Rabbit polymorphonuclear neutrophils form 35S-labeled S-sulfo-calgranulin C when incubated with inorganic [35S]sulfate.

Rabbit peritoneal polymorphonuclear neutrophils reduced inorganic [35S]sulfate to [35S]sulfite in vitro, concomitant with incorporation of 35S into a 10.68-kDa cytosolic protein as a S-[35S]sulfo-derivative. Amino-terminal sequencing of the purified protein identified calgranulin C, a member of the S100 protein family. cDNA clones of calgranulins B and C were isolated using oligonucleotide primers based on the established amino acid sequences of other mammalian calgranulins. The complete amino acid sequence of rabbit calgranulin C was deduced from the nucleotide sequence of the corresponding cDNA. It comprises 91 amino acid residues, has a calculated molecular mass of 10.52 kDa, has 74% identity with porcine calgranulin C, and shows high homology with other S100 calcium-binding proteins. Rabbit calgranulin C has a single cysteine residue at position 30, which we believe to be modified to S-[35S]sulfo-cysteine as a consequence of sulfate reduction by neutrophils. The formation of S-[35S]sulfo-calgranulin C appears to be a reaction specific to neutrophils. The specific radioactivity of calgranulin C from the neutrophil culture medium was 50-fold greater than that of the calgranulin C within the cells, suggesting that S-sulfation of calgranulin C might be associated with its secretion.

Sulfate reduction to sulfite has been well characterized in microorganisms (1,2) where the reaction is catalyzed by an enzyme, 3Ј-phosphoadenosine-5Ј-phosphosulfate-reductase, and at least one carrier protein (1). Sulfate reduction in mammalian tissues is less common (3,4), and little is known about its precise mechanism or biological function.
In a previous paper (5) we reported that rabbit polymorphonuclear neutrophils (PMN) 1 reduce inorganic sulfate to inor-ganic sulfite. We have proposed that PMN may form sulfite as an anti-bacterial agent, as an agent for cleaving disulfide bonds in proteins to facilitate proteolysis, and even as an agent for changing the antigenicity of proteins for recognition by the immune system (5). Although a large proportion of the reduced [ 35 S]sulfate was released in the form of inorganic [ 35 S]sulfite, polyacrylamide gel electrophoresis (PAGE) revealed the presence of a component that migrated with an apparent molecular mass of less than 14 kDa. 35 S label could be displaced from this macromolecule by treatment with sodium sulfite, consistent with the presence of an S-[ 35 S]sulfo-derivative.
The formation of a S-sulfo-protein might be expected to arise as a consequence of sulfite production because sulfite is known to react nonenzymically with disulfide bonds in proteins in a reversible, nonspecific manner under physiological conditions without requirement for enzyme catalysis (6). In this reaction the disulfide bond is cleaved forming a S-sulfo-derivative and a free thiol.
Because this reaction is reversible, the detection and isolation of S-[ 35 S]sulfo-proteins requires elimination of free thiols from the reaction mixture. This has been achieved by addition of excess N-ethylmaleimide (NEM), which reacts rapidly with free thiols and with inorganic sulfite (7).
Formation of a small protein associated with the majority of the protein-bound [ 35 S]sulfite in PMN cells raises the possibility that this protein might be involved in the sulfite formation mechanism. In this paper we describe the isolation and purification of this protein and its identification as the rabbit homolog of the porcine granulocyte protein calgranulin C, a member of the S100 protein family (8). Calgranulins A, B, and C are specifically expressed in neutrophils and monocytes, although calgranulins A and B are also expressed in certain cells of epithelial lineage (9 (19.1 Ci/g) was from Amersham Corp. Staphylococcus aureus V8 protease, N-ethylmaleimide, low melt agarose, sodium EDTA, phenylmethylsulfonyl fluoride, benzamidine hydrochloride, 6-aminohexanoic acid, N-dansylaziridine, and N-lauroylsarcosine were from Sigma, polyvinylidine difluoride (Immobilon-P) membranes from Millipore; Hema IEC BioQ resin (10 m) was from Alltech; PepRPC (C 18 ) HR 5/5 and Mono-Q HR5/5 columns, and Sephadex G-25 and G-50 were from Pharmacia Biotech Inc.; and molecular mass standards for electrophoresis were from Novex (New South Wales, Australia).
Cell Isolation and Culture-PMN were harvested from the peritoneal cavity of adult female laboratory rabbits of mixed breeds as described by Cohn and Hirsch (10). Rabbit PMN cells (10 7 -10 8 cells/ml) were incubated at 37°C with 0.1 mM [ 35 S]sulfate (1000 Ci/mol) in Hanks' balanced salt solution (11) containing 0.5 mM MgCl 2 in place of MgSO 4 and supplemented with 5.5 mM glucose as described previously (5).
For identification of S-[ 35 S]sulfo-proteins, rabbit PMN were incubated with [ 35 S]sulfate for 1 h and then disrupted by sonication. The lysate was divided into two equal portions; one was treated immediately with excess NEM (final concentration 50 mM), while the other was treated with 25 mM Na 2 SO 3 at room temperature for 30 min before the addition of NEM. These preparations were analyzed for 35  Purification of Calgranulin C-Rabbit PMN (10 8 cells/ml) were incubated for 4 h with [ 35 S]sulfate as described above. The incubation mixture was chilled in ice and sonicated in the presence of proteinaseinhibitors (final concentrations, 10 mM sodium EDTA, 0.1 mM phenylmethylsulfonyl fluoride, 5.0 mM benzamidine hydrochloride, 10 mM 6-aminohexanoic acid, and 50 mM NEM). After centrifugation at 5,000 rpm for 10 min at 4°C, the PMN extract was fractionated by gel filtration on a column (2 ϫ 105 cm) of Sephadex G-50 (Pharmacia, fine grade) equilibrated with 0.1 M NaHCO 3 . Elution was performed at a flow rate of 20 ml/h. Eluted fractions (5.0 ml) were assayed for radioactivity, and protein concentration was determined using the Bradford protein assay (Bio-Rad). Specific fractions were analyzed by SDS-PAGE. Fractions containing the major 35 S-labeled protein (see Fig. 3A, fractions 21-25) were pooled, diluted 2-fold, and subjected to anionexchange chromatography on a Mono Q HR 5/5 column (Pharmacia), equilibrated with 0.01 M NaHCO 3 , pH 8.0. The column was eluted at a flow rate of 1.0 ml/min with a linear concentration gradient of 0.01-1.0 M NaHCO 3 as indicated in the legend of Fig. 4. Fractions of 1.5 ml were collected and assayed for radioactivity and protein content. The major peak of 35 S-labeled protein (see Fig. 4A, fractions 52-56) was pooled and desalted by reverse-phase chromatography using a PepRPC (C 18 ) HR 5/5 column (Pharmacia) eluted at a flow rate of 1.0 ml/min with a gradient formed from solvent A (0.1% trifluoroacetic acid in water) and solvent B (0.1% trifluoroacetic acid in 80% acetonitrile, 20% water) as follows: 1% solvent B for 5 min, 1-100% solvent B in 60 min, and 100% solvent B for 5 min. Fractions of 1.0 ml were collected and assayed for radioactivity.
Detection of S-Sulfo-cysteine-A sample of the 35 S-labeled protein, recovered from anion-exchange chromatography on Mono Q, was digested with 10 mg of Pronase at 50°C for 24 h in 1.0 ml of 0.1 M Tris-HCl, pH 8.0, containing 1.0 mM CaCl 2 . An aliquot of the digested sample was mixed with standard S-sulfo-cysteine and subjected to anion-exchange HPLC on a column (4.6 ϫ 300 mm) of Hema IEC BioQ eluted isocratically with 0.2 M triethylamine acetate in 0.2 M acetic acid.
Enzymic Digestion and Peptide Purification-A sample of purified 35 S-labeled protein (0.32 mg) recovered from reverse-phase chromatography on PepRPC (Pharmacia), was lyophilized and resuspended in 0.1 M NH 4 HCO 3 , 2 mM EDTA, pH 8.0. Digestion was carried out with 12 g (10 units) of S. aureus V8 protease (Sigma), at 37°C for 24 h. The reaction was terminated by freezing and lyophilizing the reaction mixture. The resultant peptides were fractionated by reverse-phase chromatography on a PepRPC column as described above.
Electrophoresis-SDS-PAGE in Tricine buffer was performed under nonreducing conditions on a 16% gel in a Mini-PROTEIN II apparatus (Bio-Rad) as described by Schä gger and von Jagow (12). Samples for electrophoresis were dissolved in 68 mM Tris-HCl buffer, pH 6.8, containing 0.2% (w/v) SDS, 10% (v/v) glycerol, and 0.01% (w/v) bromphenol blue. The resultant gels were stained with Coomassie Brilliant Blue or silver and were analyzed for radioactivity with a PhosphorImager (Molecular Dynamics) or by fluorography. Fluorography was carried out by exposure to Kodak X-Omat AR film (Kodak, Melbourne, Australia) for 7-8 days at Ϫ80°C.
Fluorescence Measurements-Proteins labeled with the fluorescent reagent N-dansylaziridine were detected with a Hitachi (Tokyo, Japan) Fluorescence Spectrophotometer Model F-4000 using excitation at 345 nm and detection of emission at 510 nm (13). Fluorescent proteins were detected on gels after SDS-PAGE using a Spectroline UV transilluminator (Model TVC-312A, Spectronics) at 312 nm.
Mass Spectrometry-The molecular mass of calgranulin C was determined using an API III LC/MS system equipped with a triple quadropole ionspray mass spectrometer (Perkin-Elmer Sciex). Samples were run at an orifice plate voltage of 55 V, an ionspray voltage of 0.5 kV, a gas (N 2 ) flow rate of 0.6 ml/min, and a liquid flow rate of 30 l/min.
Amino Acid Analysis and Sequencing-Aliquots of appropriate fractions from reverse-phase chromatography on PepRPC were subjected to amino acid sequencing in an Advance Biosystems gas phase sequenator to determine their amino-terminal sequences. The sequences were then compared with the available protein data bases. Amino-terminal sequencing of N-dansylaziridine-labeled proteins after SDS-PAGE was performed by transfer of the proteins to a polyvinylidene difluoride membrane (Millipore) by electroblotting. Fluorescent bands were excised from the membrane and subjected to amino acid sequencing as described above.
Oligonucleotide Primers-The sequence of the primer used for PCR amplification was 5Ј-ATCATCAA(C/T)ATCTTCCATCA(G/A)TAC-3Ј corresponding to the amino-terminal positions 10 -17 of calgranulin C. The primer included 4-fold degeneracy in order to accommodate multiple codon possibilities for two of the amino acids on which the primer sequence was based.
Cloning of Rabbit Calgranulin C and B cDNA-Total cellular RNA was extracted from rabbit PMN (8 ϫ 10 8 cells) using a modification of the method of Chomczynski and Sacchi (14). cDNA samples were prepared using 2 g of total cellular RNA as a template in a reaction mix of 20 l that included Moloney murine leukemia virus reverse transcriptase (Promega, 100 units), RNAsin (Promega, 20 units), 50 M dNTP, and 1 M 30-mer R1T primer (5Ј-AAGTCGACGAATTCCA-GAT 12 -3Ј). The reaction was incubated at 40°C for 1 h, treated with RNase H at 37°C for 30 min, and stored at Ϫ20°C for later use.
The amino-terminal primer of calgranulin C and an 18-mer R1T primer (5Ј-AAGTCGACGAATTCCAGA-3Ј) were used in standard PCR reactions with 1 l of single-stranded cDNA as template in a reaction mixture of 30 l of 10 mM Tris-HCl buffer, pH 8.8, containing 2 mM MgCl 2 , 1.25 units of Taq polymerase (Promega), 200 mM mixed dNTP together with each primer at a concentration of 1 M. Optimal reaction conditions were: initial cycle of denaturation at 95°C for 2 min, annealing at 52°C for 1 min, and extension at 72°C for 1.5 min, followed by 29 cycles of denaturation at 95°C for 40 s, annealing at 52°C for 1 min, and extension at 72°C for 1.5 min. This was then followed by a single 5-min extension cycle at 72°C.
The products were subjected to electrophoresis through a 2% low melt agarose gel (Sigma) in 50 mM Tris-acetic acid buffer, pH 7.4, containing 1.0 mM EDTA. Two prominent bands of approximately 400 and 500 base pairs were excised and DNA recovered using a Wizard DNA-prep purification column (Promega). The purified DNA preparations were cloned directly into the pGEM-T vector (Promega). The ligation products were used to transform competent Escherichia coli cells (strain DH5␣). Plasmid DNA was subsequently harvested from transformants using a Wizard miniprep column (Promega).
Nucleotide Sequencing of cDNA Inserts-Clones containing the 400and 500-base pair inserts were identified by restriction enzyme digestion of the miniprep DNA with PvuII (Progen). Nucleotide sequences were determined by double-stranded sequencing using the dideoxy chain termination method of the TaqTrack or fmol Sequencing system (Promega). Preliminary sequence determination was done with the universal and reverse sequencing primers. The complete sequence of inserts was obtained using the 18-mer R1T primer and the aminoterminal primer of calgranulin C. All sequences were confirmed with at least two preparations of cDNA synthesized from independent PCR amplifications.  Proteins in the two preparations were separated by SDS-PAGE and subjected to fluorography (Fig. 2B). A number of 35 S-labeled proteins was observed in the untreated preparation (Fig. 2B, lane 2); however, densitometric scanning of the fluorogram showed that 83% of the radioactivity was associated with a protein that migrated with an apparent molecular mass of 8 kDa. After sodium sulfite-treatment, this protein and another protein of 16 kDa were shown to have lost their 35 S label (Fig. 2B, lane 3), whereas a 31-kDa protein and other proteins of even larger molecular size retained their label.

Kinetics of [ 35 S]Sulfite and [ 35 S]Macromolecule Formation in
The ability of neutral sodium sulfite to displace the 35  Purification of the 35 S-Labeled Protein-Rabbit PMN (6.3 ϫ 10 7 cells/ml) were incubated with [ 35 S]sulfate for 4 h at 37°C, sonicated, treated with excess NEM, and centrifuged. The supernatant was fractionated by gel filtration on a Sephadex G-50 column, which resolved the 35 S-labeled macromolecules into two major peaks (Fig. 3A), well separated from unincorporated label. SDS-PAGE of the two peaks of protein (Fig. 3B,  lanes 2 and 3) and analysis of the resultant gels for radioactivity using a PhosphorImager established that the highly radiolabeled 8-kDa protein was associated with the second peak (data not shown).
Fractions from Sephadex G-50 containing the 35 S-labeled 8-kDa protein were combined and subjected to ion-exchange chromatography on a Mono Q-HR 5/5 column as described under "Experimental Procedures." Three major peaks of protein (A, B, and C) were eluted by the salt gradient (Fig. 4A), but a single major radioactive peak was eluted with 0.4 M NaHCO 3 , just after the last peak of protein (peak C). The peak of radioactivity (fraction 54) did not coincide with the peak of protein (fraction 52), but SDS-PAGE (Fig. 4B) revealed that the 35 Slabeled 8-kDa protein eluted in peak C (fractions 52-56; Fig.  4B, lanes 4, 5, and 6). The two-fraction displacement of the protein profile and the radioactivity profile in the peak C protein (Fig. 4A) was assumed to indicate the slightly more anionic behavior of the S-[ 35 S]sulfated protein compared with the nonsulfated protein. This difference could not be detected after SDS-PAGE of fractions 52-56; the Coomassie Blue staining bands and the PhosphorImager bands coincided and migrated the same distance in each lane.
The fractions containing the 8-kDa protein were pooled and subjected to reverse-phase chromatography on a PepRPC column to remove minor contaminating proteins and salt. Fractions containing the single sharp peak of 35 S-labeled protein, which eluted with 50% acetonitrile (data not shown), were pooled and lyophilized for further characterization and sequence analysis. SDS-PAGE analysis of this preparation revealed a single radioactive band migrating with an apparent mass of 8 kDa.
Protein estimation showed that the labeled protein constituted 8.6% of the total cytosolic (soluble) protein of PMN. The precise molecular mass of the purified protein, determined by Primary Structure of the 35 S-Labeled Protein-The lyophilized protein from the PepRPC purification step was subjected to amino-terminal sequencing. The sequence of the first 25 residues was found to be TKLEDHLEGIINIFHQYSVRTGH-YD. Amino acid comparison with the available data bases revealed a high degree of homology between this protein and several proteins of the S100 family. The best alignment was with a porcine granulocyte protein, calgranulin C with 96% homology within the first 25 amino acid residues (Fig. 5B).
In order to obtain further amino acid sequences of rabbit calgranulin C, fragments were generated by protease V8 digestion of the purified protein. The digest was resolved into three peaks of 35 S label by reverse-phase chromatography on Pep-RPC. SDS-PAGE revealed that the first peak contained no detectable protein, the second peak (fraction 29) contained two peptides of approximately 3.0 and 4.0 kDa (designated V1 and V2), and the third peak (fraction 40) contained undigested material together with a partially digested fragment, whereas fraction 41 contained a single fragment of the same size as that in fraction 40 (designated V3). The amino acid sequences of these peptide fragments were determined by Edman degradation. The amino-terminal residues of V3 (TKLEDHLEGIINIF-HQYSVRTGHYDTLSKXE) were identical to those of undigested calgranulin C; peptide V3 thus results from digestion at the carboxyl end of calgranulin C. The full sequences of V1 (GIINIFHQYSVRTGHYDTLSKXE) and V2 (LVNTIKNTKDQ-ATVDRIFRDLDEDGDHQVDFKE) were assigned by comparing the amino acid residues detected in each sequencing cycle with the sequence obtained from V3. Peptide V1 corresponded to the sequence from residues 9 to 31, and V2 exhibited high homology with residues 40 -72 on porcine calgranulin C (Fig.  5B). No residue could be identified for residue 30 in V3, although a clear but unidentifiable signal was detected by the sequencer. A similar signal was observed for residue 22 in V1. Because V3, and V1 were both labeled with 35 S, we postulated that residue 30 could be the site of S-[ 35 S]sulfo-cysteine. 35 S-Labeled Calgranulin C-To confirm the presence of S-sulfo-cysteine, exhaustive Pronase digestion of purified 35 S-labeled calgranulin C was performed. The digest was subjected to anion-exchange HPLC. 35 S label (45% of the total) co-eluted with the S-sulfocysteine standard (data not shown). Thus [ 35 S]sulfite formed by PMN could become bonded to cysteine residues through a covalent, S-sulfate, or thiol-ester bond.

Identification of S-[ 35 S]Sulfo-cysteine in
Cloning of Rabbit Calgranulin C and Calgranulin B-An oligonucleotide primer was prepared based on the predetermined amino-terminal sequence of purified rabbit calgranulin C. This primer was used, together with a R1T primer, to amplify specific rabbit PMN cDNA by PCR. Two PCR products having the approximate sizes of 400 and 500 base pairs were cloned into the pGEM-T vector by T-A ligation. Positive clones were selected by restriction enzyme digestion, and the nucleotide sequence of the insert was determined. The resultant partial nucleotide sequences, when compared with nucleotide sequences in GenBank/EMBL data bases, indicated that the cDNA clones were of calgranulin C and calgranulin B. The partial nucleotide sequences determined and the deduced amino acid sequences of calgranulin C and B are shown in Fig.  5A. The nucleotide sequence was confirmed using clones obtained from independent PCR trials.
The partial cDNA sequence of rabbit calgranulin C, when combined with the amino-terminal sequence obtained by Edman degradation (Fig. 5A), indicated that the cDNA had an open reading frame of 273 nucleotides, predicting a protein of 91 amino acids with a calculated molecular mass of 10,520 Da. A comparison of the complete amino acid sequence with the porcine counterpart showed an overall amino acid identity of 74%. Notably, residue 30 in rabbit calgranulin C, the site postulated to form a S-sulfo-derivative during sulfate reduction by PMN, was confirmed to be cysteine. The molecular mass of 10.52 kDa calculated for rabbit calgranulin C is less than the value of 10.68 kDa determined by mass spectrometry. The discrepancy of 160 Da most probably arises from the attach-ment of NEM to the single cysteine residue because the elution profile from Mono Q (Fig. 4A) indicates that nonsulfated calgranulin C is the major form present. A less likely possibility is bound Ca 2ϩ because, using the 45 Ca 2ϩ overlay technique of Maruyama et al. (15), we have demonstrated (data not shown) that purified rabbit calgranulin C binds calcium, even after reaction with NEM and SDS-PAGE.
The partial nucleotide sequence of rabbit calgranulin B had 73, 71, and 70% identity to the nucleotide sequences of human (16), mouse (17), and rat (18) calgranulins B, respectively. The putative amino acid sequence thus determined for rabbit calgranulin B (Fig. 5A) had 70, 65, 61, and 57% identity with the corresponding proteins of bovine, human, mouse, and rat granulocytes, respectively. Fig. 5B shows the alignment of the rabbit sequences with those of some other mammalian calgranulins B and C.
Distribution of 35   The specific radioactivity of the purified calgranulin C isolated from the cell lysate and from the medium was calculated to be 8.87 ϫ 10 4 and 4.40 ϫ 10 6 dpm/mg protein, respectively, corresponding to a 50-fold increase in S-sulfation of the secreted calgranulin C compared with that which remained within the cell. The specific radioactivity of the extracellular calgranulin C was equivalent to 0.021 mol of [ 35 S]sulfite bound/ mol of calgranulin C. The specific radioactivity of the peak (fraction 54) of labeled calgranulin C eluted from ion-exchange chromatography (Fig. 4A) indicated a similarly low level of sulfation (0.017 mol of [ 35 S]sulfite bound/mol). These calculations have made no correction for dilution of the [ 35 S]sulfate with endogenous sulfate but are consistent with the elution profile from Mono Q (Fig. 4A), which indicated that the 35 Slabeled calgranulin C was a minor component of the calgranulin C. The proportion of calgranulin C recovered in the medium accounted for 23% of the total calgranulin C in the preparation (2.6 g/10 6 cells).
Specific S-[ 35 S]Sulfo-labeling on the Cysteine Residue of Calgranulin C by PMN-Rabbit PMN (6.3 ϫ 10 7 cells/ml) were incubated for 2 h at 37°C with 0.1 mM [ 35 S]sulfate (1000 Ci/mol), and then the incubation mixture was treated with 0.1 M N-dansylaziridine in isopropanol to yield a final concentration of 30 mM N-dansylaziridine. The resultant mixture was then incubated at room temperature for 16 h to label free thiol groups in the PMN proteins. Insoluble material was then removed by centrifugation at 5,000 rpm for 10 min at 4°C, and the PMN extract was fractionated by gel filtration on Sephadex G-50 followed by anion-exchange chromatography on Mono Q as described under "Experimental Procedures." Fig. 6 shows the elution profile from Mono Q revealing three peaks of protein coincident with peaks of fluorescence at 510 nm and a single sharp peak of 35 S radioactivity corresponding to the labeled calgranulin C. SDS-PAGE of fractions 26 and 30, the peak fluorescence fractions, revealed the presence of two major protein components (both fluorescent) with apparent molecular masses of 8 and 15 kDa. These bands were electroblotted onto a polyvinylidene difluoride membrane and subjected to aminoterminal sequencing. The 8-kDa component yielded the sequence PTDLENSLNSIISVYHKYSL, which has 75% identity with the amino-terminal sequence of human calgranulin A; a protein that contains cysteine residues (20). The amino terminus of the 15-kDa component seemed to be blocked, and this component remains unidentified. Separate experiments established that both of these components were labeled by reaction with [ 14 C]NEM (data not shown). A purification procedure similar to that described above was used to isolate the 14 Clabeled macromolecules. SDS-PAGE revealed two major 14 Clabeled components of 8 and 15 kDa. Although the 14 C-labeled 8-kDa component was not readily resolved from calgranulin C by SDS-PAGE, it behaved quite differently during anion-exchange chromatography and reverse-phase chromatography (data not shown).
In addition to these proteins, a separate 13-kDa protein with amino-terminal sequence homology to rabbit lysozyme C was also detected in the PMN lysate. Lysozyme C contains four disulfide bonds. These results indicate that apart from calgranulin C, rabbit PMN contain other cysteine-containing, disulfide-bonded proteins. Yet during sulfate-reduction in PMN, calgranulin C appears to be the major protein that becomes labeled in association with the formation of sulfite. Amino-terminal sequence analysis of the labeled protein identified it to be the rabbit homolog of a porcine PMN protein identified as calgranulin C by Dell'Angelica et al. (8). No cysteine residue, however, was present in porcine calgranulin C. To confirm the presence of and locate the cysteine residue(s) in rabbit calgranulin C, molecular cloning techniques were applied to generate a cDNA clone of calgranulin C by PCR. The nucleotide sequence of the cloned cDNA confirmed and extended the amino acid sequence determined by amino-terminal sequencing (Fig. 5A). Residue 30 was shown to be cysteine, and this was the only cysteine residue within the entire calgranulin C structure.
The sequence of rabbit calgranulin C is consistent with the features exhibited by other S100 calcium-binding proteins, which have two structurally conserved and well defined Ca 2ϩbinding sites: the "EF-hands" (19,20) at amino acid positions 18 -31 and 61-72 (Fig. 5B). The least conserved regions of rabbit calgranulin C are at the carboxyl-terminal region and in the hinge region that connects the two EF-hands; yet even these regions are highly conserved between porcine and rabbit calgranulins C (Fig. 5B). The divergence of the amino acid sequences at the hinge and carboxyl-terminal regions may confer functional specificity to each S100 protein (19).
The sequence of the 5Ј-oligonucleotide primer used to isolate a cDNA encoding calgranulin C was based on the highly conserved ␣-helix region common to calgranulins B and C (8), so it came as no surprise when rabbit calgranulin B cDNA was also amplified in the PCR process. The calgranulins B in each species have highly conserved ␣-helix and Ca 2ϩ -binding regions (Fig. 5B). The hinge region of these proteins also shows a high degree of homology. The carboxyl terminus of rabbit calgranulin B has a repeated GHGHGHSH tail not observed in the other mammalian homologs. The nucleotide sequence corresponding to this repeated tail is not a single repeating structure, which makes the possibility of error in the sequence determination seem unlikely. Whether the repeated carboxylterminal GH-rich tail conveys any particular function of rabbit calgranulin B has yet to be determined. It is noteworthy that the amino acid sequence for rabbit calgranulin B contained no cysteine residues.
Comparison of the rabbit calgranulin C with sequences in protein data bases revealed a maximal homology with calgranulin B, which is also referred to as cystic fibrosis antigen, L1 protein, or MRP-14 (16,21,22). Human calgranulin B occurs as a heterodimer with calgranulin A (MRP-8). The formation of the biologically active complex is Ca 2ϩ -dependent (23,24). Although the precise function of calgranulin B and its complex forms has yet to be defined, it has been suggested that the calgranulin A/B complex plays a role in the differentiation and maturation of myeloid cells (17,25,26).
Rabbit calgranulin C comprises about 8.6% of the total granulocyte cytosolic protein, consistent with the value of 8% reported for porcine granulocytes (8). In contrast, a human calgranulin C homolog (27) constitutes just 5% of the total PMN cytosolic protein, far less than human calgranulins A and B, which together constitute 45-49% (21,23). Porcine calgranulin C has been shown to undergo conformational change upon calcium binding and has been implicated in Ca 2ϩ -dependent signal transduction pathways (8), whereas human calgranulin C has been shown to translocate to the PMN plasma membrane, together with calgranulins A and B during stimulation of the cell by Ca 2ϩ -dependent stimuli (27)(28)(29). Because we have been unable to purify S-sulfo-calgranulin C from the unmodified protein, we have no evidence about the effect of S-sulfation on its calcium binding.
Dell'Angelica et al. (8) have used cross-linking experiments to demonstrate that porcine calgranulin C occurs as a homodimer. They proposed that the association in the homodimer was noncovalent in nature. Many S100 proteins have been reported to exist as dimers, both in disulfide-bonded (30,31) or noncovalently associated forms (23). Because rabbit calgranulin C does contain a cysteine residue, it could form a disulfidebonded homodimer. There is, however, no direct evidence for disulfide bonding, but the formation of a S-sulfo-derivative is consistent with that possibility. Furthermore the result of fluorography on SDS-PAGE in Fig. 2B suggests that two S-[ 35 S]sulfo-proteins were formed by PMN: the major band of calgranulin C with an apparent mass of 8 kDa and a minor band at about 16 kDa. If the latter band was a dimeric form of calgranulin C with 35 S label bonded through a S-sulfo-linkage, then the formation of this dimer does not necessarily involve disulfide bond formation.
S-Sulfation of calgranulin C is a novel reaction, the biological significance of which has yet to be determined. Rabbit calgranulin C appears to act as a major sulfite acceptor in the sulfate reduction mechanism of PMN (Fig. 2B). Whether this protein is directly involved in the sulfate-reducing mechanism, acting perhaps as a sulfite-carrier protein, or the S-sulfation of calgranulin C occurs as the primary event that then leads to sulfite formation as a result of thiol-exchange has yet to be determined. S-Sulfation of calgranulin C in rabbit PMN does, however, appear to be a specific reaction. Lysozyme, which contains disulfide bonds, was also detected in the cytosol of rabbit PMN together with rabbit calgranulin A, which contained a free thiol group; yet these proteins did not become labeled with 35 S in these experiments. In separate experiments we have shown that exogenous proteins, including bovine serum albumin (5) and egg white lysozyme, do become labeled with 35 S when incubated with PMN, but in cell-free experiments egg white lysozyme does not react with 1.0 mM inorganic [ 35 S]sulfite. These results suggest that PMN products, such as calgranulin C, may be involved in the labeling mechanism. Because we have not purified S-sulfo-calgranulin C, we have no evidence about its ability to transfer 35 S label to other proteins.
During incubation of PMN with [ 35 S]sulfate, 23% of the total calgranulin C was recovered in the medium of the culture. The specific radioactivity of the 35 S-labeled calgranulin C in the medium was found to be 50-fold greater than that which remained in the cell. This result could reflect the reducing environment, maintained by the ratio of reduced to oxidized glutathione, within the PMN cytoplasm. A reducing environment would displace [ 35 S]sulfite from cytoplasmic S-[ 35 S]sulfo-calgranulin C, yet the same protein could be more stable within the endoplasmic reticulum and the vesicles of the secretory pathway. Amino-terminal sequencing of calgranulin C isolated separately from the cell pellet and from the culture medium revealed no evidence of proteolytic modification at the amino terminus of the protein, indicating that the protein lacks a leader sequence for translocation into the endoplasmic reticulum. Because the oligonucleotide probe used to isolate calgranulin C cDNA by PCR was based on the amino-terminal sequence of the protein, we cannot exclude the involvement of a leader sequence during its secretion. Calgranulins A and B have also been reported to occur extracellularly (20,26), yet neither contains a signal peptide sequence required for membrane translocation. Intracellular transport of the calgranulins has been claimed to follow a pathway different from the endoplasmic reticulum-Golgi route (32), and it is tempting to speculate that the S-sulfation of calgranulin C might be involved in its secretion. Extracellular calgranulin C could well exhibit cytostatic activity like the calgranulin A and B complex (33)(34)(35).