MUC5B is the predominant polymeric mucin in human saliva [Thornton, Khan, Mehrotra, Howard, Veerman, Packer and Sheehan (1999) Glycobiology 9, 293–302], where it contributes to oral cavity hydration and protection. More recently, the gene for another putative polymeric mucin, MUC19, has been shown to be expressed in human salivary glands [Chen, Zhao, Kalaslavadi, Hamati, Nehrke, Le, Ann and Wu (2004) Am. J. Respir. Cell Mol. Biol. 30, 155–165]. However, to date, the MUC19 mucin has not been isolated from human saliva. Our aim was therefore to purify and characterize the MUC19 glycoprotein from human saliva. Saliva was solubilized in 4 M guanidinium chloride and the high-density mucins were purified by density-gradient centrifugation. The presence of MUC19 was investigated using tandem MS of tryptic peptides derived from this mucin preparation. Using this approach, we found multiple MUC5B-derived tryptic peptides, but were unable to detect any putative MUC19 peptides. These results suggest that MUC19 is not a major component in human saliva. In contrast, using the same experimental approach, we identified Muc19 and Muc5b glycoproteins in horse saliva. Moreover, we also identified Muc19 from pig, cow and rat saliva; the saliva of cow and rat also contained Muc5b; however, due to the lack of pig Muc5b genomic sequence data, we were unable to identify Muc5b in pig saliva. Our results suggest that unlike human saliva, which contains MUC5B, cow, horse and rat saliva are a heterogeneous mixture of Muc5b and Muc19. The functional consequence of these species differences remains to be elucidated.
- tandem MS
Saliva is a dilute, complex mixture of proteins, glycoproteins, lipids and ions secreted from the major and minor salivary glands [1–3]. Saliva has many diverse functions including aiding mastication and speech, digestion of food and maintenance of oral health . The major macromolecular components of normal, unstimulated saliva are the mucins. There are two populations of salivary mucins: the high-molecular-mass polymeric, gel-forming mucin MUC5B and the lower-molecular-mass, non-polymeric mucin MUC7 [5–7]. Two distinctive structural characteristics of these glycoproteins have important functional consequences . First, the MUC5B and MUC7 mucin polypeptides have a large central domain, with tandemly repeated sequences enriched in serine and/or threonine residues [the TR (tandem repeat) or mucin domain], which are the sites of extensive modification with O-glycans. Secondly, upstream and downstream of the MUC5B mucin domain are regions of the polypeptide similar to von Willebrand factor D domains, which are important for mucin polymer formation . Thus, as a result of their extreme size and abundance of negatively charged O-glycans, these glycoproteins play key roles in hydration and lubrication of the oral surfaces. Furthermore, mucins bind to and sequester bacteria via their glycans and protein domains [9,10].
Until 2004, four gel-forming mucins had been reported, namely MUC6, MUC2, MUC5AC and MUC5B, which are encoded by four consecutive genes on chromosome 11p15.5 . These mucins are expressed in a tissue- and cell-specific manner. For instance, in humans, MUC6 is expressed in mucous cells of submucosal glands in the stomach, MUC5B is mainly expressed by the mucous cells of the salivary glands and the submucosal glands in the airways, MUC5AC is expressed by the goblet cells of the airways and the stomach, and MUC2 is mainly expressed by the goblet cells in the intestine [12–17]. This tissue-specific distribution has been found to be relatively well conserved between species [18,19]. For example, in the mouse, Muc5b, along with Muc5ac, has been found to be expressed in the airways, Muc2 is mainly expressed in the intestine and Muc5ac and Muc6 are both found in the stomach [20–23]. However, differences in expression between mouse and human have also been reported. For example, in contrast with humans, no evidence of Muc5b (nor Muc2, Muc5ac or Muc6) expression has been reported in mouse saliva [22,23].
In 2004, MUC19 was reported as a fifth human gel-forming mucin. Although neither the complete amino acid nor mRNA sequence has been published, MUC19 is predicted to have similar structural features to the other human gel-forming mucins . MUC19 gene expression has been reported in human airways and, in particular, in human salivary glands . In the mouse, the complete sequence of the Muc19 gene has been described and it is strongly expressed in sublingual and submandibular salivary glands [24,25]. Furthermore, mRNA sequences from porcine and bovine submaxillary mucins show high sequence similarity to the mouse Muc19 sequence [26–28]. On the basis of these results, MUC19/Muc19 would be expected to be present in saliva from humans and other mammals. However, studies on human saliva have identified MUC5B as the predominant polymeric mucin [7,16], and the MUC19 glycoprotein is yet to be identified in human saliva, or for that matter, in saliva from other mammals. Moreover, in other animals, it is not known whether Muc5b is a major salivary mucin. Therefore, in the present study, our first aim was to purify the polymeric mucins from human saliva in order to identify whether MUC19 was present. Subsequently, we analysed the saliva collected from horse, rat, pig and cow to identify whether Muc5b and/or Muc19 were present.
Unstimulated saliva was collected from six healthy human volunteers (between 20 and 35 years of age) by spitting into 50 ml tubes that were kept on ice during the collection. Horse, cow and pig saliva were collected from excess secretions drooled while feeding. Necessary ethical approval was obtained for all the above animal samplings. Rats were anaesthetized with ketamine hydrochloride (100 mg/kg) and xylazine (10 mg/kg) by intraperitoneal injection and were injected intraperitoneally with the cholinergic agonist pilocarpine (25 mg/kg) to stimulate salivary secretion. The anaesthetized rats were laid on their side and saliva dripping from the mouth was collected into plastic Petri dishes. Experiments were carried out in accordance with the U.K. Home Office guidelines for animal welfare based on the Animals (Scientific Procedures) Act 1986. Equal volumes of 8 M GdmCl (guanidinium chloride) were added to each sample at the time of collection and up to 5 vol. of 4 M GdmCl were subsequently added to solubilize the mucins. This step was performed by gentle agitation for at least 24 h at 4 °C.
Purification and identification of human and horse salivary mucins
Human and equine salivary mucins, solubilized as above, were purified by CsCl (caesium chloride)/4 M GdmCl density-gradient centrifugation at a starting density of 1.4 g/ml. The samples were centrifuged for at least 68 h at 40000 rev./min and 15 °C in a Beckman Ti45 rotor. After centrifugation, the tubes were emptied from the top into 20 fractions. An aliquot of each fraction was slot-blotted on to a nitrocellulose membrane and stained with PAS (periodate–Schiff) reagent according to our method published previously . The PAS-rich fractions were pooled, reduced and carboxymethylated, dialysed into water and then freeze-dried.
The mucins present in the sample were identified by using MS/MS (tandem MS) as described previously . In brief, samples were resuspended into 0.1 M ammonium bicarbonate prior to digestion with trypsin. The low-molecular-mass tryptic peptides were isolated by Sephacryl S-100 gel-filtration chromatography and then separated by reverse-phase chromatography and analysed inline by positive-ion electrospray ionization-MS/MS using a Q-ToF (quadropole-time-of-flight) micromass spectrometer (Waters, Manchester, U.K.). MS/MS data were analysed using a custom database containing putative mucin sequences (see below).
Mucin sequence prediction
Published Muc19 and MUC5B sequences (GenBank® accession numbers NM_207243, AF005273 and AJ004862 [25,27,30]) obtained from mRNA sequencing data were used to perform a BLAST search of the genome sequence database of human, horse, bovine, pig and rat genomes. The Muc5b and Muc19 genes in each genome were determined based on gene synteny and sequence similarity. In the human (chromosome 12), horse (chromosome number not identified) and rat (chromosome 7) genomes, the LRRK2 gene (leucine rich repeat kinase 2 gene) is located upstream of Muc19, whereas the CNTN1 (Contactin 1) gene is located downstream of Muc19. In the current version of the cow genome, the Muc19 gene is located downstream of the LRRK2 gene on chromosome 5, but the CNTN1 gene is not in the same locus. However, the genome contains sequence gaps and the assembly might not be complete. In the human genome, MUC5B is part of a MUC gene complex located between the APA2 gene (adaptor-related protein complex 2, α2-subunit gene) and the Tollip gene (Toll-interacting protein gene). This was also the case in the horse genome, whereas in the rat and cow genomes, it is not yet clear whether synteny is conserved because of the gaps in the genomic sequence and the incomplete assembly of these genomes. The structure of the genes was predicted based on the conservation of the exon/intron structure and amino acid sequence, which has been shown to be very tightly conserved between polymeric mucins and across species [22,31]. The genome databases were accessed through the UCSC (University of California Santa Cruz) browser (http://genome.cse.ucsc.edu/) and the Ensembl browser (http://www.ensembl.org/index.html). It should be noted that the predicted N-terminal sequences do not represent the entire N-terminal sequence, but start from exons 4 and 2 for Muc19 and Muc5b respectively. The predicted exonic sequences were translated and the resultant amino acid sequences of putative Muc5b and Muc19 mucins were added to an in-house custom mucin database.
Tandem mass spectra were extracted, charge-state deconvoluted and de-isotoped by Masslynx version 4.0 (Waters). All MS/MS samples were analysed using Mascot (Matrix Science, London, U.K.; version 2.2.03) and X! Tandem (http://www.thegpm.org; version 2007.01.01.1). X! Tandem was set up to search a subset of our in-house custom mucin database. Mascot was set up to search the MSDB_2006-Sep-08 database (selected for Mammalian) and our in-house custom mucin database. Mascot and X! Tandem were searched with a fragmentation-mass tolerance of 0.4 Da and a parent-ion tolerance of 1.3 Da. The iodoacetamide derivative of cysteine was specified in Mascot and X! Tandem as a fixed modification. Oxidation of methionine was specified in Mascot and X! Tandem as a variable modification.
Criteria for protein identification
Scaffold (version Scaffold-01_07_00, Proteome Software Inc., Portland, OR, U.S.A.) was used to validate MS/MS-based peptide and protein identifications. Peptide identifications were accepted if they could be established at greater than 90% probability, as specified by the PeptideProphet algorithm . Protein identifications were accepted if they could be established at greater than 90% probability and contained at least one identified peptide. Protein probabilities were assigned by the ProteinProphet algorithm . Proteins that contained similar peptides and could not be differentiated based on MS/MS analysis alone were grouped to satisfy the principles of parsimony.
Identification of mucins in the saliva of cow, rat and pig
To determine whether Muc19 and Muc5b were present in cow, rat and pig saliva, we employed a less lengthy procedure to isolate the mucins. This was because we only wished to determine whether the mucin was present as opposed to obtaining maximum coverage of the polypeptide. For this approach, our aim was to enrich rather than purify the mucins. In brief, saliva from cow, rat and pig was dissolved with 8 M GdmCl and then dialysed into 6 M urea. An aliquot of this non-purified saliva was reduced and subjected to SDS/PAGE using NuPAGE 4–12% bisacrylamide gels (Invitrogen, Paisley, Renfrewshire, Scotland, U.K.). Gels were run for 2 h at 150 V in NuPAGE Mes/SDS running buffer and NuPAGE LDS sample buffer (Invitrogen). After electrophoresis, the gels were stained with PAS  and the high-molecular-mass smears at the top of the gel were excised, in-gel-digested with trypsin  and analysed by MS/MS as described above.
To determine whether Muc19 is a polymeric mucin stabilized by disulfide bonds, untreated mucins and reduced and carboxymethylated mucins from horse and rat were layered on to 6–8 M GdmCl gradients and centrifuged at 40000 rev./min in an SW40 Ti Beckman rotor for 2.5 h at 15 °C as described previously . The tubes were emptied into 24 fractions from the top and the mucin distribution was analysed by PAS staining.
MUC5B and MUC19 mRNA relative expression in human salivary glands
Normal human salivary gland total RNA (Clontech 636552, 24 male/female Caucasians, age 16–60) was reverse-transcribed into cDNA using random primers. The SYBR Green system (Eurogentec) was then used to amplify MUC5B and MUC19 cDNA by real-time quantitative PCR using the protocol supplied by the manufacturer. The oligonucleotide primers used to amplify MUC5B were MUC5B-forward primer 5′-ACCCAACGGTGCAATGTCA-3′ and MUC5B-reverse primer 5′-CGGCCACTCTCTTGTACTCAAAG-3′, and the oligonucleotide primers used to amplify MUC19 were MUC19-forward primer 5′-GAGTTCAGATGGCAAAATGCACA-3′ and MUC19-reverse primer 5′-TGCCATCAGGACAGTCAAGTACA-3′ . The efficiency of each pair of primers was determined by using five serial dilutions of the cDNA and was ≥99.8%. The identity of the amplified products was confirmed by sequencing.
The first aim of the study was to determine whether MUC19 was present in human saliva. Our approach, which we have employed previously to identify human and horse polymeric mucins [19,35], was to purify salivary mucins, and then use MS/MS of tryptic peptides to identify MUC19. However, only a partial MUC19 sequence, corresponding to the C-terminus of the mucin, has been reported . Therefore to maximize the chance of identifying MUC19, most of the polypeptide sequence of the N-terminus of MUC19 was predicted (for details see the ‘Experimental’ section). This resulted in a 1178-amino-acid sequence, which was added to an in-house mucin protein sequence database for subsequent analysis of MS/MS data.
Identification of polymeric mucins in human saliva
Human saliva, pooled from six individuals, was solubilized in 4 M GdmCl and fractionated by density-gradient centrifugation. PAS staining of fractions across the density distribution showed that the high-density fractions contained the bulk of the mucins (Figure 1). The major PAS-rich peak was pooled (fractions 11–18), reduced and carboxymethylated, dialysed against water and freeze-dried in order to concentrate the sample. The freeze-dried material was digested with trypsin and the low-molecular-mass tryptic peptides were analysed by MS/MS . This revealed 84 unique peptides that matched MUC5B (Figure 2), but no peptides that matched MUC19 (Figure 3). Since MUC19 was not found in the high-density, PAS-rich fractions, the lower density material (fractions 1–10) from the density gradient was pooled and analysed in the same manner as above. The MS/MS analysis showed the presence of five peptides that matched MUC5B, but no peptides that matched MUC19.
mRNA expression in human salivary glands
The apparent absence of MUC19 in human saliva extracts is at odds with published mRNA expression data . Therefore to confirm previous expression data and to determine the relative expression of MUC5B and MUC19 mRNA, we performed quantitative real-time PCR on pooled human salivary gland cDNA. This analysis showed that MUC19 mRNA was expressed in human salivary glands and that there was approx. 10-fold more MUC5B mRNA than MUC19 mRNA.
The presence of MUC19 mRNA in salivary glands suggests that our MS-based approach might not have been able to detect this mucin. Therefore to investigate whether our approach could identify this mucin, we collected saliva from other mammals (cow, horse, rat and pig), isolated the mucins and tested for Muc19 and the homologue of the major human salivary mucin, Muc5b, using MS/MS of tryptic peptides. However, for this to be successful, we first had to predict the polypeptide sequences for horse, cow, pig and rat Muc19 and Muc5b.
Predicted N-terminal sequences of Muc5b and Muc19 from cow, horse, rat and pig
Mucin polypeptide sequences for cow, horse, rat and pig were predicted as described in the ‘Experimental’ section. The Muc19 and Muc5b N-terminal deduced amino acid sequence alignments are shown in Figures 3 and 4 respectively. The Muc19 sequences show approx. 68% identity and that 99% of the cysteine residues (100/101) are conserved. The alignment of the Muc5b sequences shows that they share 67% identity and that all of the 90 cysteine residues are conserved.
Identification of Muc19 and Muc5b mucins from horse saliva
The purification and identification strategy used for analysis of human salivary mucins was employed for the analysis of horse saliva. The CsCl/4 M GdmCl density-gradient fractionation of horse saliva is shown in Figure 5. The high-density mucins (fractions 10–16) were reduced and carboxymethylated, digested with trypsin and subjected to MS/MS analysis. The results showed that thirteen peptides matched the predicted N-terminal polypeptide of Muc5b and 25 peptides matched the predicted N-terminal polypeptide sequence of Muc19 (Table 1, Figures 3 and 4). This validated our approach and showed the presence of Muc5b and Muc19 mucins in horse saliva. We then investigated whether one or both of these two mucins were present in the saliva of the cow, pig and rat.
Identification of Muc19 and Muc5b mucins from cow, pig and rat saliva
To identify more rapidly the mucins present in the saliva of these three animals, we exploited the low electrophoretic migration of polymeric mucins in SDS/PAGE . Solubilized saliva collected from rat, cow and pig was subjected to SDS/PAGE on a 4–12% gradient gel. After electrophoresis, gels were stained with the PAS reagent and the carbohydrate-rich smears located at the top of the gels (results not shown) were excised and digested with trypsin. MS/MS analysis showed the presence of both Muc5b and Muc19 in rat and cow saliva (Table 1, Figures 3 and 4). Muc19 was also present in pig saliva with eight peptide matches (Table 1, Figure 3). However, the presence of Muc5b in pig saliva could not be tested, because no Muc5b sequence is currently available in the pig genome.
Is Muc19 a polymeric mucin?
Muc19 is reported to be a polymeric mucin based on the presence of the von Willebrand factor-like D domains, which are necessary for the disulfide-bond-mediated oligomerization of gel-forming mucins . In order to test this, we examined whether there was a change in size distribution of horse and rat salivary mucins (both mixtures of Muc5b and Muc19) after treatment with a reducing agent (10 mM dithiothreitol), by rate-zonal centrifugation (Figure 6). The unreduced mucins were characterized by a broad range of sedimentation rates, characteristic of a polydisperse distribution of mucins [37,38], which, after reduction, exhibited a lower sedimentation rate. These results provide evidence that Muc19 (and Muc5b) is a polymeric mucin stabilized by disulfide bonds.
The finding that the MUC19 gene is expressed in human salivary glands  raised the issue that a major component of the protective barrier in the mouth has been overlooked in previous studies on saliva. For instance, we have reported that MUC5B was the polymeric mucin responsible for the properties of saliva [7,35,39]. However, using standard methods for extraction and purification of polymeric mucins , we were unable in the present study to detect MUC19 either in the main mucin preparation after density-gradient purification or in any of the other fractions from the density gradient. Thus the data presented here confirm these earlier studies and suggest that MUC5B is the predominant polymeric mucin in adult human saliva (pooled from six healthy volunteers) and that MUC19 is apparently absent. There are a number of reasons why MUC19 might not have been detected: (i) it was not solubilized in 4 M GdmCl; (ii) peptides may not have been produced by trypsin digestion, thus the mucin would not have been identified using the MS-based approach; (iii) it is not present in saliva; or (iv) MUC19 is present, but at levels below the limit of detection of the MS analysis. It seems unlikely that the first two reasons explain the lack of MUC19 since, using the same methodology, Muc19 was found in horse saliva. In addition, using the MS-based identification strategy, Muc19 was also present in rat, pig and cow saliva. We therefore propose that MUC19 is either not present in adult human saliva or is present at comparatively low levels. The latter proposal would seem to be favoured by the mRNA expression data, which showed, in pooled human salivary gland mRNA, that MUC19 mRNA is only approx. 10% of the level of MUC5B mRNA. Thus, if the two mRNAs are translated with similar efficiency, this would suggest that MUC19 is a more minor component of saliva than MUC5B. However, on the basis of the large number of peptides (84 in total) that matched MUC5B from the MS data, we might reasonably have expected to detect peptides from MUC19 if it were present at 10% of the level of MUC5B in saliva. Therefore we cannot discount other possibilities to explain the apparent absence of MUC19 from human saliva. It is possible that the mRNA is not (or is only poorly) translated into protein, the mucin is only present during development or childhood, or the protein is made but not secreted under normal circumstances. Interestingly, MUC19 levels (mRNA and protein) have been shown to be altered in the conjunctival tissue from a patient with Sjögren's syndrome , a condition that also affects saliva production. However, these authors did not investigate changes in MUC19 production in salivary glands.
Compositional differences of saliva have been reported between different mammals, for example in amylase, histatin and proline-rich proteins [42,43]. Thus the difference in mucin composition between the different animals is not surprising. The functional significance of having or not having Muc19 in saliva is at present obscure. One might speculate that it has an impact on the physical protective properties of saliva, in terms of hydration and lubrication, and/or it influences the innate defence properties of saliva, i.e. pathogen binding and sequestration of toxins. It is possible that the absence or low level of MUC19 in human saliva is an evolutionary adaptation relating to diet or oral flora.
The polymeric properties of MUC5B/Muc5b are an established feature of this mucin [16,19,44–46]; however, the oligomerization potential of Muc19 was based on its sequence similarity to the other gel-forming mucins . Here, we have provided experimental evidence that Muc19 from horse and rat forms polymers held together by disulfide bonds. While this is the first study that specifically demonstrates that Muc19 is a polymeric glycoprotein, earlier studies on recombinant N- and C-terminal fragments of PSM (pig submaxillary mucin; now known to be Muc19) showed that they were able to assemble in a disulfide-bonddependent manner [47–49], further confirming the polymeric nature of Muc19/PSM.
Horse, rat and cow saliva are a mixture of two polymeric mucins: Muc5b and Muc19. Mucus gels from other tissues have also been shown to be mixtures of two mucins: in the airways MUC5AC/Muc5ac and MUC5B/Muc5b [19,50] and in the stomach MUC5AC and MUC6 . In all of these tissues, the specific role of each mucin is not well understood. However, potential roles for the glycan structures on the mucins in the stomach have been proposed in the defence of the epithelium against the bacterium Helicobacter pylori. Leb and sialyl Lex glycans on MUC5AC can act as receptors for H. pylori  and α1-4-linked GlcNAc termini of glycans on MUC6 act as an antibiotic . Maybe Muc5b and Muc19 have similar roles in the oral cavity. Interestingly, in the rat, we do have preliminary evidence that Muc19 has higher charge density than Muc5b (K. Rousseau and D.J. Thornton, unpublished work).
In summary, our results confirm the earlier findings of Chen et al.  that MUC19 mRNA is expressed in human salivary glands; however, we have demonstrated that the MUC19 glycoprotein is not a major component of human adult saliva. Furthermore, we have shown for the first time that horse, cow and rat saliva do contain Muc19 and these secretions are mixtures of Muc5b and Muc19. Future studies are required to investigate the functional relevance of differences in mucin composition of saliva in the physiology of the oral cavity.
We thank Emma Keevill and Julian Selley for their technical assistance with MS analyses. This work was supported by the Wellcome Trust, the Horserace Betting Levy Board and the BBSRC (Biotechnology and Biological Sciences Research Council).
Abbreviations: CNTN1, contactin 1; GdmCl, guanidinium chloride; LRRK2 gene, leucine rich repeat kinase 2 gene; MS/MS, tandem MS; PAS, periodate–Schiff; PSM, pig submaxillary mucin; TR, tandem repeat
- © The Authors Journal compilation © 2008 Biochemical Society