Characterization of buffalo native pregnancy-associated glycoprotein: mass spectrometry-based glycan composition analysis, sugar-binding characteristics and proteolytic activity assay

Pregnancy-associated glycoproteins constitute a large family of extensively glycosylated secretory proteins which are specifically expressed in outer epithelial cell layer (chorion/trophectoderm) of placenta. The present study reports purification and characterization of buffalo pregnancy-associated glycoprotein (BuPAG) from foetal cotyledons during early pregnancy (< 3 months). Native BuPAG was purified by anion-exchange and Vicia villosa affinity chromatography. Three distinct bands were observed in SDS-PAGE and Western blot with higher molecular weight range of ~ 55–70 kDa in comparison to the calculated one (~ 40 kDa) which may be attributed to the presence of glycosylation in BuPAG. Glycan analysis of purified BuPAG by high-resolution mass spectrometry (ESI-qTof) revealed the presence of nine N-glycan structures and three O-glycan structures. Interaction analysis of various sugars with purified BuPAG using fluorescence quenching assay showed decrease in fluorescence intensity without any shift in emission maxima. BuPAG exhibited proteolytic activity within pH range of 2.0–5.5 with highest activity at pH 3.0. In the presence of inhibitor pepstatin A, proteolytic activity reduced by 84%. The glycan structures and sugar-binding characteristics of BuPAG are identified for the first time in this study. The activity of BuPAGs as active peptidases indicates their possible role in various molecular pathways required for sustenance of pregnancy.


Introduction
The mammalian placenta synthesizes and secretes a wide variety of steroid and peptide hormones including various proteins which have been found associated with the pregnancy. Pregnancy-associated glycoproteins (PAGs), also known as pregnancy-specific protein B (PSPB), constitute a large family of glycoproteins specifically expressed in outer epithelial cell layer (chorion/trophectoderm) of the placenta in several eutherian species (Garbayo et al. 2000;Green et al. 2000;Zoli et al. 1992a). They were first identified as novel antigens by producing antisera against homogenates of whole bovine placenta (Butler et al. 1982). PAGs belong to aspartic proteinase family, which include the proteolytic enzymes requiring acidic conditions for their optimal activity (Davies 1990). They share more than 50% amino acid sequence identity with other aspartic proteinases such as pepsin, chymosin, cathepsin D, and cathepsin E (Guruprasad et al. 1996;Xie et al. 1991). Broadly, PAGs are categorized into two groups: an ancient group secreted by trophoblast mono-and bi-nucleate cells and a modern group secreted by trophoblast bi-nucleate cells only (Xie et. 1997). The ancient PAGs are known to possess their proteolytic activity like other typical aspartic proteinases (Telugu et al. 2010;Wooding et al. 2005). However, the modern PAGs are considered to be enzymatically inactive due to key mutations Masoud Lotfan and Suman Choudhary contributed equally to this work.

Electronic supplementary material
The online version of this article (https ://doi.org/10.1007/s4248 5-018-00003 -5) contains supplementary material, which is available to authorized users. within their catalytic center (Guruprasad et al. 1996;Green et al. 2000;Xie et al. 1991). Comparative modeling studies have shown that PAGs have retained the well-known bilobed structure similar to other APs, containing a substrate-binding cleft between the N-and C-terminal lobes (Andonissamy et al. 2012;Guruprasad et al. 1996). Reports have claimed that they can bind to 6-amino acid-long peptide, pepstatin A, which is a powerful inhibitor of APs (Guruprasad et al. 1996;Patel et al. 2004). Previously, it has been reported that bovine PAG2 and PAG12 exhibit proteolytic activity at acidic pH and both were inhibited by pepstatin A (Telugu et al. 2010). Therefore, in spite of their speculated function during pregnancy, their enzymatic physiological function as proteinases needs to be explored further.
The post-translational processing is a complex phenomenon which contributes to provide specific functions, immunogenicity and stability to various proteins. The trophoblastic binucleate cells (BNCs) store several glycoproteins in cytoplasmic granules and possess specific glycosylation patterns. It has been reported that PAGs and PRP-I are the main glycoproteins of bovine BNC granules to which various lectins can bind with high specificity such as VVA (Vicia villosa agglutinin), DBA (Dolichos biflorus agglutinin) and PHA-L (Phaseolus vulgaris leucoagglutinin) (Klisch et al. 2005). Glycosylation does play a role in antigenicity of PAGs and may also have functional roles during pregnancy (Patel et al. 2004). However, detailed information is lacking on the glycan compositions of these extensively glycosylated proteins to demonstrate their functional significance.
The PAGs constitute a multigene family of proteins and more than 100 genes of this molecule are speculated with differences in their spatial and temporal expression (Brandt et al. 2007;El Amiri et al. 2004;Green et al. 2000;Garbayo et al. 2008;Szafranska et al. 2006;Telugu and Green 2008;Xie et al. 1997). Studies have shown that some PAGs were expressed during certain stages of pregnancy while others were absent Zoli et al. 1992a). Expression pattern of bovine and ovine PAGs showed that some are expressed throughout the trophoectoderm, while others are restricted to binucleate cells . Different PAG isoforms and their variants have been detected through cDNA screening and extraction from placental tissues of various animal species (Klisch et al. 2005;Kiewisz et al. 2009;Sousa et al. 2002;Szafranska and Panasiewicz 2002;Xie et al. 1997). In bovine placental tissue, 22 PAG cDNA transcripts were screened using the reverse transcriptase-PCR as early as day 18 after artificial insemination (AI) Garbayo et al. 2008;Kiewisz et al. 2008). In spite of extensive evidence of transcripts, only a limited number of distinct PAG isoforms have been identified and characterized at the protein level. The concentration of PAGs throughout pregnancy has been studied thoroughly in different species (Chentouf et al. 2008;Green et al. 2005;Ledezma-Torres et al. 2006). In bovine, higher PAG concentrations were detected in maternal as compared to fetal serum indicating that they are delivered into the maternal circulation Zoli et al. 1992b). The average PAG concentration in plasma of cattle increased progressively from week 8 to week 35 followed by a strong increase in the last week of gestation. After delivery, plasma PAG concentrations declined significantly until weeks 2-10 postpartum (Sousa et al. 2003). In another study (Barbato et al. 2017), the buffalo (Bubalus bubalis) PAG molecules were detected at week 6 in blood samples using three different radioimmunoassay (RIA) systems and their concentrations were observed to increase gradually until week 28 of pregnancy. The exact function of PAGs is largely speculated in placentogenesis, embryo protection from maternal rejection, maternal recognition and sustenance of pregnancy (Szafranska et al. 2001;Zoli et al. 1992b). Because of their uniqueness to placental origin and their presence in maternal circulation, PAGs have been studied extensively for their application in pregnancy diagnosis in ruminant species (Green et al. 1998;Szafranska et al. 2006;Willard et al. 1995). Several researchers have reported the extraction and purification of PAGs from ruminant placenta (Butler et al. 1982;Barbato et al. 2008;Green et al. 2005;Telugu and Green 2008;Zoli et al. 1992b).
Buffalo is an important dairy animal which is known to have low reproductive efficiency due to long calving intervals, late puberty, low conception rate and high incidence of anestrus. BuPAG is a reliable biomarker for early detection of pregnancy . In water buffalo (Bubalus bubalis), pregnancy-associated glycoproteins were purified from late-pregnancy placentas for the development of a radioimmunoassay for pregnancy diagnosis . However, purification of PAGs in sufficient quantities from early stage of pregnancy is necessary for their characterization and to further make them good candidates for early pregnancy diagnosis in buffalo. Therefore, in the present study, we report (1) isolation and purification of PAGs from fetal cotyledons of less than 3-month pregnancy in water buffalo (Bubalus bubalis); (2) characterization of purified PAGs in terms of their proteolytic activity, identification of glycan moieties attached to these proteins and analysis of ligand-binding properties.

Purification of buffalo pregnancy-associated glycoproteins (BuPAG)
BuPAGs were purified by following the previously described protocol (Barbato et al. 2008) with minor modifications. Gravid uteri (less than 3-month pregnant) from buffaloes (Bubalus bubalis) were collected from local slaughter house and kept in saline. The age of pregnancy was determined by measuring their crown-rump (C-R) length (Masoud et al. 2018). Fetal cotyledons were separated out and stored at − 80 °C till further use. About 700 g of foetal cotyledons was thawed, finely minced, and homogenized in 10 mM potassium phosphate buffer (pH 7.6) in a ratio of 5:1 (v/w) in the presence of protease inhibitors such as phenylmethylsulfonyl fluoride (PMSF), sodium azide (NaN3), ethylene diamine tetra acetic acid (EDTA). The homogenate was stirred for 2 h, centrifuged (16,000×g, 1 h, 4 °C) and the pellets obtained were subjected to second extraction protocol. The pellets were mixed, homogenized in phosphate buffer with protease inhibitors, gently stirred overnight at 4 °C and finally centrifuged (16,000×g, 1 h). The pellets were discarded and the retained supernatant was pooled with the first extract.
The pooled protein extract was subjected to acid precipitation by adjusting the pH to 4.5 with 0.5 M H 3 PO 4 , allowed to precipitate at 4°C for 2 h and centrifuged (16,000×g, 1 h). The pH of the supernatant was readjusted to 7.6 with 0.5 M KOH and subjected to ammonium sulphate (AS) precipitation for 16 h to obtain 0-40% AS fraction and then for 4 h to obtain 40-80% AS fraction followed by centrifugation (16,000×g, 1 h). The pellets were re-suspended in 0.01 M Tris-HCl buffer (pH 7.6), extensively dialyzed against the same buffer and finally centrifuged (16,000×g, 1 h).The dialyzed protein fractions were then subjected to anionexchange chromatography using DEAE-Sephadex A50 column (SIGMA, USA). The protein fractions were loaded onto the column pre-equilibrated with 0.01 M Tris-HCl buffer (pH 7.6) and allowed to bind. The column was washed with the same buffer till the OD decreased to 0.05 at 280 nm. The bound protein was eluted with elution buffer (Tris-HCl, pH 7.6) using concentration gradient till 1 M NaCl. The eluted fractions were subjected to 12% SDS-PAGE to check the level of purity of BuPAGs. The protein bands were visualized by CBB (Coomassie Brilliant Blue) staining. The fractions were then dialyzed overnight against the buffer 5 mM ammonium bicarbonate (pH 8.0) at 4°C. After dialysis, the sample was concentrated by stirred ultrafiltration cell (Amicon, USA) using 10 kDa cut-off membrane and lyophilized.
The lyophilized protein was dissolved in 10 mM HEPES buffer (pH 7.5) to obtain a final protein concentration of 2 mg/mL and centrifuged (27,000×g, 1 h, 4°C). The protein was loaded onto a 10 mL column (BioRad, USA) packed with agarose-bound Vicia villosa (VVA) lectin (Vector Laboratories, Burlingame, USA) and allowed to bind. The column was extensively washed with same buffer till the OD decreased to 0.05 at 280 nm. The bound protein was eluted with elution buffer containing 50 mM N acetyl galactosamine (Acros Organics, New Jersey, USA). The eluted fractions then dialyzed against 5 mM ammonium bicarbonate (pH 8.0), centrifuged at 27,000×g for 15 min at 4°C and concentrated by stirred ultrafiltration cell. The purified protein was quantified using Bradford assay (Bradford 1976). The protein fractions were subjected to 12% SDS-PAGE followed by CBB staining to check the level of purity of BuPAGs and were finally confirmed by Western blot. For Western blot, unstained protein from the SDS-PAGE gel was transferred to immobilon PVDF membrane (Millipore, USA). The membrane containing transferred proteins was blocked using NAP blocker (G-biosciences, USA, 1:1 with TBST). The membrane was incubated overnight at 4 °C with primary antibody (1: 1000 dilutions). The primary antibody used was a peptide-based antibody raised in rabbit against buffalo PAG antigenic peptide namely AS#860 and AS#859 (Prof. J.F. Beckers, University of Liege, Belgium). The membrane was washed thrice for 10 min each with TBS (Tris-buffer saline) and incubated with secondary antibody solution for 1 h at 1:2000 dilutions. The secondary antibody was horseradish peroxidase-conjugated anti-goat IgG (Merck, USA). Finally, immunoreactivity was detected by the DAB system (Bangalore Genei, India).

Glycan composition analysis of BuPAGs by mass spectrometry (LC-MS/MS)
The purified protein at 10 μg concentration was used for insolution trypsin digestion and subjected to nano-LC (Nano Advance, Bruker, Germany) followed by identification by captive spray-Maxis-HD qTOF (Bruker, Germany) mass spectrometer (MS). No electron transfer dissociation (ETD) was used for generation of the data. The MS data were analyzed using Mascot (2.4.1 Matrix Science, UK). The search criteria were as follows: taxonomy-other mammalia, trypsin digestion, allowing up to one missed cleavage, variable modification-oxidation of methionine, fixed modificationcysteine as carboxyamidomethylation or propionamide, peptide tolerance of 50 ppm, and MS/MS tolerance of 0.05 Da. The ion-score cut off was manually set to 15 to eliminate the lowest-quality 189 matches. To eliminate false positives, 1% FDR was applied at both protein and peptide levels. For identification of glycan moieties, mgf file was subjected to glycan search using Glycome database.

Carbohydrate-binding analysis by fluorescence quenching assay
The carbohydrate-binding analysis of BuPAG was done on the basis of ligand-induced quenching of intrinsic tryptophan fluorescence. Fluorescence measurements were carried out on Varian Cary Eclipse fluorescence spectrophotometer (Agilent, USA). Emission spectra were recorded from 290 to 500 nm upon excitation at 280 nm. Excitation and emission slits were maintained at 10 nm and the scan speed was set at 600 nm/min. Standard reaction mixtures were prepared using 20 µM solution of protein in 25 mM phosphate buffer saline, pH 7.2 to a final volume of 1 mL. Binding affinity was analyzed with different sugars (N-acetyl glucosamine, mannose, dextrose and galactose). The additives were preincubated with BuPAG for about 1 h to obtain the spectra. The spectrum was corrected for background emission generated by buffer or ligands and the experiments were repeated in triplicates.

Determination of optimum pH
For proteolytic activity assay of BuPAG, the optimum pH was estimated by incubating the purified protein in various buffers. All the pH activity experiments were conducted at 37 °C. A synthetic FRET cathepsin D/E substrate (Sigma) was used to investigate the activity of native BuPAGs. The substrate was dissolved in 10% DMSO (Di-methyl sulphoxide) to a final concentration of 200 μM and stored in dark at − 80 °C till further use. The protein sample at 0.1 μg/μl concentration was incubated with 20 μM of substrate. The pH optima was determined by measuring activity at 37 °C for 30 min over the pH range 2.0-5.5 using 50 mM glycine-HCl (pH 2.0-3.5), 50 mM sodium citrate (pH 4.0-4.5) and 0.1 M sodium acetate (pH 5-5.5) as assay buffers. The reaction was performed for 20 min at 37 °C in a black-colored 96-well plate (Nunc). The resultant fluorescence in the mixture was read in the Infinite 200 PRO NanoQuant (Tecan, Switzerland) at 328 nm (excitation) and 393 nm (emission). All the experiments were set up in triplicates. The results from duplicate reads and from two successive experiments were used to compile the data.

Proteolytic activity inhibition assay
The pepstatin A inhibitor (Sigma Aldrich) was used to examine its effect on proteolytic activity of BuPAG. The inhibitor at 1 μM concentration was used to observe the inhibition of protease activity at optimum pH and 37 °C.

Isolation of native BuPAGs from fetal cotyledons
As an initial step for the purification of BuPAGs, the acid precipitation procedure resulted in the clarification of tissue extract and removal of most of the abundant non-specific proteins. Subsequently, DEAE anion-exchange chromatography step was performed where the SDS-PAGE profile of the eluted fractions showed multiple protein bands in the range of size ~ 45-70 kDa including some non-specific bands of low molecular weights (Fig. 1a, b). To remove the nonspecific protein bands, further purification of BuPAG by VVA chromatography resulted in three major protein bands in the range of size ~ 55-70 kDa (Fig. 2a). In Western blot confirmation, three protein bands, representing different glycosylated forms of BuPAG, were detected in the range of size ~ 55-70 kDa by specific peptide-based antibody (Fig. 2b).

Glycan analysis of BuPAGs
We identified the N-and O-linked glycan structures of BuPAGs in the present study by high-resolution LC-MS/ MS (ESI-qTOF). The MS analysis revealed the presence of a total of 9 different N-glycan structures within BuPAG Fig. 1 Purification of native buffalo BuPAG by ion-exchange chromatography (IEX). a Chromatogram showing elution of two merged peaks (indicated by arrows) during a concentration gradient run; b SDS-PAGE analysis of IEX purified protein fractions from peak 1 and peak 2. Lane 1-2: peak 1 fractions showing non-specific protein bands, lane 3: molecular weight marker, lane 4-6: peak 2 fractions showing partially purified native BuPAGs (indicated by braces) (Fig. 3a). The MS spectra obtained for the purified BuPAG and the MS data for the identified N-and O-glycan structures are provided in (Supplementary files 1-3), respectively. The structures were identified as high-mannose type, hybrid and complex type. High-mannose-type N-glycans consisted mainly of mannose residues with GlcNAc (N-acetyl glucosamine) at the distal or reducing ends. All N-glycan structures possessed terminal GlcNAc residues at reducing ends of their chains. The complex structures consisted of mannose residues along with glucose, galactose, GlcNAc, GalNAc (N-acetyl galactosamine), and sialic acid (Neu5Ac, N-acetyl neuraminic acid). Some of the complex structures were inferred to possess sialic acid at the non-reducing ends of their antennae. One di-sialylated biantennary N-glycan structure with composition Hex5HexNAc4NeuAc2 was observed where two sialic acid residues were linked to galactose residues at outer antennae.
Other two bi-antennary N-glycans with compositions Hex5HexNAc3NeuAc1 and Hex5HexNAc4NeuAc1 were sialylated at only one of their branches. The hybrid and complex glycans were identified as mainly biantennary due to the presence of two or more branches and one triantennary glycan was identified with composition Hex5HexNAc4S1 possessing three branches. The MS analysis revealed the presence of 3 O-glycan structures in BuPAGs (Fig. 3b). The structures mainly consisted of galactose, GalNAc, Glc-NAc, and sialic acid residues. All glycans were observed to possess a GalNAc residue at their reducing ends. Two O-glycans with compositions Hex2HexNAc2NeuAc1 and Hex1HexNAc1NeuAc1 contained sialic acid residues linked to galactose at their reducing ends, while one structure with composition Hex1HexNAc1NeuAc1 had sialic acid linked to GalNAc at one of the outer antennae. The compositional analysis revealed that no mannose or glucose residue was identified in any of the O-glycan structure. The LC-MS/MS signals and composition analysis of N-and O-linked glycans of BuPAGs are shown in Table 1.

Carbohydrate binding analysis of BuPAGs
During fluorescence quenching assay, the fluorescence emission spectrum of BuPAGs revealed that addition of different sugars (N-acetyl glucosamine, mannose, dextrose and galactose) at four different concentrations, i.e., 10 mM, 15 mM, 30 mM and 50 mM to the protein solution resulted in quenching of the fluorescence. When BuPAGs were incubated with monosaccharides, quenching of the fluorescence was observed between 310 and 330 nm without any shift of wavelength emission maxima (Fig. 4). Quenching in the fluorescence was observed more at higher ligand concentrations, i.e., 30 mM and 50 mM. Beyond this concentration, there was no detectable change in the spectra.

Effect of pH on proteolytic activity of BuPAG
The pH profile of native BuPAG revealed that the enzyme was active in the pH range of 2-5.5 with maximal activity at pH 3.0. Protease activity decreased significantly below pH 3.0 and was approximately 42% of the maximal activity when assayed at pH 2.5 and 29% at pH 2.0 in glycine-HCl buffer. Enzyme activity was slightly reduced to 74% and 65% at pH 3.5 and pH 4.0, respectively, in glycine-HCl buffer. It was observed to be 57% at pH 4.5 in sodium citrate buffer, 44% at pH 5.0 and pH 5.5 in sodium acetate buffer (Fig. 5a).

Effect of inhibitor on proteolytic activity
The activity of purified BuPAG was strongly inhibited by Pepstatin A, an inhibitor that specifically and irreversibly binds to aspartate within the protease active site. The proteolytic activity of BuPAG was reduced by 84% at an inhibitor concentration of 1 μM. The activity of denatured BuPAG was observed to be reduced by 79% (Fig. 5b).  Barbato et al. reported the apparent molecular masses of the immune reactive bands by VVA purification from 59.5 to 75.8 kDa using the cotyledons from mid-pregnancy placenta (Barbato et al. 2008). Various other reports have established the existence of single PAG corresponding to molecular mass of 60.67 kDa (Sousa et al. 2002;Xie et al. 1996;Zoli et al. 1991). In bovine, the mass spectrometric analysis identified ~ 56-75 kDa bands representing different PAG isoforms (Klisch et al. 2005). As shown by our study, native BuPAGs were detected in the range of size ~ 55 kDa-70 kDa; however, the theoretical molecular mass of BuPAGs based upon the amino acid sequence is approximately 40 kDa. This clearly shows that the observed size differences are due to the presence of various N-and O-glycans with different extents of glycosylation and BuPAGs are secreted as highly glycosylated proteins in their native state. Furthermore, the enzymatic release of N-glycans from PAGs in bovine reduced the molecular weight to ~ 37 kDa which corresponds to the theoretical molecular mass and the attached N-glycans were suggested as the main determinant of molecular mass of bovine PAGs (Klisch et al. 2005). The number of potential N-linked glycosylation sites has been detected from 2 to 5 in caprine, 2-7 in sheep and 1-6 in bovine (Garbayo et al. 2000;Xie et al. 1996Xie et al. , 1997. The glycosylation pattern of a protein can depict its biological activity and contributes to the maintenance of its stability and quality. The variable degree of glycosylation in different PAGs has been reported to be an important factor in regulating the plasma half-life of these proteins and also play a role in antigenicity of PAGs (Klisch et al. 2005). We observed in our study that a higher number of different N-glycans,

Discussion
i.e., 11 are present in BuPAGs as compared to 3 different O-glycan moieties. However, the amino acid sequence analysis revealed the presence of 3-5 potential N-linked glycosylation sites in BuPAGs and 6-12 O-linked glycosylation sites. This suggests that in spite of the presence  relative activity of BuPAG with and without inhibitors. The maximum observed activity under this condition defined as 100% and the relative activities were calculated as a fraction of this value of more O-linked glycan sites, a higher variability in the compositions of N-glycans is present in BuPAGs. The Nand O-glycan chains are built up sequentially via an initial linkage of GlcNAc to the amide group of asparagine and GalNAc to the hydroxyl groups of serine/threonine residues, respectively (Varki et al. 2009). The distal or reducing ends of the chains of N-and O-glycans were found to be terminated by GlcNAc and GalNAc, respectively, in the identified structures. The sialylation of N-and O-linked glycan chains containing galactose and GalNAc residues at their non-reducing ends was identified in BuPAGs. Reports have suggested that sialic acids present in oligosaccharides of glycoproteins play essential roles in a variety of biological processes including the recognition of carbohydrate groups, mediate a variety of cell-cell interactions and also function in the masking of the terminal carbohydrate groups (Schauer 1982). The identified N-and O-linked glycan moieties may be actively involved in the process of embryo implantation. The glycan structures expressed on cell surfaces or secreted in biological fluids are involved in various functions related to cell recognition and adhesion processes (Listinsky et al. 1998;Staudacher et al. 1999). They may have significant biological functions in protecting BuPAGs against the proteolytic degradation, aggregation and in modulating various biological functions of these glycoproteins. We performed the fluorescence-quenching experiments to assess the sugar-binding property of BuPAGs. It has been reported that pregnancy-associated rat uterine glycoprotein possesses carbohydrate-binding properties and have a significant role in embryo implantation (Das et al. 1994). This led us to hypothesize that apart from being highly glycosylated, PAGs may also possess ability to bind carbohydrates to perform significant biological functions. As identified in our study, the ability of BuPAGs to bind different sugars reveals a novel structural feature of BuPAGs to possess specific sugar-binding sites. The sugar-binding ability of BuPAGs may have a role in interactions with other glycosylated protein partners to execute their specific biological functions. The receptors containing sugars such as N-acetyl glucosamine, mannose, dextrose and galactose may have a large number of biologically significant roles such as binding to various growth factors required for development of fetus during pregnancy. Our results indicate that apart from containing various sugar moieties, PAGs possess sugar-binding properties. Therefore, PAGs are not just the possible proteases but they may also act as lectins.
This is the first report which describes the proteolytic activity of buffalo PAGs isolated in native state. Previous reports have described the proteolytic properties of porcine and bovine recombinant PAGs expressed in Baculoviral expression systems (Telugu and Green 2008;2010). The purified native PAG preparations were tested for their optimal activity against a synthetic FRET cathepsin D/E substrate. This fluorescence substrate has been tested against Cathepsin D and E which are well-characterized APs (Yasuda et al. 1999). The native BuPAGs in our study exhibited an optimal proteolytic activity towards this substrate. The maximal activity for BuPAGs was observed at acidic pH 3.0. Bovine PAGs also exhibited optimal activity at acidic pH optima which is typical for most APs (Telugu et al. 2010). Porcine PAG2 showed optimal proteolytic activity around pH 3.5 (Telugu and Green 2008). The acidic environment aids in the cleavage of the propeptide in most APs providing access of the substrate to the active site (Wittlin et al. 1999). PAGs have been suggested to play a role in cleavage of the protein substrates either at the fetal-maternal interface or within the trophoblast and in proteolytic processing of proteins to play a role in secretory pathways (Telugu et al. 2010). The difference in optimal activities at different pH levels in various species may be due to the variation in substrate-binding residues within the active site. In our recent study, it was shown that BuPAG possesses a well-defined substrate-binding cleft as observed in other APs and various residues within the cleft are involved in polar and non-polar interactions with pepstatin inhibitor for its binding with a high affinity (Lotfan et al. 2018). This is further supported by our present investigation where the proteolytic activity of BuPAG was found to be strongly reduced in the presence of pepstatin inhibitor. This further paves a way for future investigations for identification of putative substrates to gain insight into the physiological role of PAGs in buffalo.

Conclusions
The mass spectrometry-based analysis of purified BuPAGs reveals the presence of a variety of glycan structures which may be responsible for their specific structural and functional attributes. The study reports preliminary analysis of glycan compositions of BuPAG in ESI-qTof without using ETD. Therefore, data generated using ETD will help in indepth analysis of glycan residues in future. Determination of the sugar compositions of these proteins will further help in revealing their surface features as well as their interaction properties with other proteins and ligands. The potential of BuPAGs to bind different sugars and ability to act at low pH range further provides new functional insights into the role of these proteins.