by the low temperature guanidinium thiocyanate extraction method (21). For RT-PCR, 5 µg of RNA were heated at 70 °C for 10 min and then mixed with 10 µl of 5 × reverse transcription buffer (Life Technologies, Inc.), 5 µl of 0.1 M dithiothreitol, 5 µl of 10 mM dNTP mixture, 1.5 µl of RNase inhibitor (40 units/µl), 100 pmol of the primer M1, and 50 units of Superscript II reverse transcriptase in a total volume of 50 µl. The reaction mixture was incubated at 45 °C for 1 h, followed by heating at 95 °C for 5 min to inactivate the reverse transcriptase and partially degrade the RNA. Ten microliters of this ``first strand cDNA'' mixture were then used for a first round of standard PCR using primers P1 and M2. The second round nested PCR utilized 5 µl of the first round product with the primers P1 and M3. After the second round nested PCR, three distinct bands were resolved on an agarose-gel. These three PCR-amplified DNA fragments were individually subcloned into Bluescript II KS(+) vectors at SmaI site, and characterized by sequence analysis. The resulting 198-bp DNA fragment (fragment A) was identified to be the cDNA derived from sialidase L. The other two PCR products were identified to be the nonspecific amplification products.

Cloning of the 5 End of Sialidase L cDNA by RACE

To clone the 5 terminal segment of sialidase L cDNA, a modified 5-RACE method was used. Two sequence specific antisense oligonucleotides (M4 and M5) based on the nucleotide sequence of fragment A were synthesized. The primer M5 was designed to contain a NotI site at the 5 end, which facilitated subsequent subcloning. The first strand cDNA was synthesized from leech total RNA using M4 primer under the same conditions as described for RT-PCR. After reverse transcription, the reaction mixture was passed through a CHROMA SPIN-400 column to remove free dNTP and excess primers. The 3 ends of the reverse transcript were tailed with homo-poly(dA) by terminal deoxynucleotidyltransferase at 37 °C for 15 min in a total volume of 100 µl consisting of 1 × reaction buffer (Promega), 0.2 mM dATP, and 40 units of terminal deoxynucleotidyltransferase. Ten microliters of the above reaction mixture were used for PCR with M5 primer and a poly(dT) oligonucleotide L1. The DNA products were digested with XhoI and NotI, passed through a CHROMA SPIN-400 column, and the entire eluate was used for subcloning the DNA in Bluescript II KS(+) vector between XhoI and NotI sites. Transformations were accomplished by electroporation using E. coli strain DH10B as host and the plasmid-bearing cells were selected on ampicillin-containing LB plates. An antisense oligonucleotide, M6, which is complementary to the nucleotide sequence near the 5-terminal segment of fragment A, was radiolabeled at its 5 end and used as the probe to screen colonies of transformed cells. Of 2,000 colonies screened, 83 positives were found. The size of the DNA insert in each positive clone was examined by PCR using the primers L1 and M6. A single bacterial colony was used in each PCR with one fifth of the standard PCR volume (20 µl). Two clones that contained the largest DNA insert were characterized by sequence analysis. Both contained the same nucleotide sequence (fragment B) and were identified to be the cDNA derived from sialidase L.

Based on the nucleotide sequence of fragment B, we synthesized two additional oligonucleotides (M7 and M8) that facilitated isolation of cDNA for the region upstream of fragment B. The first strand cDNA was reverse transcribed from leech total RNA using primer M7, followed by poly(dT) tailing. The product was subsequently subjected to PCR with primers L1 and M8 followed by cloning in Bluescript vectors and E. coli DH 10B. The radiolabeled antisense oligonucleotide M9, which is complementary to the nucleotide sequence near the 5 end of fragment B, was used to screen transformed cells by colony hybridization. The size of the DNA insert in each positive clone was examined by PCR with primers L1 and M9. Three clones that contained the largest DNA insert (fragment C) were then characterized by sequence analysis and proved to be the cDNA derived from sialidase L. The presence of an in-frame upstream translation stop codon, TAA, in fragment C indicated the completion of the 5 end coding region of sialidase L cDNA (cf. Fig. 2).

Cloning of the 3 End of Sialidase L cDNA by RLM-RACE

A linker primer, T2, was phosphorylated at its 5 end by T4 polynucleotide kinase and dideoxy adenosine monophosphate added to its 3 end by terminal deoxynucleotidyltransferase. The product was then ligated to the leech total RNA in 80 µl of reaction mixture containing 10 µg RNA, 100 pmol modified T2, 100 units of T4 RNA ligase, 40 units of RNase inhibitor, 2 µg of bovine serum albumin, and 1 × ligation buffer (New England Biolabs). After incubation at 16 °C for 8 h, the reaction mixture was passed through a CHROMA SPIN-400 column and the eluate was precipitated in ethanol, redissolved in water, and used for reverse transcription with primer T2, which is complementary to T2. Conditions for reverse transcription were as described for RT-PCR. Five microliters of the reaction products were used for two rounds of nested PCR amplification with the two paired primers, P2 with T2 and P3 with T2, respectively. After two rounds of PCR amplifications, a 1.5-kb PCR product (fragment D) was obtained and identified to be the cDNA derived from sialidase L.

Fragments A, B, C, and D (which covered the complete coding region of sialidase L cDNA) were used to construct a single sialidase L cDNA sequence by standard restriction endonuclease digestions and ligation reactions (22). The resulting 2524-bp DNA construct carrying the contiguous nucleotide sequence from the 5 end of fragment C through the 3 end of fragment D was subcloned into Bluescript II KS(+). For expression of sialidase L, a cDNA construct (E2/M10) was amplified from the 2524-bp DNA by PCR using the primers E2 and M10. The construct E2/M10 consists of the coding region starting from the 29th amino acid residue Glu through the translation termination codon. The PCR product was digested with NcoI and BamHI and subcloned into the same sites of the expression vector pET 15b (Novagen) to create a NcoI site immediately to the ATG initiation codon at the 5 end and a BamHI site 11 bp in frame with the TAA termination codon of the pET 15b vector at the 3 end. The recombinant protein was obtained by expressing the construct in E. coli strain BL21 (DE3) grown in LB medium containing 50 µg/ml ampicillin at 37 °C to an optical density of 0.7 at 600 nm, before induction with 2 mM of IPTG for 2 h.

Unless otherwise indicated, all of the operations were carried out at 4 °C. The cells from 4 liters of the IPTG-induced culture containing the sialidase L expressed by the construct E2/M10 were harvested by centrifugation at 5,000 × g for 30 min, washed with 50 mM sodium phosphate buffer (pH 7.0), and resuspened in 50 ml of the same buffer containing 1 mM each of EDTA and phenylmethylsulfonyl fluoride. The cell suspension was then passed twice through a chilled French pressure cell (SLM Aminco) at 16,000 p.s.i. Tween-20 (10% in water, w/v) was mixed gently into the cell lysate to yield a final concentration of 0.25% at room temperature. The solution was first centrifuged at 30,000 × g for 30 min, and the supernatant was further clarified by centrifugation at 100,000 × g for 30 min in a Beckman XL-90 ultracentrifuge equipped with a 50.2-Ti rotor. The supernatant from ultracentrifugation was applied to a Fractogel DEAE column (2.5 × 14 cm) that had been equilibrated with 50 mM sodium phosphate buffer, pH 7.0, at a flow rate of 60 ml/h. After washing with the same buffer, the recombinant sialidase L was eluted with 600 ml of a linear NaCl gradient from 0 to 0.5 M in the same buffer, and 7-ml fractions were collected. The fractions containing MU-NeuAc-cleaving activity were pooled and passed through an octyl-Sepharose column (1.5 × 13 cm) that had been equilibrated with 0.2 M sodium phosphate buffer, pH 7.0, at room temperature. The column was washed with the same buffer at 50 ml/h. The MU-NeuAc-cleaving activity in the unadsorbed fractions was pooled, concentrated to 25 ml by ultrafiltration using Amicon stirred cell with a PM10 membrane, and analyzed by SDS-PAGE.

RESULTS AND DISCUSSION

Initially, the M. decora leech cDNA library constructed in ZAP II system (Stratagene) was screened with the DNA probe derived from the RT-PCR product (fragment A) of sialidase L. By screening approximately 2 × 10⁶ phages from the leech cDNA library, no positive clone was detected. Northern blot analysis also did not reveal any positive signal. These results indicate that the sialidase L mRNA is in very low abundance in the leech. A combination of the PCR-based methods was therefore used to clone a 2,524-bp composite cDNA of sialidase L as described under ``Experimental Procedures.'' A detailed restriction map of sialidase L cDNA and the locations of the four cDNA fragments (fragment A, B, C, and D) on the cDNA sequence are shown in Fig. 1.

Fig. 2 shows the nucleotide sequence of the composite cDNA and the deduced amino acid sequence of sialidase L. No poly(A)⁺ tail was found at the 3 end. The 2,286-bp open reading frame encodes 762 amino acid residues with a calculated molecular mass of 82,982 kDa, which is very close to the value of 84 kDa estimated from the native protein by SDS-PAGE (5). All nine peptide sequences derived from the native protein by CNBr cleavage and tryptic digestion (5) were found in the deduced amino acid sequence of sialidase L. The Asp box, Ser-X-Asp-X-Gly-X-Thr-Trp, which is conserved in reported bacterial and mammalian sialidases, was repeated at four positions at amino acid residues 330-337, 513-520, 573-580, and 622-629 of the deduced amino acid sequence of sialidase L (Fig. 2). The function of the Asp box is still not known. The crystal structure of Salmonella typhimurium sialidase indicates that the Asp boxes are exposed at the sialidase surface, away from the active site, and thus may be important for protein secretion or folding (23). Another stretch of conserved sequence found in microbial sialidases, the ``FRIP region'' (24), which consists of a tetrapeptide sequence, Phe-Arg-Ile-Pro, located on the upstream region from the first Asp box. This motif is also present at the corresponding position of the deduced amino acid sequence of sialidase L at residues 292-295 (Fig. 2). The Arg and Pro residues of the FRIP region are always conserved among microbial sialidases (24). Substitution of Arg³⁷ by Lys in C. perfringens sialidase causes significant change in K_m, and V_max values, as well as K_i with the sialidase inhibitor 2-deoxy-2,3-dehydro-NeuAc, suggesting the involvement of this amino acid residue in substrate binding (25). As shown in Fig. 3, the Kyte-Doolittle hydropathy plot of sialidase L suggests that the N-terminal sequence of this enzyme could be a signal sequence that directs secretion. Analysis of this putative signal sequence by the PC/Gene program reveals a possible peptide cleavage site between amino acid residues 28 and 29. In addition, this N-terminal region resembles the putative peptide sequence of Clostridium septicum sialidase (8) (cf. Fig. 6).

Alignment of the amino acid sequence and the predicted -strands of sialidase L (

Sialidase L cDNA construct E2/M10 devoid of the putative signal peptide sequence was overexpressed at a level of about 100 mg/liter of the culture using a T7 RNA polymerase expression vector system. The sialidase L activity was found to be exclusively in the cytosol of E. coli. Fig. 4 shows the SDS-PAGE analysis of the proteins in the uninduced cell lysate (lane 2), the IPTG-induced cell lysate (lane 3), and the purified recombinant sialidase L after octyl-Sepharose chromatography (lane 4). The molecular size of the recombinant sialidase L as shown by SDS-PAGE was estimated to be 84 kDa, which agrees well with that of the native enzyme. The specific activity of the homogeneous recombinant sialidase L was also found to be comparable to that of the native enzyme.

Fig. 5 shows the specificity of the recombinant sialidase L toward the natural substrates with three different sialosyl linkages. As in the case of the native enzyme, the recombinant sialidase L hydrolyzed only the NeuAc23Gal linkage in 3-sialyllactose, and the cleavage product was 2,7-anhydro-NeuAc (Fig. 5, lane 4). The NeuAc26Gal linkage in 6-sialyllactose and the NeuAc28NeuAc linkage in G_D3 were completely resistant to this enzyme (Fig. 5, lanes 7 and 10). These results indicate that the recombinant sialidase L possesses the strict NeuAc23Gal linkage specificity and releases 2,7-anhydro-NeuAc as the cleavage product.

TLC showing the hydrolysis of 3-sialyllactose, 6-sialyllactose, and G

Since sialidase L is derived from a eukaryotic source, the primary structure of sialidase L was first compared to two mammalian sialidases, those from rat skeletal muscle cytosol (27) and Chinese hamster ovary cells (28). We observed that sialidase L shares little primary sequence identity (<10%) with these two mammalian sialidases. The only common features among the three sialidases are the consensus repeat Asp box and the conserved FRIP motif.

In contrast, sialidase L shares several features in common with sialidases from bacterial and protozoan sources (24). Fig. 6 shows the alignment of the amino acid sequence of sialidase L with two bacterial sialidases, those from C. septicum and S. typhimurium. Among all reported sialidases, the C. septicum sialidase shows the highest sequence identity (18%) with sialidase L. In the region between the FRIP motif and the first Asp box, sialidase L exhibits a remarkable sequence identity (~50%) with the bacterial sialidases. In addition, the distances between the four Asp boxes of sialidase L (175, 52, and 41 residues apart, respectively) are very similar to the Asp box spacing in many reported bacterial sialidases (29). Furthermore, several single amino acid residues Gly³⁰³, Arg³¹², Val⁴³⁹, Asp⁴⁴³, Tyr⁵⁷², Glu⁵⁹⁵, Glu⁶⁰⁰, Arg⁶¹¹, Arg⁶⁷³, Tyr⁷⁰², Tyr⁷¹³, Leu⁷¹⁶, and Glu⁷²⁹ in the deduced amino acid sequence of sialidase L are conserved at the corresponding positions in the sequences of bacterial and protozoan sialidases which were aligned by Roggentin et al. (24). Some of these conserved amino acid residues (marked by asterisks in Fig. 6) have been shown to be involved in the active site of the S. typhimurium sialidase (23). Crystallographic studies on the active site of S. typhimurium sialidase reveals that the arginine triad which consists of Arg³⁷ in the FRIP region, Arg²⁴⁶ and Arg³⁰⁹, interacts and stabilizes the carboxylic acid group of all natural sialic acid derivatives (23). It is noteworthy that these three Arg residues are also conserved in sialidase L (Arg²⁹³, Arg⁶¹¹, and Arg⁶⁷³, respectively) and may have a similar role to their counterparts in S. typhimurium sialidase. Another interesting observation is that the conservation of Tyr⁷¹³ in sialidase L which corresponds to Tyr³⁴² in the active site of S. typhimurium sialidase. This tyrosine residue in the active site of Salmonella sialidase has been proposed to stabilize the carbonium ion transition state intermediate, since the tyrosine hydroxyl group is very close to the C1 and C2 carbons of the sialic acid derivative (23). The conservation of this tyrosine residue in sialidase L may imply that the transition state for the formation of 2,7-anhydro-NeuAc by this enzyme also involves a carbonium ion intermediate.

The sequence alignment shown in Fig. 6 suggests that sialidase L has two domains, an N-terminal domain and a catalytic domain. The latter aligned well with the S. typhimurium sialidase, including all the conserved single amino acid residues, FRIP region, and Asp boxes. The secondary structure of the catalytic domain of sialidase L, as predicted by the method of Garnier et al. (30), shows -strands positioned near the conserved sequence motifs containing the catalytic residues in an arrangement similar to that observed in the crystal structure of Salmonella sialidase (Fig. 6) (23). Possibly, sialidase L has a similar folding topology in the catalytic domain as that of the Salmonella enzyme. Analysis of the primary and the predicted secondary structures of sialidase L indicates that this enzyme shows many characteristic features of the bacterial sialidase and that it may share a three-dimensional structure with this group of sialidases.

The formation of 2,7-anhydro-NeuAc by sialidase L requires a conformational change of the pyranose ring from ²C₅ to ⁵C₂ during the transition state (31). The ⁵C₂ conformation is thermodynamically unstable since all the substituents are at the axial positions of the pyranose ring except the carboxyl function. It is still unclear how this enzyme brings about the necessary conformational change of the pyranose ring and catalyzes the formation of the intramolecular glycosidic linkage of 2,7-anhydro-NeuAc. Interestingly, the potent competitive inhibitor of microbial and mammalian sialidases 2-deoxy-2,3-dehydro-NeuAc has very little inhibitory effect on the activity of sialidase L (5). Thus, some aspects of the mechanism of action of sialidase L must be different from the ordinary sialidase. The biological function of 2,7-anhydro-NeuAc is still unknown. The availability of recombinant sialidase L should facilitate our understanding of the structure, function, and mechanism of action of the enzyme as well as allow the exploration of the biological roles of 2,7-anhydro-NeuAc and the NeuAc23Gal linkage in sialoglycoconjugates.