Reinvestigation of the internal glycan rearrangement of Lewis a and blood group type H1 epitopes.

Protonated ions of fucose-containing oligosaccharides are prone to undergo internal glycan rearrangement which results in chimeric fragments that obfuscate mass-spectrometric analysis. Lack of accessible tools that would facilitate systematic analysis of glycans in the gas phase limits our understanding of this phenomenon. In this work, we use density functional theory modeling to interpret cryogenic IR spectra of Lewis a and blood group type H1 trisaccharides and to establish whether these trisaccharides undergo the rearrangement during gas-phase analysis. Structurally unconstrained search reveals that none of the parent ions constitute a thermodynamic global minimum. In contrast, predicted collision cross sections and anharmonic IR spectra provide a good match to available experimental data which allowed us to conclude that fucose migration does not occur in these antigens. By comparing the predicted structures with those obtained for Lewis x and blood group type H2 epitopes, we demonstrate that the availability of the mobile proton and a large difference in the relative stability of the parent ions and rearrangement products constitute the prerequisites for the rearrangement reaction.

N-glycoproteomic analyses of human intestinal enteroids, varying in histo-blood group geno- and phenotypes, reveal a wide repertoire of fucosylated glycoproteins.

Human noroviruses, globally the main cause of viral gastroenteritis, show strain specific affinity for histo-blood group antigens (HBGA) and can successfully be propagated ex vivo in human intestinal enteroids (HIEs). HIEs established from jejunal stem cells of individuals with different ABO, Lewis and secretor geno- and phenotypes, show varying susceptibility to such infections. Using bottom-up glycoproteomic approaches we have defined and compared the N-linked glycans of glycoproteins of seven jejunal HIEs. Membrane proteins were extracted, trypsin digested, and glycopeptides enriched by hydrophilic interaction liquid chromatography and analyzed by nanoLC-MS/MS. The Byonic software was used for glycopeptide identification followed by hands-on verifications and interpretations. Glycan structures and attachment sites were identified from MS2 spectra obtained by higher-energy collision dissociation through analysis of diagnostic saccharide oxonium ions (B-ions), stepwise glycosidic fragmentation of the glycans (Y-ions), and peptide sequence ions (b- and y-ions). Altogether 694 unique glycopeptides from 93 glycoproteins were identified. The N-glycans encompassed pauci- and oligomannose, hybrid- and complex-type structures. Notably, polyfucosylated HBGA-containing glycopeptides of the four glycoproteins tetraspanin-8, carcinoembryonic antigen-related cell adhesion molecule 5, sucrose-isomaltase and aminopeptidase N were especially prominent and were characterized in detail and related to donor ABO, Lewis and secretor types of each HIE. Virtually no sialylated N-glycans were identified for these glycoproteins suggesting that terminal sialylation was infrequent compared to fucosylation and HBGA biosynthesis. This approach gives unique site-specific information on the structural complexity of N-linked glycans of glycoproteins of human HIEs and provides a platform for future studies on the role of host glycoproteins in gastrointestinal infectious diseases.

The disordered protein SERF promotes α-Synuclein aggregation through liquid-liquid phase separation.

The aggregation of α-Synuclein (α-Syn) into amyloid fibrils is the hallmark of Parkinson's disease. Under stress or other pathological conditions, the accumulation of α-Syn oligomers is the main contributor to the cytotoxicity. A potential approach for treating Parkinson's disease involves preventing the accumulation of these α-Syn oligomers. In this study, we present a novel mechanism involving a conserved group of disorderly proteins known as small EDRK-rich factor (SERF), which promotes the aggregation of α-Syn through a cophase separation process. Using diverse methods like confocal microscopy, fluorescence recovery after photobleaching assays, solution-state NMR spectroscopy, and Western blot, we determined that the N-terminal domain of SERF1a plays a role in the interactions that occur during cophase separation. Within these droplets, α-Syn undergoes a gradual transformation from solid condensates to amyloid fibrils, while SERF1a is excluded from the condensates and dissolves into the solution. Notably, in vivo experiments show that SERF1a cophase separation with α-Syn significantly reduces the deposition of α-Syn oligomers and decreases its cellular toxicity under stress. These findings suggest that SERF1a accelerates the conversion of α-Syn from highly toxic oligomers to less toxic fibrils through cophase separation, thereby mitigating the biological damage of α-Syn aggregation.

Species-Specific N-Glycomes and Methylation Patterns of Oysters Crassostrea gigas and Ostrea edulis and Their Possible Consequences for the Norovirus-HBGA Interaction.

Noroviruses, the major cause of acute viral gastroenteritis, are known to bind to histo-blood group antigens (HBGAs), including ABH groups and Lewis-type epitopes, which decorate the surface of erythrocytes and epithelial cells of their host tissues. The biosynthesis of these antigens is controlled by several glycosyltransferases, the distribution and expression of which varies between tissues and individuals. The use of HBGAs as ligands by viruses is not limited to humans, as many animal species, including oysters, which synthesize similar glycan epitopes that act as a gateway for viruses, become vectors for viral infection in humans. Here, we show that different oyster species synthesize a wide range of N-glycans that share histo-blood A-antigens but differ in the expression of other terminal antigens and in their modification by O-methyl groups. In particular, we show that the N-glycans isolated from Crassostrea gigas and Ostrea edulis exhibit exquisite methylation patterns in their terminal N-acetylgalactosamine and fucose residues in terms of position and number, adding another layer of complexity to the post-translational glycosylation modifications of glycoproteins. Furthermore, modeling of the interactions between norovirus capsid proteins and carbohydrate ligands strongly suggests that methylation has the potential to fine-tune the recognition events of oysters by virus particles.

Direct Blockade of the Norovirus Histo-Blood Group Antigen Binding Pocket by Nanobodies.

Noroviruses are the leading cause of outbreaks of acute gastroenteritis. These viruses usually interact with histo-blood group antigens (HBGAs), which are considered essential cofactors for norovirus infection. This study structurally characterizes nanobodies developed against the clinically important GII.4 and GII.17 noroviruses with a focus on the identification of novel nanobodies that efficiently block the HBGA binding site. Using X-ray crystallography, we have characterized nine different nanobodies that bound to the top, side, or bottom of the P domain. The eight nanobodies that bound to the top or side of the P domain were mainly genotype specific, while one nanobody that bound to the bottom cross-reacted against several genotypes and showed HBGA blocking potential. The four nanobodies that bound to the top of the P domain also inhibited HBGA binding, and structural analysis revealed that these nanobodies interacted with several GII.4 and GII.17 P domain residues that commonly engaged HBGAs. Moreover, these nanobody complementarity-determining regions (CDRs) extended completely into the cofactor pockets and would likely impede HBGA engagement. The atomic level information for these nanobodies and their corresponding binding sites provide a valuable template for the discovery of additional "designer" nanobodies. These next-generation nanobodies would be designed to target other important genotypes and variants, while maintaining cofactor interference. Finally, our results clearly demonstrate for the first time that nanobodies directly targeting the HBGA binding site can function as potent norovirus inhibitors. IMPORTANCE Human noroviruses are highly contagious and a major problem in closed institutions, such as schools, hospitals, and cruise ships. Reducing norovirus infections is challenging on multiple levels and includes the frequent emergence of antigenic variants, which complicates designing effective, broadly reactive capsid therapeutics. We successfully developed and characterized four norovirus nanobodies that bound at the HBGA pockets. Compared with previously developed norovirus nanobodies that inhibited HBGA through disrupted particle stability, these four novel nanobodies directly inhibited HBGA engagement and interacted with HBGA binding residues. Importantly, these new nanobodies specifically target two genotypes that have caused the majority of outbreaks worldwide and consequently would have an enormous benefit if they could be further developed as norovirus therapeutics. To date, we have structurally characterized 16 different GII nanobody complexes, a number of which block HBGA binding. These structural data could be used to design multivalent nanobody constructs with improved inhibition properties.

Loci and motifs of the GalNAcα1 → 3/O related glycotopes in the mammalian glycoconjugates and their lectin recognition roles.

Galα1 → and GalNAcα1 → are the two essential key sugars in human blood group AB active glycotopes, in which GalNAcα1 → related sequences are located at both sides of the nonreducing and the reducing ends of human blood group A active O-glycans. It is also found at the nonreducing ends of GlcNAc N-glycans and glycosphingolipid(GSL) of human blood group A active glycotopes (Ah) and Forssman antigen (Fp). When monosaccharides and their α, ß anomers are involved in basic units to express the complex size of the combining sites of the GalNAcα1 → specific lectins, they can be divided into a cavity site to accommodate the GalNAcα → key sugar and a subsite with a wide and broad range of recognition area to adopt the rest part of sugar sequences or glycotopes. The function of the subsite is assumed to act as an enhancement factor to increase its affinity power. The following three points are the theme of this mini review: (1) the loci and distribution of the GalNAcα1 → related glycotopes in mammalian glycoconjugates are illustrated and their chemical structures are advanced by the expression of the disaccharide units and code system; (2) the sizes and motifs of GalNAcα1 → specific lectin-glycan interactions are given and (3) the role of the polyvalent blood group Ah and Bh glycotopes as blood group AB antigens are proposed. These three highlights should provide an essential background required for the advances in this field.

Bovine natural antibody IgM inhibits the binding of human norovirus protruding domain to its HBGA receptors.

Human norovirus (HuNoV) is the primary viral pathogen that causes acute gastroenteritis (AGE) in humans. The protruding (P) domain of HuNoV interacts with cell surface histo-blood group antigens (HBGAs) to initiate infection. Owing to the lack of an effective in vitro culture method and a robust animal model, our understanding of HuNoVs is limited, and as a result, there are no commercial vaccines or antivirals available at present against the virus. In an attempt to develop a preventative measure, we previously identified that bovine colostrum (bCM) contains functional factors that inhibit the binding of HuNoV P domain to its HBGA receptors. In this study, a candidate functional factor in bCM was identified as immunoglobulin M (IgM) using mass spectrometry, followed by database comparison. The natural antibody IgM was further verified to be a functional protein that inhibited HuNoV P protein binding to HBGA receptors through receptor-binding inhibition experiments using bCM, commercial IgM, and fetal bovine serum. Our findings provide a foundation for future development of natural IgM into an antiviral drug, which may help to prevent and/or treat HuNoV infection.

Lewis fucose is a key moiety for the recognition of histo-blood group antigens by GI.9 norovirus, as revealed by structural analysis.

Noroviruses have been identified as major causative agents of acute nonbacterial gastroenteritis in humans. Histo-blood group antigens (HBGAs) are thought to play a major role among the host cellular factors influencing norovirus infection. Genogroup I, genotype 9 (GI.9) is the most recently identified genotype within genogroup I, whose representative strain is the Vancouver 730 norovirus. However, the molecular interactions between host antigens and the GI.9 capsid protein have not been investigated in detail. In this study, we demonstrate that the GI.9 norovirus preferentially binds Lewis antigens over blood group A, B, and H antigens, as revealed by an HBGA binding assay using virus-like particles. We determined the crystal structures of the protruding domain of the GI.9 capsid protein in the presence or absence of Lewis antigens. Our analysis demonstrated that Lewis fucose (α1-3/4 fucose) represents a key moiety for the GI.9 protein-HBGA interaction, thus suggesting that Lewis antigens might play a critical role during norovirus infection. In addition to previously reported findings, our observations may support the future design of antiviral agents and vaccines against noroviruses.

Norovirus-glycan interactions - how strong are they really?

Infection with human noroviruses requires attachment to histo blood group antigens (HBGAs) via the major capsid protein VP1 as a primary step. Several crystal structures of VP1 protruding domain dimers, so called P-dimers, complexed with different HBGAs have been solved to atomic resolution. Corresponding binding affinities have been determined for HBGAs and other glycans exploiting different biophysical techniques, with mass spectrometry (MS) and nuclear magnetic resonance (NMR) spectroscopy being most widely used. However, reported binding affinities are inconsistent. At the extreme, for the same system MS detects binding whereas NMR spectroscopy does not, suggesting a fundamental source of error. In this short essay, we will explain the reason for the observed differences and compile reliable and reproducible binding affinities. We will then highlight how a combination of MS techniques and NMR experiments affords unique insights into the process of HBGA binding by norovirus capsid proteins.

Glycan Recognition in Human Norovirus Infections.

Recognition of cell-surface glycans is an important step in the attachment of several viruses to susceptible host cells. The molecular basis of glycan interactions and their functional consequences are well studied for human norovirus (HuNoV), an important gastrointestinal pathogen. Histo-blood group antigens (HBGAs), a family of fucosylated carbohydrate structures that are present on the cell surface, are utilized by HuNoVs to initially bind to cells. In this review, we describe the discovery of HBGAs as genetic susceptibility factors for HuNoV infection and review biochemical and structural studies investigating HuNoV binding to different HBGA glycans. Recently, human intestinal enteroids (HIEs) were developed as a laboratory cultivation system for HuNoV. We review how the use of this novel culture system has confirmed that fucosylated HBGAs are necessary and sufficient for infection by several HuNoV strains, describe mechanisms of antibody-mediated neutralization of infection that involve blocking of HuNoV binding to HBGAs, and discuss the potential for using the HIE model to answer unresolved questions on viral interactions with HBGAs and other glycans.

The structure of a family 110 glycoside hydrolase provides insight into the hydrolysis of α-1,3-galactosidic linkages in λ-carrageenan and blood group antigens.

α-Linked galactose is a common carbohydrate motif in nature that is processed by a variety of glycoside hydrolases from different families. Terminal Galα1-3Gal motifs are found as a defining feature of different blood group and tissue antigens, as well as the building block of the marine algal galactan λ-carrageenan. The blood group B antigen and linear α-Gal epitope can be processed by glycoside hydrolases in family GH110, whereas the presence of genes encoding GH110 enzymes in polysaccharide utilization loci from marine bacteria suggests a role in processing λ-carrageenan. However, the structure-function relationships underpinning the α-1,3-galactosidase activity within family GH110 remain unknown. Here we focus on a GH110 enzyme (PdGH110B) from the carrageenolytic marine bacterium Pseudoalteromonas distincta U2A. We showed that the enzyme was active on Galα1-3Gal but not the blood group B antigen. X-ray crystal structures in complex with galactose and unhydrolyzed Galα1-3Gal revealed the parallel ß-helix fold of the enzyme and the structural basis of its inverting catalytic mechanism. Moreover, an examination of the active site reveals likely adaptations that allow accommodation of fucose in blood group B active GH110 enzymes or, in the case of PdGH110, accommodation of the sulfate groups found on λ-carrageenan. Overall, this work provides insight into the first member of a predominantly marine clade of GH110 enzymes while also illuminating the structural basis of α-1,3-galactoside processing by the family as a whole.

Monoclonal antibody specific to the Dia blood group antigen generated by phage display technology.

BACKGROUND: Alloanti-Dia can be implicated in mild to severe blood transfusion reactions. Given the concomitance of a high prevalence of the Dia antigen and antibody circulating in some populations, an anti-Dia typing reagent is required in order to enable safe blood transfusions. Limitations of hybridoma technology to produce such a reagent led to the use of phage display technology to generate an anti-Dia monoclonal antibody. MATERIALS AND METHODS: A library of phages displaying murine single-chain variable fragment antibody (scFv-phages) was consecutively adsorbed with different panels of Di(a-b+) red cells to eliminate scFc-phages that potentially bind irrelevant blood group antigens. Thereafter, the subtractive library was specifically selected for the scFv-phages that bound Dia antigen by sequentially biopanning the library with Di(a+b+) cell ghosts and Di(a+b-) intact red cells. A specific interaction between the selected scFv-phages and Dia epitope was validated with the Dia peptide by a competitive haemagglutination inhibition assay and confirmed with the red cells by flow cytometry. RESULTS: An scFv-phage clone specifically bound the Dia epitope, as shown by its binding competition with the human anti-Dia to the Dia peptide in a haemagglutination inhibition test. Moreover, it was highly reactive to Di(a+b+) red cells but not to Di(a-b+) red cells, as determined by flow cytometry. DISCUSSION: In this study, a Dia-specific scFv-phage antibody was successfully produced. The selection protocol might be a prototypic platform for producing monoclonal antibodies to relevant blood group antigens. The scFv-phage produced in this way warrants further development for use as a reagent for Dia red cell typing.

Two novel antithetical KN blood group antigens may contribute to more than a quarter of all KN antisera in Europe.

BACKGROUND: All antigens described in the KN blood group system are located in the long homologous repeat D (LHR-D) of complement receptor 1 (CR1). While there have been reports that some sera react only with the long homologous repeat C (LHR-C), the antigens in LHR-C are unknown. STUDY DESIGN AND METHODS: Recombinant LHR-C and LHR-D were used to identify antibodies directed against LHR-C of CR1, into which a point mutation was introduced to characterize the underlying blood group antigens. In addition, database studies to define haplotypes of CR1 were performed. RESULTS: Several antisera were identified that were specific against CR1 p.1208His and against CR1 p.1208Arg, located in LHR-C. Fifteen KN haplotypes were found in the Ensembl genome browser. It was shown that due to a linkage disequilibrium anti-CR1 p.1208His may be mistaken for anti-KCAM. CONCLUSION: A novel antithetical KN blood group antigen pair was found at position p.1208 of CR1, for which the names DACY and YCAD are proposed. Antibodies against these two novel antigens seem to contribute to more than a quarter of all KN sera in Europe.

Disruption of the tumour-associated EMP3 enhances erythroid proliferation and causes the MAM-negative phenotype.

The clinically important MAM blood group antigen is present on haematopoietic cells of all humans except rare MAM-negative individuals. Its molecular basis is unknown. By whole-exome sequencing we identify EMP3, encoding epithelial membrane protein 3 (EMP3), as a candidate gene, then demonstrate inactivating mutations in ten known MAM-negative individuals. We show that EMP3, a purported tumour suppressor in various solid tumours, is expressed in erythroid cells. Disruption of EMP3 by CRISPR/Cas9 gene editing in an immortalised human erythroid cell line (BEL-A2) abolishes MAM expression. We find EMP3 to associate with, and stabilise, CD44 in the plasma membrane. Furthermore, cultured erythroid progenitor cells from MAM-negative individuals show markedly increased proliferation and higher reticulocyte yields, suggesting an important regulatory role for EMP3 in erythropoiesis and control of cell production. Our data establish MAM as a new blood group system and demonstrate an interaction of EMP3 with the cell surface signalling molecule CD44.

Biotechnologically produced fucosylated oligosaccharides inhibit the binding of human noroviruses to their natural receptors.

Norovirus infections cause severe gastroenteritis in millions of people every year. Infection requires the recognition of histo-blood group antigens (HBGAs), but such interactions can be inhibited by human milk oligosaccharides (HMOs), which act as structurally-similar decoys. HMO supplements could help to prevent norovirus infections, but the industrial production of complex HMOs is challenging. Here we describe a large-scale fermentation process that yields several kilograms of lacto-N-fucopentaose I (LNFP I). The product was synthesized in Escherichia coli BL21(DE3) cells expressing a recombinant N-acetylglucosaminyltransferase, ß(1,3)galactosyltransferase and α(1,2)fucosyltransferase. Subsequent in vitro enzymatic conversion produced HBGA types A1 and B1 for norovirus inhibition assays. These carbohydrates inhibited the binding of GII.17 virus-like particles (VLPs) to type A1 and B1 trisaccharides more efficiently than simpler fucosylated HMOs, which were in turn more effective than any non-fucosylated structures. However, we found that the simpler fucosylated HMOs were more effective than complex molecules such as LNFP I when inhibiting the binding of GII.17 and GII.4 VLPs to human gastric mucins and mucins from human amniotic fluid. Our results show that complex fucosylated HMOs can be produced by large-scale fermentation and that a combination of simple and complex fucosylated structures is more likely to prevent norovirus infections.

Blood Group Antigen Shielding Facilitated by Selective Cell Surface Engineering.

Production of red blood cells (RBCs) without immunogenicity of blood group antigens is of special interest in blood transfusion therapy in clinical chemistry. In this study, a selective cell surface engineering method was developed for the preparation of antigen-shielded RBCs based on molecular imprinting. Using an epitope imprinting method, biocompatible molecularly imprinted nanogels (MIgels) were prepared with a high affinity to the blood group antigens of RBCs. The antigen-shielded RBCs could avoid the agglutination caused by blood group mismatch, resulting in the antigen-shielded RBCs in efficiently substituting RBCs in case of a shortage of blood supply. Moreover, the antigen-shielded RBCs could maintain the normal physiological structure and functions of the original RBCs. We believe that the selective cell surface engineering presented in this work may offer significant benefits in specific cell protection for biomedical application.

Snow Mountain Virus recovery by synthetic human histo-blood group antigens is heavily influenced by matrix effects.

Noroviruses are known to bind to histo-blood group antigens (HBGAs) and the specific binding patterns depend on the virus genotype. However, the development of point-of-care diagnostic assays based on this binding has been challenging due to low assay sensitivity. This study utilized a well-defined stool collection from a GII.2 Snow Mountain Virus (SMV) human challenge study to investigate virus recovery from stool and emesis samples using HBGA-coated beads. SMV was recovered from H type III-coated beads for 13 stool specimens out of 27 SMV-positive specimens tested. After adjusting for non-specific binding to PEG-coated beads, the mean percent recovery by H type III-coated beads was 308.11% +/- 861.61. Recovery by H type III ligands was subject-specific and weakly correlated with stool consistency. Input virus titer was not correlated with SMV recovery. The results suggest that the generally low virus recovery we observed may be due to bead saturation or hindrance by existing glycans in the matrix that precluded the virus from being captured by the synthetic glycans. These results indicate a strong role for subject-specific and matrix effects in HBGA binding by SMV. Further investigation of the nature of this interference is needed to facilitate development of high sensitivity diagnostic assays.

BC2L-C N-Terminal Lectin Domain Complexed with Histo Blood Group Oligosaccharides Provides New Structural Information.

Lectins mediate adhesion of pathogens to host tissues, filling in a key role in the first steps of infection. Belonging to the opportunistic pathogen Burkholderia cenocepacia, BC2L-C is a superlectin with dual carbohydrate specificity, believed to mediate cross-linking between bacteria and host cells. Its C-terminal domain binds to bacterial mannosides while its N-terminal domain (BCL2-CN) recognizes fucosylated human epitopes. BC2L-CN presents a tumor necrosis factor alpha (TNF-) fold previously unseen in lectins with a novel fucose binding mode. We report, here, the production of a novel recombinant form of BC2L-CN (rBC2L-CN2), which allowed better protein stability and unprecedented co-crystallization with oligosaccharides. Isothermal calorimetry measurements showed no detrimental effect on ligand binding and data were obtained on the binding of Globo H hexasaccharide and l-galactose. Crystal structures of rBC2L-CN2 were solved in complex with two blood group antigens: H-type 1 and H-type 3 (Globo H) by X-ray crystallography. They provide new structural information on the binding site, of importance for the structural-based design of glycodrugs as new antimicrobials with antiadhesive properties.

Characterization of a Histo-Blood Group Antigen-like Substance in Romaine Lettuce That Contributes to Human Norovirus Attachment.

Human noroviruses (HuNoVs) are among the main pathogens causing acute nonbacterial gastroenteritis. Histo-blood group antigens (HBGAs) are widely accepted receptors for HuNoV specific binding. HBGA-like substances in produce are also considered as the critical ligands for capture of HuNoVs. However, the composition of viral ligands from food substrates remains unknown. In this study, an oligosaccharide (H2N2F2) was captured and isolated from romaine lettuce extract by a bacterial surface display system. Using electrospray ionization mass spectrometry and tandem mass spectrometry, it was shown that H2N2F2 was most likely to be a chimera of type A, H, and Lewis a HBGAs. The composition was consistent with our ELISA results using a panel of monoclonal antibodies against HBGAs. Our results revealed a possible interaction mechanism between HuNoVs and romaine lettuce. Better understanding of the interaction of HuNoVs with easily contaminated produce will ultimately aid in the control of and reduction in disease outbreaks.

HBGA binding modes and selectivity in noroviruses upon mutation: a docking and molecular dynamics study.

Norovirus, also called winter vomiting bug, is the most common cause for gastroenteritis and severe childhood diarrhea disease. High mutation rates cause drug resistance and thus complicate the development of an effective therapeutics against virus infection. The virus protein enters the host cell via the interaction with histo-blood group antigens (HBGAs), formed by oligosaccharides. To date, the crystal structures of numerous complexes of virus proteins with different antigens have been reported. The HBGAs bind to the two distinct regions of the virus proteins. Herein, the affinity of different variants of virus protein to some common glycans has been computationally analyzed. Molecular docking studies as combination of docking scores and rmsd values revealed that the binding region 1 is more attractive for the ligands in variants of categories 1-5, but selectivity is drastically shifted to region 2 due to in category 6. In addition, molecular dynamics simulations were unraveled when the region 1 is hindered (in category 6); the blocking loop has less fluctuation than that of unblocked in other categories.