This is the first study quantifying the novel observations made in our laboratory by Habte et al.  that crude saliva, from uninfected individuals and its mucins MUC5B and MUC7 inhibit HIV-1 in an in vitro assay. Crude saliva from HIV positive patients was not studied by Habte et al.[2, 3, 21]. The method we used was strictly according to that of Habte et al. , who did not test for mucin in the insoluble debris. This debris was shown to have some mucin in saliva preparations, which was not added to that in the supernatant in this study. Any HIV particles in the pellet would be rendered denatured by the 6 M guanidinium hydrochloride in the extraction media. Besides confirming Habte’s finding on normal saliva and salivary mucin , we have gone on to show that crude saliva and its purified mucins from HIV infected individuals also inhibited HIV-1, in contrast to the findings of Habte et al.[2, 3, 21], who reported that HIV positive mucins did not inhibit the virus.
We consider this study to be quantitative in the sense that it differs from previous studies of Habte et al.  in that there were two major groups in this study, a diagnostically confirmed normal (n = 20) and infected group (n = 20), and there was no pooling of sample in either group. We do not claim to quantify mucins strictly by assay because of the difficulties such a measurement poses, due to the polydisperse or even heterogenous nature of crude mucus secretions.
This quantitative study had several advantages over the previous qualitative study of Habte et al. , in our laboratory. Firstly, the HIV negative group of twenty individuals were tested for HIV, unlike previously where samples were taken on the basis of trust from patients declaring a risk-free lifestyle with respect to sexual habits and other risky behaviours such as the abuse of drugs. Furthermore this study compared the inhibitory potential of crude saliva in both groups together with that of the mucins MUC5B and MUC7. Since Habte et al.  did not test crude saliva from HIV positive individuals and focussed only the salivary mucins MUC5B and MUC7 purified from a pool of HIV positive saliva , our findings raise new questions, encouraging us to design future studies that will take into account the treatment status of patients, the extent of the infection in relation to the CD4 counts and a wider genetic study of mucins in a population in this region. It is interesting that a few samples in both groups did not inhibit the virus. This may be due to inter-individual variation in mucin concentration of the crude saliva samples. The establishment of a dose–response curve, which we are busy with, will help answer such questions. It had not been done because of limitations in terms of time constraints, sample yield (we used individual samples) and location (these experiments were carried out in the Division of Virology at the University of Stellenbosch in which there was this established assay made available to us and in which laboratory space was shared due to restricted conditions).
Crude saliva is conveniently separated into its mucin components with MUC5B eluting in the V0 and MUC7 in the Vi of a Sepharose CL-4B gel filtration column. However, there was inter-individual variation within normal and infected groups, especially for MUC5B which in some instances seemed hardly detectable, whilst in other cases eluted as a small but broad peak from the column. When comparing the biochemical properties of mucins from normal saliva and that from HIV-infected patients, both groups had more MUC7 material than MUC5B. The reason for this is unknown but it is not an exact measure in the strict sense of a mucin assay (with reference to a standard curve), but rather a more general assessment based on the size of peaks eluting from the column, an indicator of amounts of material on a broad comparative basis, like has been previously reported in other studies on mucins in the stomach [25, 26]. All samples showed a larger MUC7 peak, the shape of which differed from sample to sample and was associated with considerable amounts of protein compared to the MUC5B of the void volume. In some instances this peak was split, suggesting a variation in the MUC7 population, in keeping with Habte’s findings of two bands of purified MUC7 on Western blotting after 4-20% gradient gel electrophoresis .
Crude saliva and purified mucins from both groups inhibited the virus in an in vitro assay. Mucins were purified by density gradient ultra-centrifugation in CsCl, a long-established procedure shown to free mucins in complex secretions from contaminating protein [27–29]. Salivary mucins MUC5B and MUC7, purified from the saliva of HIV positive patients, were previously shown by 4-20% gradient SDS-PAGE to be pure and Western Blotting confirmed the identity of these mucins . The amino acid analysis is characteristic of purified O-glycosylated mucin in its serine, threonine and proline (S, T and P) content , which was slightly higher in the HIV positive group. The significance of this is not known and since only 3 samples from each group were analysed, a statistical analysis was not done.
It may be worthwhile to determine further whether the roles of MUC5B and MUC7 vary in the inhibition of HIV-1. A study by Thomsson et al  highlights the differences in glycosylation between MG1 (MUC5B) and MG2 (MUC7). They identified that MG1-derived oligosaccharides (sugar side chains attached to the protein molecule) were significantly longer than those of MG2. This is in keeping with the findings in this study that MUC5B was a larger molecule than MUC7 and eluted in the void volume (V0) of the Sepharose Cl-4B column, although the length and size of the polypeptide may not necessarily relate to the length of oligosaccharides. A greater extent of glycosylation may be indicative of a more relevant role in the inhibition of HIV-1 transmission. Interestingly it is a feature of cancer-related mucins to be under-glycosylated and to have an altered glycosylation pattern when compared with normal mucins [11, 31]. A more glycosylated molecule with longer oligosaccharides could more effectively aggregate virus particles. Some of these questions would form the basis for future research.
Western blotting detection of purified salivary mucin samples revealed a large amount of MUC5B in HIV negative samples compared with trace amounts of MUC5B in HIV positive samples. MUC7 was detected in both HIV negative and HIV positive samples (Figure 3b). Here however the charge of the oligosaccharides on the apomucin can influence migration of the mucin through an agarose gel.
Purified salivary mucins MUC5B and MUC7 with dilutions of 10-1, 10-3, 10-5, 10-10, 10-20 and 10-40, provided a wider range than that used by Habte et al. , and showed the potency of the mucin in the inhibition of the virus at very low concentrations and high dilutions. The finding of Habte et al. that MUC5B and MUC7 from the saliva of patients who are HIV positive did not inhibit mucins is an essential difference to our finding in this study. If any variation existed in the inhibitory potential of mucins from sample-to-sample, this would be masked by the pooling of these individual samples. There is a large amount of other protein contaminants in crude saliva and their inhibitory potential needs to be investigated. These include cystatins (inhibitors of cysteine proteases), antibodies (sIgA) and, in addition to mucins, other high molecular weight glycoproteins such as salivary agglutinin (SAG) [9, 32]. SAG has been shown to have a specific inhibitory effect through interaction with viral capsid glycoprotein gp120 . A larger study is being planned to investigate these questions, especially a comparison of the inhibitory properties of mucins versus other components such as gp340 in the inhibition of the virus.
DNA analysis of tandem repeat regions in the genes of MUC5B and MUC7 from HIV negative and HIV positive donors revealed no association of HIV-infection status and gene polymorphisms. Polymorphisms are distributed between both groups. There is no apparent link between heterozygosity or homozygosity in either the MUC5B or MUC7 tandem repeat alleles and HIV-infection. This suggests that there is no genetic predisposition for the susceptibility to HIV-infection. An investigation of the variation in the genetic and protein structure of mucins MUC5B and MUC7 and their association with infection with HIV is a study that could be engendered by these findings.
One area of research aimed at the reduction of infection in this region is the development of microbicides which can enhance the body’s first line of defence against the virus. A formulation or preparation, in which mucins could form a part of a barrier substance, may be an idea worth exploring. Secreted mucins such as MUC5B form gels in the respiratory and other internal tracts of the body which protect epithelia lining the mucosae of these tracts from harmful factors in a hostile milieu. One such example is the crude mucus gel lining the surface of the gastric mucosa, a 200 μM thick barrier which forms an unstirred layer on the gastric mucosal surface and protests it from hydrochloric acid (down to pH 1), pepsin activity and the shear forces associated with digestion . Our findings suggest that the virus could be trapped by mucins in the saliva. We suggest that the biochemical structure of these mucins, together with the physical properties of the crude mucus, could make a mucus-based formulation (a substance or preparation that contains mucins or even crude mucus as a component of the preparation) an effective barrier, protecting the vaginal mucosa against the shear associated with sexual intercourse and remaining intact long enough to resist the infection of the mucosal cells by the HI virus. Studies amenable to identifying a peptide or short polypeptide with activity that could be developed into a microbicide are also in progress.