Human sialic acid acetyl esterase: Towards a better understanding ...

Glycan Metabolism

Human sialic acid acetyl esterase: Towards

a better understanding of a puzzling enzyme

Flavia Orizio2, Eufemia Damiati3, Edoardo Giacopuzzi3, Giuliana Benaglia2,

Stefano Pianta2,6, Roland Schauer4, Reinhard Schwartz-Albiez5,

Giuseppe Borsani3, Roberto Bresciani2, and Eugenio Monti1,2

2Unit of Biotechnology, 3Unit of Biology and Genetics, Department of Molecular and Translational Medicine (DMTM),Viale Europa 11, Brescia 25123, Italy, 4Institute of Biochemistry, Christian-Albrechts University, Otto-Hahn-Platz 9, Kiel24118, Germany, and 5Department of Translational Immunology, Unit of Glycoimmunology, German Cancer ResearchCenter, Im Neuenheimer Feld 280, Heidelberg 69120, Germany

1To whom correspondence should be addressed: Tel: +39 303717544; Fax: +39 303701157; e-mail: [email protected]

6Present address: Centro di Ricerca E. Menni (CREM), Fondazione Poliambulanza-Istituto Ospedaliero, Via Bissolati, 57,25124 Brescia, Italy.

Received 10 October 2014; Revised 5 May 2015; Accepted 17 May 2015

Abstract

Sialic acid acetyl esterase (SIAE) removes acetyl moieties from the hydroxyl groups in position 9 and

4 of sialic acid. Recently, a dispute has been opened on its association to autoimmunity. In order to

get new insights on human SIAE biology and to clarify its seemingly contradictory molecular prop-

erties, we combined in silico characterization, phylogenetic analysis and homology modeling with

cellular studies in COS7 cells. Genomic and phylogenetic analysis revealed that in most tissues only

the “long” isoform, originally referred to lysosomal sialic acid esterase, is detected. Using the hom-

ology modeling approach, we predicted a model of SIAE 3D structure, which fulfills the topological

features of SGNH-hydrolase family. In addition, the model and site-directed mutagenesis experi-

ments allowed the definition of the residues involved in catalysis. SIAE transient expression revealed

that the protein is glycosylated and is active in vitro as an esterasewith a pH optimum corresponding

to 8.4–8.5. Moreover, glycosylation influences the biological activity of the enzyme and is essential

for release of SIAE into the culturemedium. According to these findings, co-localization experiments

demonstrated the presence of SIAE inmembranous structures corresponding to endoplasmic reticu-

lum andGolgi complex. Thus, at least in COS7 cells, SIAE behaves as a typical secreted enzyme, sub-

jected to glycosylation and located along the classical secretory route or in the extracellular space. In

these environments, the enzyme could act on 9-O-acetylated sialic acid residues, contributing to the

fine-tuning of the various functions played by this acidic sugar.

Key words: 3D model, in silico characterization, Neu5Ac, O-acetylation, phylogenesis, sialic acid acetyl esterase-SIAE, subcellularlocalization

Introduction

Sialic acids as terminal sugars of N- and O-glycosylated macromole-cules are decisive for the functions of glycoconjugates both in solubleform and integrated into cell membranes (Varki 2008; Schauer 2009).

A wide range of structural diversity characterizes sialic acids, withO-acetylation at the hydroxyl group of carbonC9 as themost frequentmodification. Functions and biosynthesis of O-acetylated sialic acids

Glycobiology, 2015, vol. 25, no. 9, 992–1006doi: 10.1093/glycob/cwv034

Advance Access Publication Date: 28 May 2015Original Article

© The Author 2015. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: [email protected] 992

Downloaded from https://academic.oup.com/glycob/article-abstract/25/9/992/1988759by gueston 11 February 2018

http://www.oxfordjournals.org

have been recently reviewed (Mandal et al. 2012) and among the greatvariety of phenomena involving these modified sialic acids, severalhave been described for the immune system (Schauer et al. 2011), inparticular for human B-cell differentiation (Erdmann et al. 2006;Wip-fler et al. 2011) and autoimmunity (Pillai et al. 2009). In this perspec-tive, the balance betweenO-acetylation and de-O-acetylation of sialicacid as an on/off switch mechanism is able to regulate distinct pro-cesses such as apoptosis (Malisan et al. 2002; Kniep et al. 2006; Mu-kherjee et al. 2008). The two enzymes responsible for acetylation andde-acetylation have been described as sialate O-acetyltransferase(SOAT) and sialate O-acetylesterase (SIAE), respectively (Mandalet al. 2012). Recently, the Cas1 protein has been reported as a possiblecandidate for sialic acid specific O-acetytransferase of sialic acids(Arming et al. 2011). The first description of an esterase acting on4-O- and 9-O-acetyl groups of free or bound sialic acid dates backto 1983 (Shukla and Schauer 1983) and from 1986 onwards (Varkiet al. 1986) several studies have been published on SIAE (Mandalet al. 2012). SIAE (EC 3.1.1.53) is largely diffuse in nature and severalforms of the enzyme with regards to its origin are reported (EnzymeDatabase-BRENDA, www.brenda-enzymes.info/) ranging fromviruses and bacteria to vertebrates. A functional link between SIAE ac-tivity and B-cell development, signaling and immunological tolerancehas recently been described in rodents (Cariappa et al. 2009). Indeed,engineered mice with catalytically deficient SIAE showed enhancedB-cell receptor (BCR) activation, defects in peripheral B-cell develop-ment, presence of autoantibodies against chromatin and developedglomerular deposits of immune-complexes. At the molecular level,this effect could be mediated by the 9-O-acetylation state of sialicacid, which in turn regulates the function of CD22/Siglec-2. Actually,SIAE deacetylates α(2–6)-linked sialic acid on N-linked glycans ofBCR, allowing the interaction with CD22, that functions in vivo asan inhibitor of BCR signaling (Dorner et al. 2012; Pillai et al.2012). Starting from 2010, functionally defective SIAE variants havebeen associated to susceptibility to autoimmune disorders in humansubjects (Surolia et al. 2010; Hirschfield et al. 2012). However, a sub-sequent large study on ∼67,000 subjects failed to replicate the associ-ation with autoimmune and chronic immune diseases (Hunt et al.2012), pointing out the challenges of associating low-frequency var-iants with complex traits and disease. Recently, catalytically defectiverare variants of SIAE have been again associated with autoimmunity,reopening the debate about the role played by this enzyme in immuno-logical response (Chellappa et al. 2013). From the very beginning,SIAE has been described as an enzyme presenting two (iso-)forms,the cytosolic sialic acid esterase (Cse) and the lysosomal sialic acid es-terase (Lse)/membrane-associated enzyme (Varki et al. 1986; Higaet al. 1987; Diaz et al. 1989; Butor et al. 1993, 1995). Subsequentstudies on the enzyme resulted in contrasting findings (Guimaraeset al. 1996; Stoddart et al. 1996; Cariappa et al. 2009) and several as-pects of SIAE biology remain still unsolved. For example, differentmolecular-weight subunits of the cytosolic enzyme, originated by pro-teolytic cleavage of the lysosomal protein during maturation in theendosome/lysosome compartment, have been reported for SIAE(Takematsu et al. 1999) as well as differential splicing events havebeen described for SIAE (Stoddart et al. 1996). Also the functional im-pact of the subcellular/extracellular localization of the protein(s) re-mains unclear (Mandal et al. 2012). In 2004, the cDNA encodingfor the human SIAE has been isolated by using differential displayRT-PCR from testis, revealing a high degree of sequence identity tothe mouse enzyme and an apparent lysosomal localization of the pro-tein in a human pancreatic adenocarcinoma cell line (Zhu et al. 2004).Although several other DNA sequences encoding for SIAE orthologs

have been identified from various organisms, a detailed phylogeneticanalysis of this enzyme(s) has never been carried out. Moreover, basedon the presence of conserved residues and on functional data, includ-ing the use of inhibitors (Schauer et al. 1989), SIAE enzyme(s) are clas-sified as serine esterases, belonging to a super family of hydrolaseswith >7000members (see for details UniProt KW-0719). Surprisingly,no structural data are available for SIAE enzymes from higher organ-isms, whereas detailed studies have been carried out on microbial es-terases acting on sugar residues (Pfeffer et al. 2013), on a SIAE fromEscherichia coli (Rangarajan et al. 2011) and further extensive studieson viral SIAE like from influenza C, corona and toroviruses (Herrlerand Klenk 1991; de Groot 2006). Trying to get new insights on humanSIAE biology and in order to clarify its debated molecular properties,we conducted a multidisciplinary study on this enzyme. We combinedin silico characterization, phylogenetic analysis and homology model-ing with cellular studies in COS7 cells revealing new details on thebiology of SIAE enzyme.

Results

Characterization of the human SIAE gene

The human SIAE gene is located on chromosome 11 at position11q24.2 (Zhu et al. 2004). As summarized in Introduction, previousstudies report that it encodes two different transcript isoforms namedvariant 1 and 2 also listed in the Gene database at National Center forBiotechnology Information (NCBI). Supplementary data, Figure S1shows the two variants as displayed in the UCSC Genome Browser.Variant 1 (RefSeq NM_170601) spans 10 exons and represents theshorter transcript but encodes the longer isoform 1 of the protein,also referred to as Lse. Variant 2 (RefSeq NM_001199922.1) spans12 exons that differs in the 5′ UTR, and its first coding exon corre-sponds to the second of variant 1. The resulting isoform 2 of the pro-tein has a different internal N-terminus and is also referred to as theCse. For these reasons, Lse is 523 aa long (RefSeq NP_733746),while Cse has 488 aa (RefSeq NP_001186851) lacking the first 35aa of the lysosomal form. A bioinformatics analysis revealed the pres-ence of 78 expressed sequence tags (ESTs) generated from varioushuman tissues and cell types supporting the SIAE transcript variant1, while no ESTs corresponding to SIAE transcript variant 2 werefound (Supplementary data, Figure S2A).

The Ensembl genome browser reports only two SIAE transcriptswith protein-coding capabilities: SIAE-001, corresponding to the Lseisoform (RefSeq NM_170601, NP_733746), and SIAE-201 which isidentical to the previous one with the exception that the second inframe ATG codon is considered as translation starting site. Althoughthis would result in a 488 aa polypeptide identical to the Cse encodedby the above-mentioned variant 2 (RefSeq NM_001199922.1,NP_001186851), a bioinformatic analysis carried out with algorithmsfor identifying the initiation codons in cDNA sequences strongly sup-port the first in frame ATG as the most reliable translation start site.No protein-coding transcript with an alternative 5′ exon correspond-ing to the RefSeq NM_001199922.1 cytosolic SIAE isoform is listedby Ensembl. Furthermore, the analysis of RNASeq data from Illumi-na’s Human BodyMap 2.0 project indicates that in 15 of the 16human tissues analyzed exon–exon boundaries correspond to variant1, while there is no transcript model supporting variant 2 (Supplemen-tary data, Figure S2B).

Previous studies indicate that the Siae mouse gene behaves quitesimilarly, encoding for two isoforms that differ in length likely as aconsequence of an alternate promoter (Takematsu et al. 1999). An

A study of SIAE using a multidisciplinary approach 993


www.brenda-enzymes.info/




http://glycob.oxfordjournals.org/lookup/suppl/doi:10.1093/glycob/cwv034/-/DC1




in silico analysis showed that also in mice, only the Lse transcript vari-ant is present in the ESTs database with 55 sequences (data notshown). Furthermore, NCBI Gene database reports as gene product

a single reference sequence (RefSeq) of 541 aa (NP_035864), corre-sponding to Lse. It has to be noted that five mouse ESTs (four fromwhole joint and one from corpora quadrigemina) are present in

Fig. 1. LOGO representation of multiple alignment of SIAE orthologous proteins. Themultiple sequence alignment of 32 sequences from SIAE orthologous proteins

is represented in LOGO format. Peptide sequence and amino acid numbering refers to human SIAE (NP_733746.1). Black lines indicate the 11 highly conserved

sequence blocks, while the red line indicates the four conserved blocks typical of family II SGNH hydrolases. Secondary structure elements are reported as

black arrows, indicating β-sheets, or black spirals, indicating α-helices. They are named accordingly to the topological representation in Figure 3A. The other

predicted features are also reported: †, glycosylation sites; ‡, phosphorylation sites; *, catalytic residues; +, residues potentially interacting with substrate. This

figure is available in black and white in print and in color at Glycobiology online.

994 F Orizio et al.


dbEST having an alternative 5′ exon compared with the Siae Lse tran-script. These ESTs correspond to a transcript variant named X1 en-coding a 553 aa polypeptide (XP_006510217), with the first 35differing from the Lse isoform. Overall, these data indicate that inadult human and mouse tissues expression of SIAE transcripts is al-most exclusively limited to the lysosomal SIAE isoform.

RNA-Seq data generated from 27 human tissues and made avail-able through the Expression Atlas database have been used to infer theexpression level of the SIAE gene. Supplementary data, Figure S3 indi-cate that the gene is expressed in all tissues tested, although at differentlevels. The highest levels of expression have been observed in colon,kidney and thyroid gland.

In silico characterization of SIAE protein

Genome sequencing projects recently completed on several organismsallow the retrieval of valuable information about molecular phylogen-esis of the gene of interest (Mulder et al. 2008; Genome 10K Commu-nity of Scientists 2009). This allowed us to carry out the firstphylogenetic analysis of SIAE, pointing out several interesting featuresof this enzyme.

To identify highly conserved motifs and residues, we used MUS-CLE to generate a multiple sequence alignment with sequences from

32 protein orthologs of human SIAE (Supplementary data, FiguresS4, S5 and Table SI). According to the criteria defined in Materialsand methods, we were able to identify 11 highly conserved sequenceblocks (Figure 1). Among them, we also identified four sequenceblocks characterizing family II of SGNH hydrolase superfamily (ester-ase blocks) and the two essential residues required for catalysis, repre-sented by the fully conserved amino acids S127 and H377. Thepredicted N-terminal signal peptide is highly conserved in vertebrates,with the exception of Taeniopygia guttata and Xenopus laevis (Sup-plementary data, Figure S5A).

To complete the in silico characterization of the human SIAE,we also investigated the presence of N- and O-glycosylation sites,as well as phosphorylation sites in the protein sequence. Usingthe tools described in Materials and methods, we predicted 6 N-glycosylation, 10 O-glycosylation (4 with high confidence) and24 phosphorylation sites (Table I, more details in Supplementarydata, Tables SII–SIV). We investigated the location of these sites inthe predicted protein structure. The sites T58, T73 and N107 are intheN-terminal portion, not included in the final structure. The other5 N-glycosylation sites and 7 O-glycosylation sites are, as expected,exposed on the protein surface, while T383 is located inside the pre-dicted structure (Supplementary data, Figure S7). Analysis of theconservation at these positions on the primary structure identifiedtwo glycosylation sites (N138 and S177) highly conserved throughMetazoa, and five phosphorylation sites (S30, S227, S228, T221 andS255) highly conserved in all the 32 protein sequences used in ourmultiple alignment (Figure 2). Using subfamily LOGO to asses forconservation in specific subgroups, we identified also two potentialglycosylation sites conserved only in Vertebrata and three conservedonly in Mammalia, as well as five phosphorylation sites conservedonly in Metazoa and five only in Mammals (Figure 2). In silico pre-dictions revealed the presence of an N-terminal signal peptide inhuman SIAE corresponding to amino acids M1-A18. This result isin agreement with data from the Signal Peptide Database (http://

Fig. 2. Conservation of glycosylation and phosphorylation predicted sites. The conservation of the predicted glycosylation and phosphorylation sites were assessed

from the multiple alignments of SIAE orthologous proteins. Subfamily LOGO analysis performed as described in Materials and methods suggested differential

conservation occurring at 5 N-glycosylation (A) and 2 O-glycosylation sites (B). The most conserved predicted phosphorylation sites thorough Eukarya and

Metazoa are also reported (C). This figure is available in black and white in print and in color at Glycobiology online

Table I. Predicted posttranslational modifications on human SIAE

protein

Modification Predicted sites

N-Glycosylation N107, N138, N267, N290, N401, N422O-Glycosylation S177, S231, S242, T511Phosphorylation S30, Y31, T52, S74, S79, S144, S156, S158, S217,

T221, S227, S228, S247, S255, T303, S308, T312,S328, S331, S332, Y349, S371, T461, S470











http://www.signalpeptide.de/



www.signalpeptide.de/, last accessed 29 May 2015), reporting aslightly longer signal peptide (M1–A23). Secretome 2.0 predictionshowed that human SIAE has a high score (NN-score 0.78) as asecreted protein.

Interestingly, an amino acid stretch of variable length (23–28 resi-dues) typical of Muroidea has been found between conserved block 5and 6 (Supplementary data, Figures S5 and S8A). Localization of theseresidues in the predicted model indicates that they belong to the loopconnecting the β-strand 4 and the α-helix C (see for details Figure 3Aand Supplementary data, Figure S8B). It is tempting to speculate suchan insertion could confer some peculiar biochemical feature(s) to SIAEof this large superfamily of Rodents (Jiang and Blouin 2007).

Considering the relevance given by some studies to SIAE geneticvariants and their association to immune disease (Pillai 2013), wesearched the Exome Aggregation Consortium (ExAC) data set of60,706 unrelated individuals and produced a catalog (Supplementarydata, Table SV) of the genetic variability of SIAE gene in humansfocused on single-nucleotide variants (SNVs) affecting residues in-volved in either post-translational modification (Table I) or catalyticactivity (S127, T132 and H377).

SIAE structure prediction

Homology modeling based on the structure of known esterases fromEscherichia coli (3PT5), Clostridium acetobutylicum (1ZMB) andArabidopsis thaliana (2APJ) allowed the generation of a complete3D model for the human SIAE protein (Supplementary data,Figure S6). The N-terminal (M1–V112) and C-terminal (L402–K523) regions of the human protein were difficult to align to the tem-plate structure, resulting in long, mainly disorganized loops and werethus removed from subsequent analysis of the predicted structure. Thetopology of the predicted model (Figure 3A) shows the presence of anα/β/α structure typical of SGNH-hydrolase family, composed by 5 par-allel and 1 anti-parallel sets of β-sheets and 12 α-helices (Molgaardet al. 2000). The parallel β-sheets are buried in the protein core, sur-rounded by the α-helical structures (Figure 3B and C), which contrib-ute to most of the protein surface. The two residues S127 and H377,together with the four esterase blocks of conserved residues (Molgaardet al. 2000) are placed on the predicted model to assess the position ofthe putative active site. Together with the conserved block 5 (G219–S228), these portions of the protein constitute a surface suitable to re-present the catalytic crevice (Figures 3 and 4A, B).

Fig. 3. Predicted structure of human SIAE. The structure of the human SIAE protein was predicted using homology modeling based on the known structures 2APJ,

3PT5 and 1ZMB. (A) Topological representation of the predicted structure comprise a core region, with seven β-sheets, depicted as orange arrows (1–7), and nine

α-helices (A–I) and depicted as sea green cylinders. The C-terminal portion that poorly aligns with the template structures comprises three α-helices (J–K) and is

represented in blue. The core region of the structure, as seen in top (B) and lateral (C) cartoon representations, resembles the classical organization of family II SGNH

hydrolases with seven β-sheets buried within the α-helices. The 9-O-acetyl-Neu5Acmolecule is represented as fuchsia sticks and placed as resulted from automated

molecular docking, indicating the position of the putative active site. This figure is available in black and white in print and in color at Glycobiology online

996 F Orizio et al.











Automated blind docking of 9-O-acetyl-Neu5Ac substrate on thestructural model of human SIAE placed the molecule in the predictedactive site as one of the best hits (estimated ΔG =− 6.59 kcal/mol), fur-ther supporting the predicted conformation. The distance between theC9-O bond of the substrate molecule and the two catalytic residuesS127 and H377 is 3.8 and 4.3 Å, respectively. Moreover, using thegenerated molecular complex, we were able to identify the residuespossibly involved in interactions with the substrates (indicated with +in Figure 1): T132, Q135, Q175, W176, S177, K178, T180, M192,W218, G219, G220 and S332. Among these, the residues T132 andS332 are placed near the essential H377 and represent the best candi-dates as possible third element involved in the esterase catalytic triad.Mutagenesis experiments (see below) demonstrated that the threeresidues T132, S332 and H377 compose the actual catalytic triad.The architecture of the predicted active site of human SIAE is shownin Figure 4.

Cloning and expression of SIAE in COS7 cells

The in silico analysis of various transcriptome databases indicates thatthe so-called lysosomal SIAE or Lse is by large the prevalent isoformexpressed in all tissues investigated, as previously described. We thuscloned the lysosomal SIAE coding region (1572 bp) into the expres-sion vector pMT21 allowing the expression of a myc-tagged fusionprotein. The resulting construct (pMT21-SIAE) was used to follow ex-pression of the protein in transiently transfected COS7 cells for up to48 h by means of enzymatic activity and western-blot analysis. In cellextracts derived from pMT21-SIAE-specific COS7 transfectants, anincrease in SIAE activity was already detected after 24 h transfectionand after 48 h the enzymatic activity was increased roughly 5- to6-fold compared with non-transfected cells, indicating an enzymati-cally active overexpressed protein (Table II). Of note, in the condi-tioned culture media of cells transfected for 24 h, only a negligibleincrease in enzymatic activity was detected, whereas after 48 h expres-sion, almost a 5-fold increase was observed with respect to the mediaof non-transfected cells (Table II). Finally, using cell extracts as en-zyme source, we found an in vitro alkaline pH optimum of 8.4–8.5,with a negligible enzyme activity at pH <6.0 (Figure 5A). Taken to-gether, these data indicate that after 48 h expression, SIAE is releasedinto the growth medium and that the balance between intracellular-

to-extracellular activities is in favor of the secreted pool in a ratio of1 : 3.5, respectively.

Western-blot analysis confirmed the presence of SIAE, both in cel-lular extracts and conditioned culture media (Figure 5B). Interestingly,intracellular SIAE was detected already after 24 h expression as a62 kDa immunoreactive band, a molecular weight higher than the ex-pected 58 kDa, together with a second and less intense band of49 kDa and after 48 h the intracellular amount of SIAE increased sig-nificantly with the same protein pattern. In order to assess whether in-crease in the molecular weight of SIAE could be due to glycosylation,to COS7 cells expressing the enzyme for 24 h tunicamycin was addedto the medium for further 24 h. Tunicamycin treatment resulted in thecomplete disappearance of the 62 kDa band in cell extracts, in favor ofthe lower molecular weight band of 49 kDa (Figure 5C), suggestingthat the increase in the molecular weight is the result of a glycosylationprocess. Concerning the release of SIAE from cells, after 24 h expres-sion a faint immunoreactive band of 62 kDa was detected in thegrowth medium and, as expected from the enzyme activity (Table II),at 48 h expression its intensity is significantly increased (Figure 5B).Compared with the 24 h expression, tunicamycin treatment resultedin the increase of the 62 kDa band in the conditioned medium, prob-ably corresponding to release of the high-molecular weight form ofSIAE produced by the cells in the first 24 h of expression. To bettercharacterize synthesis and processing of newly synthesized SIAE,

Fig. 4. Predicted active site of human SIAE. Analysis of the predicted structure and molecular docking, allowed the identification of a putative catalytic site. The

conformation of the catalytic surface (A) is suitable to accept the 9-O-acetyl-NeuAc molecule, used as substrate, and resulted in a swallow pocket able to accept

also complex glycans terminating with the acetylated NeuAc. The red portion of the surface is composed by the residues S127, T132 and H377 identified as the

catalytic triad. Their placement (B) is compatible with the knownmechanism of action on theO linkage at carbon 9. For the functional role played by T123 and H377,

see also Figure 6. This figure is available in black and white in print and in colour at Glycobiology online

Table II. SIAE activity in cell extracts and culture media: effect

of tunicamycin and cycloheximide treatment

Cell extracts (mU/mg) Media (mU/mL)

NT 1.65 ± 0.35 1.61 ± 0.0824 h 5.21 ± 0.69 1.55 ± 0.2148 h 9.31 ± 0.66 8.57 ± 0.97Tunicamycin 1.7 ± 0.54 3.3 ± 0.5Cycloheximide 3.6 ± 0.22 3.7 ± 0.04

COS7 cells were transfected and allowed to express SIAE for 24 and 48 h or,after 24 h expression, tunicamycin or cycloheximide was added to the mediaand cells grown for further 24 h. NT, non-transfected cells.

Values represent the mean of three independent experiments ± standarddeviation (for details, see Materials and methods).



COS7 cells were treated with cyclohexamide under the experimentalconditions previously described. After a 24 h treatment, the drugleads to the disappearance of the 49 kDa band in favor of the62 kDa form, with a corresponding increase of the same band in theculture media. Overall, these data suggest that, at least in our expres-sion system, the 58 kDa form of SIAE is rapidly proteolitically cleavedto a 49 kDa form or glycosylated to the 62 kDa form. Moreover, onlythe high molecular weight of the protein is secreted into the growth

medium, indicating a possible trafficking of the glycosylated proteinthrough the exocytic route.

We also investigated whether glycosylation influences the enzym-atic activity of SIAE. For this purpose, we measured the enzyme activ-ity both in cells extracts and medium under the experimentalconditions previously described. As shown in Table II and Figure 5B,in cell extracts SIAE-specific activity correlates with the intensity of the62 kDa band detectable in western-blot experiments, under normal

Fig. 5. Expression of SIAE in COS7 cells. (A) Rate of hydrolysis of 4-nitrophenyl-acetate in the range of pH from 3.0 to 9.0. Two different buffer systems were used:

citrate buffer from pH 3.0 to pH 6.0 (black triangle) and phosphate buffer from pH 6.0 to pH 9.0 (black circle). (B) Representative western blot analysis of crude cell

extracts (20 µg) and culture media (20 µL) from nontransfected cells (NT) and cells transiently expressing SIAE for 24 and 48 h in absence or in presence of

tunicamycin (Tun) or cycloheximide (Cyc) treatment (24 h + 24 h). (C) Left panel: western blot analysis of crude cell extracts (10 µg) from NT and cells transiently

expressing SIAE for 48 h with or without PNGase-F treatment. Right panel: SIAE activity measured in the same samples. For details, see Material and methods.

998 F Orizio et al.


expression conditions or in presence of tunicamycin or cycloheximide.The same evidence is obtained by considering the enzyme activityreleased in the culture medium. To further confirm these results, cellextracts deriving from COS7 cells expressing SIAE for 48 h were sub-jected to peptide-N-glycosidase-F (PNGase-F) treatment under non-denaturing conditions. As shown in Figure 5C, this treatment resultedin the disappearance of the 62 kDa band in favor of the 49 kDa bandand again, the pattern of SIAE-specific activity is roughly superimpos-able with western-blot results.

These data indicate that SIAE glycosylation is essential for its secre-tion into the growth medium and, as demonstrated by tunicamycinand PNGase-F treatment, influence the enzyme activity and/or stability.

Role of T132, S332 and H377 amino acid residues

in determining the enzymatic activity of SIAE

In order to determine the possible role played in the catalytic activity ofSIAE by the amino acid residues T132, S332 and H377 (for details,see SIAE structure prediction), we generated alanine-substitution mu-tants in these positions by site-directed mutagenesis and expressedthem in COS7 cells. After 48 h expression, cells and respectivemedia were collected and analyzed. The western-blot analysis evi-denced the presence of the 62 kDa band in both cell extracts and inthe culture media, demonstrating that mutants are indeed expressed(Figure 6). The alanine substitution at position S332 did not influencesignificantly the activity of the enzyme, whereas mutagenesis at pos-ition T132 and H377 abolished the enzymatic activity that resultedcomparable with non-transfected cells. These data clearly demonstratethat amino acids T132 and H377 play an essential role in the catalytic

activity of SIAE, defining the catalytic triad of the enzyme as com-posed, beside S127 (Surolia et al. 2010), by T132 and H377.

SIAE subcellular localization

Immunofluorescence experiments have been carried out in order tostudy the subcellular localization of the myc-tagged fusion proteinexpressed upon transient transfection with pMT21-SIAE of COS7cells. Laser confocal microscopy analysis showed SIAE associatedwith intracellular tubular and vesicular structures (Figure 7). Co-localization experiments using PDI and GM130 as typical ER andGolgi complex markers, respectively, demonstrated the presence ofthe enzyme in both membranous structures (Figure 7A and B). Pres-ence of SIAE in these subcellular compartments is in agreement withthe results previously reported i.e. the enzyme is a glycosylated andsecreted protein. In contrast to what previously reported on lysosomalSIAE (Takematsu et al. 1999; Zhu et al. 2004), no SIAE signal wasdetected in vesicular structures positive for LAMP1 (Figure 7A), atypical lysosomal marker.

Discussion

Despite the interest on SIAE and its role in a variety of importantbiological events, several features of the enzyme remain unsolved(Mandal et al. 2012). Here, we describe the characterization of the en-zyme encoded by SIAE transcript variant 1 (NM_170601.4) that,based on the literature (Zhu et al. 2004), encodes for the so-calledlysosomal isoform of sialic acid esterase (Lse). The in silico analysiswe performed strongly indicates that this is the prevalent transcript

Fig. 6. Expression of SIAE mutants in COS7 cells. Enzyme activity (upper panel) and representative western-blot analysis (lower panel) of (A) crude cell extracts

(10 µg) and (B) culture media (20 µL) obtained from nontransfected cells (NT), cells transiently expressing for 48 h the wild-type (WT) SIAE and T132A, S332A

and H377A mutants. Alanine substitution at position 132 and 377 resulted in a significant reduction in SIAE activity (P < 0.005, marked with ***).



isoform expressed in most human tissues. These data suggest that the523 amino acids SIAE protein is responsible for the acetyl esterase ac-tivity on Neu5Ac in most tissues and cell types.

Considering the number of studies on genetic variants of SIAE andtheir association with an increased risk for various autoimmune dis-eases, we also compiled a catalog of single-nucleotide variants frompublic repository of large next-generation sequencing projects. In par-ticular, we focused on variants of residues that are potentially relevantfor the biological activity of the enzyme (i.e. post-translational modi-fication or catalysis). Our analysis indicates that these residues presenteither no sequence variation in the data set analyzed of >60,000 indi-viduals, or the frequency of variants is very low (<0.001212, see Sup-plementary data, Table SV), supporting the functional role of the sitesconsidered.

We also evaluated the intolerance score of SIAE, based on an ap-proach that assesses whether genes have relatively more or less func-tional genetic variation than expected based on the apparently neutralvariation found in the gene (Petrovski et al. 2013). Interestingly, theSIAE gene has a residual variation intolerance score of 0.38, which in-dicates the presence of an excess of common functional variation, a

feature that is often shared by genes linked to immunological disor-ders.

Phylogenetic analysis of SIAE proteins in Metazoa revealed thepresence of 11 highly conserved sequence blocks (Figure 1; Supple-mentary data, Figure S5). Among them, the alignment with alreadycharacterized esterases allowed the identification of the four blockstypical of the SGNH hydrolases (Figure 1) (Molgaard et al. 2000)and of the two essential residues S127 and H377. The serine and his-tidine residues are always present in the SGNH hydrolases. The cata-lytic triad is completed by an acidic residue or the main-chain carbonylgroup of serine, threonine or tryptophan, that participate in the chargerelay system necessary for the activation of the serine nucleophile(Akoh et al. 2004). This third member is less conserved and variesacross different SGNH hydrolases (Wei et al. 1995; Rangarajanet al. 2011).

Homology modeling of the human SIAE based on the knownstructures of esterases from Arabidopsis thaliana (Bitto et al. 2005),E. coli (Rangarajan et al. 2011) and Clostridium acetobutilicum(PDB 1ZMB) allowed the reconstruction of the enzyme structure (Sup-plementary data, Figure S6). Evaluation of this model revealed that the

Fig. 7. Subcellular localization of SIAE in COS7 cells. Cells were transiently transfected with pMT21-SIAE-Myc for 48 h and subjected to immunofluorescence

staining and confocal microscopy analysis. To analyze lysosomal (A), ER (B) and Golgi apparatus (C) localization cells were double stained using anti-PDI,

anti-GM130 and anti-LAMP1 antibodies, respectively. Scale bar: 10 μm. For details, see Material and methods. This figure is available in black and white in print

and in colour at Glycobiology online

1000 F Orizio et al.








human protein is composed by a core portion (aa 113–401) (Figure 3Band C) highly similar to those of other esterases, and a C-terminal do-main (L402–K523) containing three α-helices that poorly aligns withtemplate structures. The core structure of human SIAE is composed by1 anti-parallel and 6 parallel β-sheets surrounded by 12 α-helices (Fig-ure 3), a structure peculiar of SGNH hydrolase, family II (Molgaardet al. 2000; Lo et al. 2003), further supporting its classification inthis protein family. The placement of the two essential residues S127and H377 in the structural model is compatible with their catalyticalfunction. These residues, together with residues from conserved block5 (G219–S228), define a surface suitable to represent the enzyme ac-tive site (Figure 4). Molecular docking analysis using SwissDockplaced 9-O-acetyl-Neu5Ac substrate in the predicted active site asthe fourth most favorable result. However, in the top three predictedcomplexes, the substrate molecule interacts with the unstructuredC-terminal of the human SIAE and thus the predicted active site repre-sents the best placement in the enzyme core structure. Using PyMol toassess residues near the substrate molecule, we were able to identify 12residues putatively involved in interactions with the substrate that couldplay a role in substrate specificity (marked with + in Figure 1). Overall,the predicted active site consists of a solvent-exposed shallow pocket,similar to that ofNanS fromE. coli (Rangarajan et al. 2011) that is con-sistent with the binding of various complex substrates containing acety-lated sialic acid (Figure 4A). We then analyzed the predicted active siteto identify the actual catalytic triad. The residues S127 and H377 arefully conserved in Metazoa and the two residues T132 and S332 areplaced near H377, making them good candidates to fulfill the role ofthe third member in the catalytic triad (Figure 4B). Mutation of theS127 residue has already been described to severely reduce SIAE en-zyme activity (Surolia et al. 2010), supporting its essential role for ca-talysis. Using site-directed mutagenesis, we investigated thecontribution of the other putative members of the active triad and de-monstrated, upon overexpression of the mutants H377A and T132A,that both are essential for enzyme activity (Figure 6). These findingsdefine the actual catalytic triad of human SIAE to be composed byS127, T132 and H377 residues (Figure 4B).

Bioinformatics analysis also predicted 6 N-glycosylation and 10O-glycosylation sites (Table I and Supplementary data, Tables SII–SIV). Glycosylation is a key modification for secreted proteins(Benham 2012) and our data suggest that these modifications arealso involved in SIAE enzymatic activity. Various glycosylation pat-terns have been demonstrated occurring on human SIAE (Yin et al.2013) and the N402 and N422 glycosylation sites have been also ex-perimentally validated by previous studies (Chen et al. 2009; Tanget al. 2010). However, our analysis revealed that these two sites areweakly conserved inMetazoa, with the N422 conserved only in mam-mals (Figure 2A) and N402 only in primates. The predicted glycosyla-tion sites N138 and S177 were fully conserved inMetazoa (Figure 2Aand B) and thus represent good candidates for future studies on SIAEglycosylation. Moreover, all of the most conserved sites, as well asN402 and N422 that have been demonstrated to be glycosylated invivo, are exposed on the protein surface in our 3Dmodel (Supplemen-tary data, Figure S7), supporting both their role as potential glycosy-lation sites and the overall likelihood of the proposed model structure.To complete the in silico characterization of the protein, we alsopredicted phosphorylation sites (Table I and Supplementary data,Table SIV). Although the role of phosphorylation for secreted proteinsis not well understood, this post-translational modification could beinvolved in protein secretion (Tagliabracci et al. 2013) as phosphory-lated proteins are present in the secretome (Carrascal et al. 2010).Moreover, recent studies have also demonstrated the presence of active

kinases in the extracellular environment that act on various secretedproteins to modulate their biological activity (Tagliabracci et al.2012). Our analysis identified five top conserved phosphorylationsites and five additional sites conserved throughout Metazoa(Figure 2C). These are strong candidate sites for phosphorylation ofhuman SIAE and could be involved in regulation of the enzyme activ-ity and/or subcellular localization. In addition, an N-terminal signalpeptide is predicted in silico and reported in the Signal Peptide Data-base and SecretomeP 2.0 server predicts human SIAE as a secretedprotein.

The expression of the myc-tagged protein chimera in fibroblast-like COS7 cells provided interesting information. First of all, usingthis cell line, the enzyme is to the largest extent secreted into the culturemedia. Accordingly, immunofluorescence staining demonstrates ERand Golgi apparatus localization, overlapping the typical secretionpathway (Figure 6). Our finding is in agreement with recent results ob-tained by using a (glyco-)proteomic approach on the secretome ofhuman endothelial cells (Yin et al. 2013), as well as with those ob-tained in COS7 (Guimaraes et al. 1996) and U2OS cells (Cariappaet al. 2009) upon expression of the SIAE of mouse origin and inHEK293T cells with the enzyme of human origin (Chellappa et al.2013). Concerning the secretion mechanism(s), SIAE has been foundin urinary exosomes (Gonzales et al. 2009; Principe et al. 2013;Prunotto et al. 2013). In detail, expression of human SIAE in COS7cells, inhibition of its glycosylation and the block of its synthesis de-monstrated that the enzyme is either processed to a low-molecularweight form, corresponding to the 49 kDa band, or is glycosylatedgiving rise to the biologically active form of the enzyme, correspond-ing to the 62 kDa band.Moreover, we demonstrated that only the gly-cosylated, 62 kDa high-molecular-weight form of SIAE is secreted intothe medium and PNGase-F treatment under non-denaturing conditionsuggests that glycosylation represents an important step for the cata-lytic activity of the enzyme.

SIAE glycosylation has been initially demonstrated in the mouseform (Guimaraes et al. 1996) and more recently in the human proteinand, in particular, for N401 andN422 residues through a detailed gly-coproteomic analysis of liver tissue (Chen et al. 2009). In the case ofN401, glycosylation has been further confirmed in human hepatocel-lular carcinoma cell line 7703 (Tang et al. 2010). In this perspective,SIAE is expected to act mainly on cell surface or extracellular sub-strates bearing 9-O-acetylated sialic acid residues on their glycan moi-eties. Surprisingly and despite the literature (Butor et al. 1995; Zhuet al. 2004), in our experimental conditions, there is no evidence ofcytosolic or lysosomal localization of SIAE. Our results are in agree-ment with those obtained upon stable transfection of murine SIAE inhuman osteosarcoma cells, where an almost complete lack of lyso-somal localization was observed and the protein was secreted and re-covered on the cell surface (Cariappa et al. 2009). In addition, the invitro pH optimum observed, corresponding to 8.4–8.5, is not consist-ent with an activity of the enzyme within the acidic environment oflysosomes. Moreover, the absence of a typical lysosomal localizationis consistent with the subcellular localization of endogenous SIAE re-ported in The Human Protein Atlas (Uhlen et al. 2015). Nevertheless,one must consider that subcellular localization may be dependent onthe respective cell-type investigated. The SIAE expression levels andthe non-synonymous common variants could somehow influence theoverall behavior of the enzyme (i.e. intracellular localization or secre-tion into the extracellular environment). For example, a study carriedout using 293T human embryonic kidney cell, BJAB and Ramos B-celllines and peripheral blood mononuclear cells showed a peculiarbehavior of the enzyme (Chellappa et al. 2013). Actually, the enzyme









appears mainly located intracellularly, and only wild-type protein andactive variants are secreted upon transfection of the correspondingcDNAs, whereas inactive or less active variants are barely detectableoutside the cells. Despite some contrasting evidences (Muthusamyet al. 2005; Polanski and Anderson 2007), the presence of SIAE hasbeen demonstrated in human colostrum (Palmer et al. 2006) and plas-ma (Farrah et al. 2011), and in this perspective the lack of release intothe extracellular environment could reflect cellular retention due tostructural defects leading to misfolding and to an inactive/less activeenzyme. In COS7 cells, we observed a 3.4-fold ratio between extracel-lular/intracellular SIAE, supporting the notion of an enzyme acting onsubstrates localized in either intracellular vesicles or extracellular/cellsurface environment. A recent, detailed investigation on glycosphingo-lipids in colorectal cancer tissues revealed a series of acetylated gang-liosides that could represent possible natural substrates of the enzyme(Holst et al. 2013). One of the most interesting target molecules forSIAE activity is the ganglioside O-acetylated GD3. Differences inacetylation state of this ganglioside are reported to influence a varietyof biological effects, from apoptosis to extravasation of tumor cells,depending also on its subcellular localization (i.e. cell surface or intra-cellular compartments) (Schauer et al. 2011). Acetylated sialic acidshave been found also in GM1, GD2, GD1, as well as in other ganglio-sides with glycans characterized by up to eight sugars bearing one ortwo sialic acid residues (Holst et al. 2013). Concerning possible glyco-protein substrates, acetylated sialic acid residues have been found incolon mucins, namely in the carbohydrate moieties of MUC1 andMUC2, and a decrease in their acetylation levels is found in colon car-cinoma metastases (Mann et al. 1997). Perhaps, the most prominentprotein influenced by acetylation of sialylated glycoconjugates is thesialic acid-binding immunoglobulin-like CD22, expressed by B cellsand acting as a negative regulator of B-cell activation (Crispin et al.2013). Acetylation in position 9 of sialic acid impairs recognition ofthese glycoconjugates by CD22, making SIAE a key enzyme in theregulation of B-cell response and a valuable candidate in a series ofautoimmune diseases (Gan et al. 2012; Pillai et al. 2012; Varki andGagneux 2012; Pillai 2013; Yamamoto et al. 2014). Overall, our re-sults describe several novel features of human SIAE, showing activityin a range of pH rather high for a typical lysosomal enzyme andsupporting its localization in the ER, along the secretory pathwaycompartments, and mostly as a glycosylated, secreted protein. Add-itionally, our genomic and phylogenetic analysis, together with mo-lecular modeling studies, provide novel insight on SIAE protein, anenzyme that acting on chemical modification such as acetylation, con-tributes to the fine tuning of sialic acid biology.

Materials and methods

In general, standard molecular biology techniques were carried out asdescribed by Green and Sambrook (2012). Reagents and powderswere from Sigma unless otherwise indicated.

SIAE gene investigations

SIAE gene features have been investigated using both the UCSC (http://genome.ucsc.edu, last accessed 31 May 2015) and Ensembl (http://www.ensembl.org/, last accessed 31 May 2015) Genome Browsersusing the February 2009 human reference sequence (GRCh37).

Nucleotide and amino acid sequences were compared with thenon-redundant sequences present at the NCBI using the BLAST algo-rithm (Altschul et al. 1990).

RNA-Seq data have been obtained from the Illumina’s HumanBodyMap 2.0 project made available through Ensembl. In addition,RNA-Seq expression data on human SIAE gene have been retrievedfrom Expression Atlas, a database maintained by EMBL-EBI (http://www.ebi.ac.uk/gxa/home, last accessed 31 May 2015).

The catalog of SNVs of the SIAE gene shown in Supplementarydata, Table SV has been generated using data from the ExAC, Cam-bridge, MA (http://exac.broadinstitute.org, last accessed 31 May2015), as of February 2015.

Multiple sequence alignment and phylogenetic analysis

Human SIAE protein sequence (NP_733746.1) was used as query inDELTA-BLAST search to retrieve the putative orthologs of SIAE in 27different organisms throughout Eukarya, plus five bacterial species(Supplementary data, Table SI). Multiple alignment of the 32 proteinsequences was constructed with MUSCLE and then manually refinedwith JalView (Waterhouse et al. 2009), which was also used for thecalculation of conservation score at each position. Highly conservedblocks were defined as protein regions with at least 3 residues showingconservation score >8 and with no >2 consecutive residues scoringbelow the threshold. Visualization of the multiple alignment andLOGO calculation was performed using LaTex and the TexShadepackage v1.22 (Beitz 2000). Specific conservation of residues in fourdifferent evolutionary groups (Metazoa, Chordata, Vertebrata andMammalia) was evaluated using subfamily LOGO calculation (Beitz2006) with a significance threshold of 2 bits. Protein sequences weregrouped accordingly to the taxonomic groups defined in Supplemen-tary data, Table SI. The presence ofN- orO-glycosylation sites on thehuman SIAE protein was predicted using NetNGlyc 1.0 (http://www.cbs.dtu.dk/services/NetNGlyc/, last accessed 31 May 2015) and Ne-tOGlyc 4.0 (Steentoft et al. 2013), while phosphorylation sites werepredicted using NetPhos 2.0 (Blom et al. 1999). The presence ofN-terminal signal peptide was predicted using SignalP 4.1 (Petersenet al. 2011) and the overall probability of being a secreted proteinwas assessed using Secretome 2.0 (Bendtsen et al. 2004).

In addition, themouseNP_035864 sequencewas used as a query in aBLAST search to retrieve 11 additional SIAE polypeptides from rodents.

Molecular modeling

Human SIAE protein sequence (NP_733746.1) was used in DELTA-BLAST against PDB database to retrieve suitable template structuresfor homology modeling. FASTA sequences of the three top hits(2APJ, 3PT5 and 1ZMB) were aligned with human SIAE using MUS-CLE and the alignment was manually refined to match the conserva-tion blocks identified from the multi-species alignment.

The refined alignment was then used as template in Modeller v.9.12(Sali and Blundell 1993) to predict a structural model of human SIAEbased on the crystal structure of 2APJ, 3PT5 and 1ZMB, using tworounds of automatic loop refinement. Putative 3D structure of humanSIAE was saved in pdb format and then visualized using PyMol v.1.7(The PyMOLMolecular Graphics System, Version 1.5.0.4 Schrödinger,LLC). Image of the topology of the secondary structure was generatedusing Pro-Origami (Stivala et al. 2011) based on the pdb file. Multiplealignment with known protein structures from the three esterases 2APJ,3PT5 and 1ZMB was used for the identification in human SIAE of thefour sequence blocks characterizing family II SGNH hydrolases (ester-ase blocks) and of the residues involved in the catalytic process.

We analyzed the position of these four conserved blocks and the fourputative catalytic residues in the predicted structure of human SIAE toreconstruct the hypothetical conformation of the catalytic crevice.



http://genome.ucsc.edu






http://www.ensembl.org/






http://www.ebi.ac.uk/gxa/home









http://exac.broadinstitute.org








http://www.cbs.dtu.dk/services/NetNGlyc/







To further support this region as the enzyme active site, we per-formed an automated molecular docking to identify the best inter-action site between the human SIAE-predicted structure and theenzyme substrate (9-O-acetyl-Neu5Ac). The structure of the enzymesubstrate was generated in mol2 file format using Chemtool v.1.6.14and then used in SwissDock (Grosdidier et al. 2011), together with thepredicted structure of human SIAE. The automated blind docking wasperformed allowing a 3 Å optimization of the input structures. Thebest predicted complexes (lower ΔG value) were examined in PyMolto evaluate the interaction between the substrate, the two putativecatalytic residues (S127 and H377) and the surface defined by thefour-esterase blocks.We also selected as possibly involved in the activesite all those residues placed within 5 Å from the substrate in themolecular model. Analysis of the surface defined by these residuesallowed a better definition of the putative active site conformation.

Cloning and expression of SIAE in COS7 cells

A Mammalian Gene Collection cDNA clone corresponding to theSIAE lysosomal transcript (Lse) (MGC:87009; IMAGE:5278370)has been obtained from Geneservice Ltd. The coding region was amp-lified by PCR using cloned Pfu Turbo DNA polymerase (Stratagene).The resulting PCR product was then digested with the appropriate re-striction enzymes and cloned into pMT21 vector (Kolodkin et al.1997) to generate an open reading frame (ORF)-encoding SIAE witha C-terminal myc-tag. The following forward (F) and reverse (R)oligonucleotides were used in PCR amplification of SIAE ORF,HsSIAE F and HsSIAE R, bearing EcoRI and SalI restriction site, re-spectively, in order to obtain a directional cloning of the amplifiedproduct (Supplementary data, Table SVI). Amplification was carriedout according to cycling parameters indicated for Pfu Turbo DNApolymerase for 35 cycles. Annealing step was performed at 55°C.The resulting recombinant vector (pMT21-Hs SIAE-Myc) waschecked by automated DNA sequencing. pMT21-Hs SIAE-Mycconstruct was used to transfect COS7 cells (ATCC® Number: CRL-1651™) kept in culture with Dulbecco’s modified Eagle’s medium(Gibco-BRL) supplemented with 10% (v/v) fetal calf serum, 4 mML-glutamine, 100 U/mL penicillin and 100 µg/mL streptomycin. Fortransient transfection, cells were seeded in 60 mm diameter Petridish or on sterile glass coverslips (VWR International) in 24-wellplates (Corning Life Science) and grown for 24 h before transfection.Transfection experiments were performed overnight in serum-freemedium (OptiMEM, Gibco-BRL) using 2 μg of plasmid DNA and Fu-GENE HD (Promega) reagent according to manufacturer’s guideline.Untransfected cells were used as negative controls. After 48 h expres-sion, cells were washed in PBS (Gibco-BRL), collected by scraping, re-suspended in sterile PBS supplemented with protease inhibitors(Protease Inhibitor Cocktail, Roche) and subjected to mild sonication(two 15″ pulse at 4°C, 15% power, using a probe Sonicator Sonoplus-Bandelin electronic). The supernatant obtained after a centrifugationat 800 × g for 10 min represented the crude cell extract and was usedfor protein quantitation, enzyme assay and western-blot analysis.

Site-directed mutagenesis

Using the Quick-change II Site-directed Mutagenesis kit (Agilent) andspecific mutagenesis primers (Supplementary data, Table SVI), wemodified the three amino acids putatively involved in the enzymeactive site. Starting from 50 ng of pMT21-SIAE representing the wild-type form of the enzyme, we generated three mutant constructscontaining the following amino acid substitutions: T132A (pMT21T132A), S332A (pMT21-S332A) and H377A (pMT21-H377A).

Each construct was cloned in XL-1 E. coli, purified with QIAfilterPlasmidMidi kit (Qiagen) and sequence verified. The resulting recom-binant plasmids were used to transiently transfect COS7 cells asdescribed in the previous paragraph.

Western-blot analysis

Media and cell extract samples were subjected to 12% (w/v) SDS–PAGE and subsequently transferred by electroblotting ontoHybond-Pblotting membrane (GE Life Sciences). Membranes were blocked for30 min in PBS containing 0.1% (v/v) Tween 20 (PBST) and 5% (w/v)dried milk and subsequently incubated for 1 h at 4°C with rabbitanti-Myc primary antibody (Millipore), diluted 1 : 500 in PBST con-taining 1% (w/v) dried milk. After several washes with PBST, mem-branes were incubated for 45 min at room temperature (RT) withdonkey anti-rabbit HPR-conjugated secondary antibody (GE LifeSciences), diluted 1 : 5000 in PBST. After final washing in PBST, detec-tion of immunocomplexes was carried out using SuperSignal WestPico Chemiluminescent Substrate detection kit (Thermo Scientific).

Treatment with tunicamycin, cycloheximide

and PNGase-F

In vivo glycosylation was analyzed incubating COS7 cells with Tuni-camycin (Sigma). Briefly, COS7 cells were transfected with pMT21-HsSIAE-Myc and after 24 h expression tunicamycin (2 µg/mL) wasadded to the cell medium and cells were grown for further 24 h.Cells andmediawere then collected and subjected towestern-blot ana-lysis as previously described. To clarify if the secretion of the proteinwas the normal fate of the SIAE or if it was just due to the overexpres-sion, after 24 h from the transfection we treated the COS7 cells withcycloheximide (75 µM; Sigma) and let them grow for additional 24 h.Any modifications in expression or glycosylation of SIAE protein weredetected by western blot as previously described.

Finally, to determinate if glycosylation is necessary for the activityof SIAE, cells extract were treated with PNGase-F (NEB) in native con-dition. Briefly, cells extracts corresponding to 20 µg were incubatedover night at 37°C in G7 reaction buffer (0.5 M sodium phosphate,pH 7.5; NEB) with or without PNGase-F. The enzyme activities ofthe samples were then analyzed as described below, while glycosylatedand unglycosylated SAIE forms were detected by western blot.

Untransfected cells were treated as above and used as control.

SIAE enzymatic assay

The enzymatic activity of SIAE in total cell lysates and in cell lysate andin media was determined using the general esterase artificial substrate4-nitrophenyl acetate (pNPA) (Schauer et al. 1988; Higa et al. 1989).Briefly, pNPA dissolved in acetonitrile (80 mM) was added to 200 µLof phosphate buffer (0.1 M), pH 8.4, to a 0.6 mMpNPA final concen-tration. For SIAE activity in cell extracts, 60 μg proteins were used; forSIAE activity in growth media, 15 µL were used as enzyme source. Re-lease of 4-nitrophenol (pNP) was followed at 20°C at 405 nm using anELISA plate reader (Scientific Technology, Inc.). Enzyme activity wascalculated using the extinction coefficient of pNP corresponding to18.5 mM−1/cm−1 and expressed as nmol/min−1 (mU) mg−1 and mU/mL−1 for cell extracts and media, respectively.

Indirect immunofluorescence analysis

COS7 cells were grown on coverslips and transfected as previously de-scribed. After 48 h, cells were rinsed with ice-cold PBS, immediatelyfixed with 3% (w/v) paraformaldehyde in PBS for 15 min at RT and





washed three times in PBS. Paraformaldehyde was quenched incubat-ing samples with 50 mM NH4Cl in PBS for 15 min. After three PBSwashes, cells were permeabilized with 0.3% saponin in PBS (PBS/Sap) for 30 min and then incubated with primary antibodies dilutedin PBS/Sap for 1 h at RT. The following primary antibodies havebeen used: rabbit anti-Myc for SIAE (1 : 400, Millipore), mouseanti-GM130 for Golgi complex (1 : 300, BD), mouse anti-PDI forER (1 : 400, Stressgen) andmouse anti-LAMP1 for lysosomes (1 : 300,BD). After washing three times in PBS/Sap, cells were incubated for45 min at RT with donkey anti-rabbit Alexa 555 and donkey anti-mouse Alexa 488 secondary antibodies (1 : 400, Molecular Probes,Life Technology). Finally, after three washes with PBS/Sap followedby three washes with PBS, specimens were mounted using FluorescentMountingMediun (Dako Cytomation), containing DAPI for nuclei la-beling. Specimens were examined using a Confocal Laser System LSM510 META (Zeiss) equipped with Axiovert 200 microscope (Zeiss).Image processing was performed with Adobe Photoshop software.

Supplementary data

Supplementary data for this article are available online at http://glycob.oxfordjournals.org/.

Funding

This work was partly supported by Regione Lombardia and aNetwork Enable Drug Discovery (NEDD) grant to E.M. The financialsupport of Ministero dell’Istruzione, dell’Università e della Ricerca(MIUR) is acknowledged.

Conflict of interest statement

All the authors have no conflict of interest to declare.

Abbreviations

BCR, B-cell receptor; Cse, cytosolic sialic acid esterase; ESTs, expressedsequence tags; ExAC, Exome Aggregation Consortium; FCS, fetal calf serum;Lse, lysosomal sialic acid esterase; NCBI, National Center for BiotechnologyInformation; ORF, open reading frame; PNGase, peptide-N-glycosidase-F;pNP, 4-nitrophenol; RT, room temperature; SIAE, sialic acid acetyl esterase;SNVs, single-nucleotide variants; SOAT, sialate O-acetyltransferase.

References

Akoh CC, Lee GC, Liaw YC, Huang TH, Shaw JF. 2004. GDSL family of serineesterases/lipases. Prog Lipid Res. 43:534–552.

Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ. 1990. Basic local align-ment search tool. J Mol Biol. 215:403–410.

Arming S, Wipfler D, Mayr J, Merling A, Vilas U, Schauer R,Schwartz-Albiez R, Vlasak R. 2011. The human Cas1 protein: A sialic acid-specific O-acetyltransferase? Glycobiology. 21:553–564.

Beitz E. 2006. Subfamily logos: Visualization of sequence deviations at align-ment positions with high information content. BMC Bioinformatics. 7:313.

Beitz E. 2000. TEXshade: Shading and labeling of multiple sequence alignmentsusing LATEX2 epsilon. Bioinformatics. 16:135–139.

Bendtsen JD, Jensen LJ, Blom N, Von Heijne G, Brunak S. 2004. Feature-basedprediction of non-classical and leaderless protein secretion. Protein EngDesign Selection. 17:349–356.

Benham AM. 2012. Protein secretion and the endoplasmic reticulum. ColdSpring Harbor Perspect Biol. 4:a012872.

Bitto E, Bingman CA, McCoy JG, Allard ST, Wesenberg GE, Phillips GN Jr.2005. The structure at 1.6 Angstroms resolution of the protein product ofthe At4g34215 gene from Arabidopsis thaliana. Acta Crystallogr D BiolCrystallogr. 61:1655–1661.

Blom N, Gammeltoft S, Brunak S. 1999. Sequence and structure-based predic-tion of eukaryotic protein phosphorylation sites. J Mol Biol. 294:1351–1362.

Butor C, Diaz S, Varki A. 1993. High level O-acetylation of sialic acids onN-linked oligosaccharides of rat liver membranes. Differential subcellulardistribution of 7- and 9-O-acetyl groups and of enzymes involved in theirregulation. J Biol Chem. 268:10197–10206.

Butor C, Griffiths G, Aronson NN Jr, Varki A. 1995. Co-localization of hydro-lytic enzymes with widely disparate pH optima: Implications for the regula-tion of lysosomal pH. J Cell Sci. 108(Pt 6):2213–2219.

Cariappa A, Takematsu H, Liu H, Diaz S, Haider K, Boboila C, Kalloo G,Connole M, Shi HN, Varki N, et al. 2009. B cell antigen receptor signalstrength and peripheral B cell development are regulated by a 9-O-acetylsialic acid esterase. J Exp Med. 206:125–138.

Carrascal M, Gay M, Ovelleiro D, Casas V, Gelpi E, Abian J. 2010. Character-ization of the human plasma phosphoproteome using linear ion trap massspectrometry and multiple search engines. J Proteome Res. 9:876–884.

Chellappa V, Taylor KN, Pedrick K, Donado C, Netravali IA, Haider K,Cariappa A, Dalomba NF, Pillai S. 2013. M89V sialic acid acetyl esterase(SIAE) and all other non-synonymous common variants of this gene arecatalytically normal. PLoS ONE. 8:e53453.

Chen R, Jiang X, Sun D, Han G, Wang F, Ye M, Wang L, Zou H. 2009.Glycoproteomics analysis of human liver tissue by combination of multipleenzyme digestion and hydrazide chemistry. J Proteome Res. 8:651–661.

Crispin JC, Hedrich CM, Tsokos GC. 2013. Gene-function studies in systemiclupus erythematosus. Nat Rev Rheumatol. 9:476–484.

de Groot RJ. 2006. Structure, function and evolution of the hemagglutinin-esterase proteins of corona- and toroviruses. Glycoconj J. 23:59–72.

Diaz S, Higa HH, Hayes BK, Varki A. 1989. O-acetylation and de-O-acetyl-ation of sialic acids. 7- and 9-o-acetylation of alpha 2,6-linked sialic acidson endogenous N-linked glycans in rat liver Golgi vesicles. J Biol Chem.264:19416–19426.

Dorner T, Shock A, Smith KG. 2012. CD22 and autoimmune disease. Int RevImmunol. 31:363–378.

Erdmann M, Wipfler D, Merling A, Cao Y, Claus C, Kniep B, Sadick H,Bergler W, Vlasak R, Schwartz-Albiez R. 2006. Differential surface expres-sion and possible function of 9-O- and 7-O-acetylated GD3 (CD60 b and c)during activation and apoptosis of human tonsillar B and T lymphocytes.Glycoconj J. 23:627–638.

Farrah T, Deutsch EW, Omenn GS, Campbell DS, Sun Z, Bletz JA, Mallick P,Katz JE, Malmstrom J, Ossola R, et al. 2011. A high-confidence humanplasma proteome reference set with estimated concentrations in PeptideA-tlas. Mol Cell Proteomics. 10:M110 006353.

Gan EH, MacArthur K, Mitchell AL, Pearce SH. 2012. The role of functionallydefective rare germline variants of sialic acid acetylesterase in autoimmuneAddison’s disease. Eur J Endocrinol. 167:825–828.

Genome 10K Community of Scientists. 2009. Genome 10K: A proposal to ob-tain whole-genome sequence for 10,000 vertebrate species. J Heredity.100:659–674.

Gonzales PA, Pisitkun T, Hoffert JD, Tchapyjnikov D, Star RA, Kleta R,WangNS, KnepperMA. 2009. Large-scale proteomics and phosphoproteo-mics of urinary exosomes. J Am Soc Nephrol. 20:363–379.

Green MR, Sambrook J. 2012. Molecular Cloning: A Laboratory Manual. 4rded. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press.

Grosdidier A, Zoete V,Michielin O. 2011. SwissDock, a protein-small moleculedocking web service based on EADock DSS. Nucleic Acids Res. 39:W270–W277.

Guimaraes MJ, Bazan JF, Castagnola J, Diaz S, Copeland NG, Gilbert DJ,Jenkins NA, Varki A, Zlotnik A. 1996.Molecular cloning and characteriza-tion of lysosomal sialic acid O-acetylesterase. J Biol Chem. 271:13697–13705.

Herrler G, Klenk HD. 1991. Structure and function of the HEF glycoprotein ofinfluenza C virus. Adv Virus Res. 40:213–234.






Higa HH, Diaz S, Varki A. 1987. Biochemical and genetic evidence for distinctmembrane-bound and cytosolic sialic acid O-acetyl-esterases: Serine-active-site enzymes. Biochem Biophys Res Commun. 144:1099–1108.

Higa HH,Manzi A, Varki A. 1989. O-acetylation and de-O-acetylation of sialicacids. Purification, characterization, and properties of a glycosylated ratliver esterase specific for 9-O-acetylated sialic acids. J Biol Chem. 264:19435–19442.

Hirschfield GM, Xie G, Lu E, Sun Y, Juran BD, Chellappa V, Coltescu C,Mason AL,Milkiewicz P, Myers RP, et al. 2012. Association of primary bil-iary cirrhosis with variants in the CLEC16A, SOCS1, SPIB and SIAE immu-nomodulatory genes. Genes Immunity. 13:328–335.

Holst S, Stavenhagen K, Balog CI, Koeleman CA, McDonnell LM,Mayboroda OA, Verhoeven A, Mesker WE, Tollenaar RA, Deelder AM,et al. 2013. Investigations on aberrant glycosylation of glycosphingolipidsin colorectal cancer tissues using liquid chromatography and matrix-assisted laser desorption time-of-flight mass spectrometry (MALDI-TOF-MS). Mol Cell Proteomics. 12:3081–3093.

Hunt KA, Smyth DJ, Balschun T, Ban M, Mistry V, Ahmad T, Anand V,Barrett JC, Bhaw-Rosun L, Bockett NA, et al. 2012. Rare and functionalSIAE variants are not associated with autoimmune disease risk in up to66,924 individuals of European ancestry. Nat Genetics. 44:3–5.

JiangH, Blouin C. 2007. Insertions and the emergence of novel protein structure:A structure-based phylogenetic study of insertions. BMC Bioinformatics.8:444.

Kniep B, Kniep E, Ozkucur N, Barz S, Bachmann M, Malisan F, Testi R,Rieber EP. 2006. 9-O-acetyl GD3 protects tumor cells from apoptosis. IntJ Cancer. 119:67–73.

Kolodkin AL, Levengood DV, Rowe EG, Tai YT, Giger RJ, Ginty DD. 1997.Neuropilin is a semaphorin III receptor. Cell. 90:753–762.

Lo YC, Lin SC, Shaw JF, Liaw YC. 2003. Crystal structure of Escherichia colithioesterase I/protease I/lysophospholipase L1: Consensus sequence blocksconstitute the catalytic center of SGNH-hydrolases through a conservedhydrogen bond network. J Mol Biol. 330:539–551.

Malisan F, Franchi L, Tomassini B, Ventura N, Condo I, Rippo MR, Rufini A,Liberati L, Nachtigall C, Kniep B, et al. 2002. Acetylation suppresses theproapoptotic activity of GD3 ganglioside. J Exp Med. 196:1535–1541.

Mandal C, Schwartz-Albiez R, Vlasak R. 2012. Functions and biosynthesis ofO-acetylated sialic acids. Topics Curr Chem. Epub ahead of print.

Mann B, Klussmann E, Vandamme-Feldhaus V, Iwersen M, Hanski ML,Riecken EO, Buhr HJ, Schauer R, Kim YS, Hanski C. 1997. LowO-acetylation of sialyl-Le(x) contributes to its overexpression in colon car-cinoma metastases. Int J Cancer. 72:258–264.

Molgaard A, Kauppinen S, Larsen S. 2000. Rhamnogalacturonan acetylesteraseelucidates the structure and function of a new family of hydrolases.Structure. 8:373–383.

Mukherjee K, Chava AK, Mandal C, Dey SN, Kniep B, Chandra S, Mandal C.2008. O-acetylation of GD3 prevents its apoptotic effect and promotes sur-vival of lymphoblasts in childhood acute lymphoblastic leukaemia. J CellBiochem. 105:724–734.

Mulder NJ, Kersey P, Pruess M, Apweiler R. 2008. In silico characterization ofproteins: UniProt, InterPro and Integr8. Mol Biotechnol. 38:165–177.

Muthusamy B, Hanumanthu G, Suresh S, Rekha B, Srinivas D, Karthick L,Vrushabendra BM, Sharma S, Mishra G, Chatterjee P, et al. 2005. PlasmaProteome Database as a resource for proteomics research. Proteomics.5:3531–3536.

Palmer DJ, Kelly VC, Smit AM, Kuy S, Knight CG, Cooper GJ. 2006. Humancolostrum: Identification of minor proteins in the aqueous phase by proteo-mics. Proteomics. 6:2208–2216.

Petersen TN, Brunak S, von Heijne G, Nielsen H. 2011. SignalP 4.0: Discrim-inating signal peptides from transmembrane regions. Nat Methods.8:785–786.

Petrovski S, Wang Q, Heinzen EL, Allen AS, Goldstein DB. 2013. Genic intoler-ance to functional variation and the interpretation of personal genomes.PLoS Genetics. 9:e1003709.

Pfeffer JM, Weadge JT, Clarke AJ. 2013. Mechanism of action of Neisseria go-norrhoeae O-acetylpeptidoglycan esterase, an SGNH serine esterase. J BiolChem. 288:2605–2613.

Pillai S. 2013. Rethinking mechanisms of autoimmune pathogenesis. JAutoimmunity. 45:97–103.

Pillai S, Cariappa A, Pirnie SP. 2009. Esterases and autoimmunity: The sialicacid acetylesterase pathway and the regulation of peripheral B cell tolerance.Trends Immunol. 30:488–493.

Pillai S, Netravali IA, Cariappa A, Mattoo H. 2012. Siglecs and immune regu-lation. Annu Rev Immunol. 30:357–392.

Polanski M, Anderson NL. 2007. A list of candidate cancer biomarkers for tar-geted proteomics. Biomarker Insights. 1:1–48.

Principe S, Jones EE, Kim Y, Sinha A, Nyalwidhe JO, Brooks J, Semmes OJ,Troyer DA, Lance RS, Kislinger T, et al. 2013. In-depth proteomic analysesof exosomes isolated from expressed prostatic secretions in urine.Proteomics. 13:1667–1671.

Prunotto M, Farina A, Lane L, Pernin A, Schifferli J, Hochstrasser DF,Lescuyer P,Moll S. 2013. Proteomic analysis of podocyte exosome-enrichedfraction from normal human urine. J Proteomics. 82:193–229.

Rangarajan ES, Ruane KM, Proteau A, Schrag JD, Valladares R, Gonzalez CF,Gilbert M, Yakunin AF, Cygler M. 2011. Structural and enzymaticcharacterization of NanS (YjhS), a 9-O-Acetyl N-acetylneuraminicacid esterase from Escherichia coli O157:H7. Protein Sci. 20:1208–1219.

Sali A, Blundell TL. 1993. Comparative protein modelling by satisfaction of spa-tial restraints. J Mol Biol. 234:779–815.

Schauer R. 2009. Sialic acids as regulators of molecular and cellular interac-tions. Curr Opin Struct Biol. 19:507–514.

Schauer R, Reuter G, Stoll S. 1988. Sialate O-acetylesterases: Key enzymes insialic acid catabolism. Biochimie. 70:1511–1519.

Schauer R, Reuter G, Stoll S, Shukla AK. 1989. Partial purification and charac-terization of sialate O-acetylesterase from bovine brain. J Biochem.106:143–150.

Schauer R, Srinivasan GV, Wipfler D, Kniep B, Schwartz-Albiez R. 2011.O-Acetylated sialic acids and their role in immune defense. Adv Exp MedBiol. 705:525–548.

Shukla AK, Schauer R. 1983. Analysis of sialic acids and study of enzyme reac-tions of sialic acid metabolism by ion-exchange high performance liquidchromatography. In: Popper H, Reutter W, Gudat F, Köttgen E, editors.Structural Carbohydrates in the Liver. Lancaster, UK: MTP Press.p. 565–576.

Steentoft C, Vakhrushev SY, Joshi HJ, Kong Y, Vester-Christensen MB,Schjoldager KT, Lavrsen K, Dabelsteen S, Pedersen NB, Marcos-Silva L,et al. 2013. Precision mapping of the human O-GalNAc glycoproteomethrough SimpleCell technology. EMBO J. 32:1478–1488.

Stivala A, Wybrow M, Wirth A, Whisstock JC, Stuckey PJ. 2011. Automaticgeneration of protein structure cartoons with Pro-origami. Bioinformatics.27:3315–3316.

Stoddart A, Zhang Y, Paige CJ. 1996.Molecular cloning of the cDNA encodinga murine sialic acid-specific 9-O-acetylesterase and RNA expression in cellsof hematopoietic and non-hematopoietic origin. Nucleic Acids Res.24:4003–4008.

Surolia I, Pirnie SP, Chellappa V, Taylor KN, Cariappa A, Moya J, Liu H,Bell DW, Driscoll DR, Diederichs S, et al. 2010. Functionally defectivegermline variants of sialic acid acetylesterase in autoimmunity. Nature.466:243–247.

Tagliabracci VS, Engel JL, Wen J, Wiley SE, Worby CA, Kinch LN, Xiao J,Grishin NV, Dixon JE. 2012. Secreted kinase phosphorylates extracellularproteins that regulate biomineralization. Science. 336:1150–1153.

Tagliabracci VS, Xiao J, Dixon JE. 2013. Phosphorylation of substratesdestined for secretion by the Fam20 kinases. Biochem Soc Trans. 41:1061–1065.

Takematsu H, Diaz S, Stoddart A, Zhang Y, Varki A. 1999. Lysosomal andcytosolic sialic acid 9-O-acetylesterase activities can be encoded by onegene via differential usage of a signal peptide-encoding exon at the N ter-minus. J Biol Chem. 274:25623–25631.

Tang J, Liu Y, Yin P, Yao G, Yan G, Deng C, Zhang X. 2010. ConcanavalinA-immobilized magnetic nanoparticles for selective enrichment of glycopro-teins and application to glycoproteomics in hepatocelluar carcinoma cellline. Proteomics. 10:2000–2014.



Uhlen M, Fagerberg L, Hallstrom BM, Lindskog C, Oksvold P, Mardinoglu A,Sivertsson A, Kampf C, Sjostedt E, Asplund A, et al. 2015. Proteomics.Tissue-based map of the human proteome. Science. 347:1260419.

Varki A. 2008. Sialic acids in human health and disease. Trends Mol Med.14:351–360.

Varki A, Gagneux P. 2012.Multifarious roles of sialic acids in immunity.AnnNY Acad Sci. 1253:16–36.

Varki A, Muchmore E, Diaz S. 1986. A sialic acid-specific O-acetylesterase inhuman erythrocytes: Possible identity with esterase D, the genetic marker ofretinoblastomas and Wilson disease. Proc Natl Acad Sci USA. 83:882–886.

Waterhouse AM, Procter JB, Martin DM, Clamp M, Barton GJ. 2009. JalviewVersion 2–a multiple sequence alignment editor and analysis workbench.Bioinformatics. 25:1189–1191.

Wei Y, Schottel JL, Derewenda U, Swenson L, Patkar S, Derewenda ZS. 1995. Anovel variant of the catalytic triad in the Streptomyces scabies esterase.Nat

Struct Biol. 2:218–223.

Wipfler D, Srinivasan GV, Sadick H, Kniep B, Arming S, Willhauck-FleckensteinM, VlasakR, Schauer R, Schwartz-Albiez R. 2011. Differentiallyregulated expression of 9-O-acetyl GD3 (CD60b) and 7-O-acetyl-GD3(CD60c) during differentiation and maturation of human T and B lympho-cytes. Glycobiology. 21:1161–1172.

Yamamoto M, Iguchi G, Bando H, Fukuoka H, Suda K, Takahashi M,Nishizawa H, Matsumoto R, Tojo K, Mokubo A, et al. 2014. A missensesingle-nucleotide polymorphism in the sialic acid acetyl esterase gene isassociated with anti-PIT-1 antibody syndrome. Endocrine Journal.61:641–644.

Yin X, Bern M, Xing Q, Ho J, Viner R, Mayr M. 2013. Glycoproteomic ana-lysis of the secretome of human endothelial cells. Mol Cell Proteomics.12:956–978.

Zhu H, Chan HC, Zhou Z, Li J, Zhu H, Yin L, XuM, Cheng L, Sha J. 2004. Agene encoding sialic-acid-specific 9-O-acetylesterase found in human adulttestis. J Biomed Biotechnol. 2004:130–136.



Human sialic acid acetyl esterase: Towards a better understanding ...

Documents

Transcript of Human sialic acid acetyl esterase: Towards a better understanding ...