OUP user menu

The effect of ingestion of milk supplemented with salivaricin A-producing Streptococcus salivarius on the bacteriocin-like inhibitory activity of streptococcal populations on the tongue

Karen P. Dierksen, Chris J. Moore, Megan Inglis, Philip A. Wescombe, John R. Tagg
DOI: http://dx.doi.org/10.1111/j.1574-6941.2006.00228.x 584-591 First published online: 1 March 2007


The colonization efficacies of salivaricin A (SalA)-producing Streptococcus salivarius strains 20P3 and 5 were compared when given in milk to 219 children, using either 2-day or 9-day dosing regimens. Colonization levels overall were superior for strain 5, and the 9-day dosing schedule resulted in higher levels of both initial colonization and strain persistence. The indigenous streptococcal tongue populations of 20 (10.9%) of the 189 children in the 2-day trial showed markedly increased SalA-like inhibitory activity following use of the S. salivarius-supplemented milk. All 20 of these children were found to have had relatively small (<5% of total S. salivarius) indigenous tongue populations of SalA-producing S. salivarius, and the relative proportions and/or inhibitory activity of these SalA producers on the childrens' tongues increased following ingestion of the S. salivarius-supplemented milk. Because SalA is known to be strongly inhibitory to Streptococcus pyogenes, an important implication of this study is that the consumption of SalA-producing probiotic S. salivarius could potentially help to effect a sustained increase in SalA-mediated protection against S. pyogenes infection.

  • Streptococcus salivarius
  • bacteriocin
  • bacteriocin-like inhibitory substance
  • microbial interference
  • probiotic
  • colonization


Probiotic bacterial replacement therapy, the practice of ingesting commensal bacteria to confer health benefits, has long been an accepted method for delivering beneficial microorganisms to the gastrointestinal tract of humans (Sullivan & Nord, 2002), poultry (Patterson & Burkholder, 2003), and swine (Abe et al., 1995). In recent years, however, researchers have begun to extend the use of probiotics to effect positive outcomes at a wider range of body sites (Tagg & Dierksen, 2003). Our principal goal is the application of streptococcal probiotics to afford protection against infections of the oral cavity, especially streptococcal pharyngitis.

Probiotic bacteria can exert antipathogen effects through a variety of competitive exclusion mechanisms that include (i) blocking attachment of other bacteria, (ii) sequestration of essential nutrients and metal ions, and (iii) production of inhibitory compounds such as acids, bacteriocins, and bacteriocin-like inhibitory substances (BLIS). Our laboratory in the Department of Microbiology and Immunology, University of Otago, Dunedin, New Zealand utilizes a deferred antagonism method on blood agar to test streptococci for BLIS production against nine standard indicators (Tagg & Bannister, 1979). By use of this P (producer)-typing procedure it has been shown that the production of BLIS is both frequent and varied within the genus Streptococcus. Many of the streptococcal BLIS activities detected by P-typing have now been attributed at least in part to antibiotic peptides of the bacteriocin class, and several of these, including SA-FF22 (Jack et al., 1994), streptin (Wescombe & Tagg, 2003) and salivaricin A (SalA) (Ross et al., 1993) belong to the lantibiotic subclass.

Streptococcus salivarius is a numerically prominent member of the indigenous oral microbiota of humans, and is only rarely associated with infections in healthy individuals (Burton et al., 2006a). Our initial P-typing studies of S. salivarius defined six prototype BLIS-producing strains, referred to as 5, 6, 9, 36, MPS and 20P3 (Dempster & Tagg, 1982). Streptococcus salivarius of P-type 676 or the very similar P-type 677 (the prototypes of which are strains 5 and 20P3) was detected in saliva specimens from c. 10% of 180 tested subjects (Tagg et al., 1983). Later, Ross et al. (1993) showed that S. salivarius 20P3 produces SalA, and subsequently it was found (P. Wescombe, unpublished data) that this lantibiotic is also produced by S. salivarius 5. Tompkins & Tagg (1989) found that individuals harbouring large numbers of P-type 677 S. salivarius (comprising >90% of their total S. salivarius) had coexisting alpha-hemolytic streptococci on their tongues that were significantly less sensitive in vitro to P-type 677 BLIS producers. This was taken to indicate that sufficient SalA had been produced in the oral cavity to effect in situ selection of SalA-resistant populations of some commensal oral streptococci.

Biologically active, extracellularly excreted SalA, like a number of other antibiotics, is capable of upregulating its own production, but perhaps even more significantly it can also cross-induce production of SalA homologues encoded by some strains of other oral streptococcal species including S. pyogenes, S. agalactiae and S. dysgalactiae (Upton et al., 2001; Wescombe et al., 2006). An in vitro assay has been used to detect specific inducing activity attributable to SalA in saliva specimens from human subjects either colonized with or naturally harboring populations of SalA-producing S. salivarius (Wescombe et al., 2006). Inducing activity could only be detected when the subjects' saliva contained a minimum of c. 8 × 105 CFU mL−1 of the SalA-producing S. salivarius (Wescombe et al., 2006).

As part of a series of trials designed to establish an effective protocol for colonizing children with probiotic S. salivarius to provide them with a degree of protection against S. pyogenes infections (Tagg & Dierksen, 2003), we wanted to evaluate milk as a potential delivery medium for the bacterial cells. Milk freshly supplemented with freeze-dried BLIS-producing S. salivarius is referred to here as ‘BLIS-milk’. The original objective of these colonization trials was simply to compare the efficacies of oral implantation of two strains of SalA-producing S. salivarius using either 2-day or 9-day BLIS-milk dosing schedules. This was carried out and the results are reported here. Much more interesting, however, we discovered upon retrospective analysis of the colonization data that in a number of subjects there was an apparent large increase, following their consumption of BLIS-milk, in the total inhibitory activity of the streptococcal component of their tongue microbiotas. The extent of this enhancement of inhibitory activity was far greater than could be accounted for by the relatively small numbers of the newly introduced colonizing strain. The conclusion that we arrived at is that the drinking of BLIS-milk containing SalA-producing S. salivarius can potentially lead to the enhancement of the inhibitory activity of the oral microbiota by at least two distinct mechanisms, namely (1) the additional activity contributed by the newly implanted probiotic strain; and (2) specific amplification of the SalA-producing activity of pre-existing components of the host's indigenous microbiota.

Materials and methods


Two hundred and nineteen children, aged between 5 and 10 years old, were recruited from five Dunedin primary schools and provided with BLIS-milk according to the schedule in Table 1. Streptococcus salivarius strains 5 and 20P3, both known to produce SalA, were compared for their colonization efficacies using either 2-day (Trial 1) or 9-day (Trial 2) dosing regimens. Approval was obtained from the Otago Ethics Committee.

View this table:

Distribution of children by trial group and BLIS colonizing strain

Trial detailsNo. of subjectsFemaleMale
Trial 1: 2-day course of BLIS-milk
Strain 20P3814734
Strain 5814833
Strains 20P3/5*271215
Trial 2: 9-day course of BLIS-milk
Strain 20P31569
Strain 515510
  • * BLIS-milk consisted of equal CFUs of each strain.


Large batch cultures of S. salivarius were grown in milk by the Cell Production Unit of the Dairy Research Institute (Palmerston, New Zealand), freeze-dried, and supplied in vacuum-sealed pouches. Containers of BLIS-milk powder were prepared for daily use by the children in the trials. Each contained a blend of freeze-dried cells (3 g), skim milk powder (7 g), and Nestle Quik® chocolate powder (5 g). The 20P3/5 BLIS-milk contained a 1 : 1 mixture (total 3 g) of S. salivarius strains 20P3 and 5. Fresh BLIS-milk was prepared by mixing 50 mL of cold tap water with the dehydrated powder and then refrigerating the suspension. The children were asked to sip the BLIS-milk slowly at least four times during the day, especially after meals. The count of S. salivarius in each of the freshly reconstituted BLIS-milks (c. 3 × 107 CFU mL−1) was determined by spiral-plating serial 10-fold dilutions (in 0.85% NaCl) onto Mitis-Salivarius agar (MSA) (Becton, Dickinson and Company, NJ) +1 mg mL−1 streptomycin (MSA+S). All three types of BLIS-milk were found to contain SalA activity (titre 8 AU mL−1).

Streptococcus salivarius and BLIS indicator strains

Naturally occurring variants (resistant to 1 mg mL−1 streptomycin) of S. salivarius 20P3 and 5 were used as the colonizing strains (Dempster & Tagg, 1982). The use of streptomycin-resistant cells was to facilitate accurate assessment of the levels of colonization. The BLIS P-types (677 and 676 respectively) given by these two strains indicate that, of the nine standard indicators, only Indicator 3 (Streptococcus anginosus strain T-29) is resistant to strain 20P3, whereas both Indicator 3 and Indicator 9 (Streptococcus dysgalactiae ssp. equisimilis strain T-148) are resistant to strain 5. Both strain 20P3 and strain 5 are SalA+ (produce salivaricin A), and in the P-typing test their BLIS activities interfere with the growth of standard indicator strains Micrococcus luteus (Indicator 1); Streptococcus pyogenes FF22, M-type 52 (Indicator 2); 71–679, M-type 4 (Indicator 5); 71–698, M-type 28 (Indicator 7); W-1, M-type 87 (Indicator 8); Streptococcus uberis ATCC 27958 (Indicator 4); and Lactococcus lactis T-21, (Indicator 6).

Testing of streptococcal tongue populations for BLIS activity

An assessment was made of the predominant in vitro BLIS activity of the streptococcal populations on each of the subjects' tongues using samples obtained by firmly swabbing the tongue surface from the back to the front. These samples were used to inoculate MSA, streaking out to obtain large numbers of single colonies. In preparing MSA, 1 mL of 1% potassium tellurite was added per litre of medium just prior to pouring the plates as per the manufacturer's instructions. MSA is used for the selective culture of streptococcal and enterococcal species. All culture incubations in the present study were for 18 h at 37°C in a 5% CO2-enriched atmosphere.

A sterile cotton swab was used to sample the mixed (in our experience predominantly streptococcal) population growing in the primary (confluent growth) region of the MSA culture. The charged swab was then used to deliver a 1-cm-wide diametric inoculum across the surface of buffered blood agar [Columbia Agar Base (Becton, Dickinson and Company) containing 0.1% calcium carbonate (to reduce acid-mediated inhibition) and 5% (v/v) whole human blood], the test medium that we routinely use for streptococcal BLIS P-typing (Tagg & Bannister, 1979). After incubation of the streak culture, the macroscopic bacterial growth was removed from the agar and the surface sterilized by inverting the medium-containing portion of the plate for 30 min over a circular cloth pad (placed on a sheet of glass) saturated with chloroform. The agar surface was then aired for 15 min (to remove residual chloroform), and the set of nine standard indicator bacteria cross-streaked across the line of the original inoculum and the plate reincubated. The indicators were inoculated by swabbing from Todd-Hewitt Broth (Becton, Dickinson and Company) cultures. After incubation, the growth of the indicator cultures was examined. Definite inhibition of the indicator culture was recorded as +. For the purposes of P-typing, the inhibitory activity against the nine standard indicators was recorded in code form (the P-type) as previously described (Tagg & Bannister, 1979). The proportion of the indigenous S. salivarius population producing SalA-like inhibitory activity was estimated as described previously (Tagg et al., 1983; Tompkins & Tagg, 1989) by picking c. 100 individual S. salivarius colonies (recognized by their characteristic large and mucoid morphology on MSA) into buffered blood agar freshly seeded with a lawn inoculum of M. luteus (Indicator 1) and then incubating. Stab-cultures surrounded by inhibition zones of >5 mm diameter were presumptive producers of SalA. Confirmation was obtained by recovering representative S. salivarius producing different-sized inhibition zones and testing these for their inhibitory profile (P-type) and for the presence of salA using PCR (‘Detection of salA-positive streptococci by PCR’).

Specific enumeration of the colonizing strain and of indigenous streptococci exhibiting resistance to streptomycin

In order to estimate the proportions of the total MSA-cultivable populations from each of the subjects' tongues that were resistant to 1 mg mL−1 streptomycin (both pre- and posttreatment) the following protocol was followed. A cotton swab was rubbed (as above) across the surface of the tongue and then placed into a bottle containing 1 mL of 1% peptone (Becton, Dickinson and Company), to be transported on ice to the laboratory for processing within 2 h. The bottle was vortex-mixed for 2–3 s, and then 50 μL of a 10−2 dilution in 0.85% NaCl was spiral-plated (Spiral Systems Inc., Model D) onto MSA (for the total cultivable count of streptococci/enterococci). Undiluted samples of the 1% peptone suspension were spiral-plated onto MSA+S (streptomycin 1 mg mL−1). Because none of the precolonization samples of the childrens' tongue microbiotas grew any S. salivarius-like colonies on this medium, all such colonies grown from the postcolonization samples were presumed to be the colonizing strain. Colonization levels were recorded as the percentage of streptomycin-resistant colonies (on MSA+S) of the total S. salivarius population (enumerated from the 10−2 dilution on MSA). Samples were collected 1, 4, and 9 days postcolonization.

Detection of salA-positive streptococci by PCR

DNA was extracted from 100 μL of each peptone suspension using a previously described boiling lysis method (Beall et al., 1996). Three microliters of template was used in subsequent PCR reactions with primers salAF 5′-GAT ATT TTG AAC AAT GCT ATC GAA G-3′ and salAR 5′-ACT AAT AGA AGT ATC TAG TAT GCT G-3′, as described previously (Upton et al., 2001). 16S primers were included in the mix to rule out PCR failure in the case of salA-negative reactions (Wilson et al., 1990).

Results and discussion

Trial 1: Dosing for 2 days with BLIS-milk containing SalA-producing strains 5 and/or 20P3

The oral cavity retention rates of strains 5 and/or 20P3 were compared in children given BLIS-milk for 2 days. More of the children receiving strain 5 still had detectable levels of that strain 1, 4 and 9 days postcolonization than did children given strain 20P3 (Table 2). Subjects dosed with a 1 : 1 mixture of strains 5 and 20P3 had retention rates intermediate between those given either strain 20P3 or strain 5 alone. Similarly, the actual colonization levels achieved 24 h after completion of the dosing regimen were higher for the subjects who received strain 5 (either alone or when combined with strain 20P3) (Table 3). Indeed, 17% of subjects given strain 5 had colonization levels of at least 1% at 24 h compared with only 5% of subjects dosed with strain 20P3 and 7% of those coinoculated with both strain 5 and strain 20P3. It is not clear why strain 5 is a more efficient colonizer than strain 20P3. Handley and coworkers have demonstrated the importance of fimbriae (group K antigen) in S. salivarius adhesion (Handley et al., 1984). However, both strain 5 and strain 20P3 react positive when tested for Lancefield group K antigen (Johnstone, 1987).

View this table:

Retention of colonizing strain following a 2- or 9-day course of BLIS-milk

Colonization regimeSample dayProportion of subjects whose tongue swabs yielded streptomycin-resistant Streptococcus salivarius after taking BLIS- milk supplemented with either 20P3, 5 or 20P3 + 5
Trial 1: 2-day application00/8100/8100/270
Trial 2: 9-day application00/1500/150NDND
  • * BLIS-milk consisted of equal CFUs of each strain.

  • Number of subjects positive for the colonizing strain (determined by the growth of S. salivarius on streptomycin-containing media)/total number of subjects.

  • Not determined as this strain combination was not used in the 9-day colonization regime.

  • P value 0.001 (Fischer's exact test).

View this table:

Level of colonization achieved 24 h after taking either 2-day or 9-day courses of BLIS-milk

Colonization regimeStrain used for colonizationNo. of subjects in groupNo. of subjects in group for whom the level of colonization* was
Trial 1: 2-day application20P3771311371240
Trial 2: 9-day application20P315027510
  • * Expressed as a percentage of the total Streptococcus salivarius population.

  • P=0.071 (Fischer's exact test) for the difference between the proportion of subjects colonized by strain 20P3 (i.e. 4/77) or by strain 5 (i.e. 13/78) at a level >1% of the total S. salivarius population.

In this study there was no attempt made to reduce the levels of the children's indigenous S. salivarius populations before they drank the BLIS-milk, and this undoubtedly contributed to the relatively low levels of colonization achieved. More recently we have found that with use of chlorhexidine pretreatment to effect reduction of the indigenous S. salivarius population, mean cell counts of 5.8 × 107 CFU mL−1 saliva are achieved 14 days postcolonization for the probiotic strain S. salivarius strain K12 (Burton et al., 2006b). Streptomycin-resistant S. salivarius strains were used in the present study to facilitate accurate assessment of their colonization levels. No tongue swabbings obtained either before or after colonization grew (streptomycin-resistant) streptococcal colonies on MSA+S, other than for the appropriate colonizing strain(s) themselves. It is known that the expression of antibiotic resistance can confer a fitness cost on bacteria, and, in the absence of a persistent selective antibiotic pressure, it has been observed that resistant bacteria are relatively disadvantaged in comparison with antibiotic-susceptible members of the same species (Andersson & Levin, 1999). Support for this observation can be found in the present study in the poor persistence of both of the streptomycin-resistant S. salivarius strains following colonization, and this appeared particularly marked when their initial colonization levels were low. By contrast, individuals harbouring naturally acquired populations of streptomycin-sensitive SalA-producing S. salivarius appear generally to retain the BLIS-producing bacteria for periods of at least 6 months (Tompkins & Tagg, 1989).

Trial 2: Dosing for 9 days with BLIS-milk containing SalA-producing strains 5 and/or 20P3

When the period of daily dosing with BLIS-milk containing either strain 5 or strain 20P3 was increased from 2 to 9 days (Table 2), the proportion of individuals (1) initially colonized (day 1 following completion of the colonization protocol) and (2) exhibiting persistent colonization was found to increase. Whereas in Trial 1, strain 20P3 was undetectable 9 days after colonization was completed, in Trial 2 80% of the children still had detectable levels of the colonizing strain 9 days after they had stopped taking the 20P3 BLIS-milk, and by day 16, 14% still remained positive. Similarly, strain 5 was still detected 9 days after taking strain 5 BLIS-milk in 87% of the Trial 2 children compared with only 28% positive at the corresponding time in Trial 1. In addition, the proportion of children harbouring relatively high initial (day 1 postcolonization) levels (i.e.>0.1% of their total streptococcal populations) was higher for the 9-day course (40% of those taking strain 20P3% and 57% for strain 5) than for the 2-day course (21% and 54% respectively) (Table 3).

Colonization with BLIS+S. salivarius may stimulate the BLIS activity of the pre-existing streptococcal microbiota

Just prior to taking BLIS-milk (day 0), 53 (65%) of the strain 20P3-dosed children, 50 (62%) of the strain 5 group, and 16 (59%) of the combined (strain 5/strain 20P3) group in Trial 1 had P-type 000 (i.e. apparently BLIS-negative) MSA-cultivable tongue populations as assessed by deferred antagonism against the set of nine standard BLIS indicator strains. However, the P-type profile of the tongue populations of 14 (8%) of these children changed to BLIS-positive soon (day 1) after they had taken BLIS-milk (Table 4). A further six subjects, although not having P-type 000 presamples, displayed substantial relative increases in the magnitude (zone width or number of affected indicators) of the P-type patterns given by their day 1 samples. In spite of its relatively less efficient colonization ability, strain 20P3 seemed at least as effective as strain 5 in bringing about this apparent enhancement of the total BLIS activity of the childrens' tongue microbiotas.

View this table:

Effect of 2 days of taking milk containing Streptococcus salivarius strains 5 and/or 20P3 on the BLIS activity of the streptococcal microbiota of the tongue

SampleProportion of BLIS+ reactions detected in tongue microbiota samples from subjects colonized with S. salivarius strain
Precolonization28/81 (35%)31/81 (38%)11/27 (41%)70/189 (37%)
Postcolonization34/81 (42%)36/79 (46%)14/27 (52%)84/187 (45%)
No. (%) changing to BLIS+6 (+7%)5 (+8%)3 (+7%)14 (+8%)
  • * Pre and postcolonization BLIS status was determined by P-typing samples of the MSA tongue swab cultures. A conversion from BLIS− to BLIS+ indicates increased BLIS production as a result of colonization, increased endogenous production, or both. Numbers (and %) of BLIS+ tongue microbiotas/total children tested are indicated for each group of subjects.

It has been our experience (results not shown) that SalA-producing S. salivarius such as strains 5 and 20P3 need to be present in proportions of at least c. 10% of the total test population in order for their inhibitory activity to be reliably detected in deferred antagonism tests such as used here to screen the BLIS activity of the childrens' streptococcal tongue populations. Hence the enhanced levels of BLIS activity detected in the 20 subjects described above could not be accounted for by the relatively small numbers of the colonizing strains (0.14% or less of the total streptococcal population) present in their day 1 samples.

Our previous studies have shown that both strain 5 and strain 20P3, when grown in human saliva, can produce auto-inducing levels of SalA (P. Wescombe, unpublished data), and furthermore that inducing levels of SalA can be detected in the saliva of individuals colonized with the SalA-positive probiotic S. salivarius K12 (Wescombe et al., 2006). Examination of samples of BLIS-milk showed that both the strain 5- and strain 20P3-supplemented milk had SalA titres of 8 AU mL−1. The control (unsupplemented milk) was inhibitor-negative in this assay. We speculated that the 20 children whose tongue populations converted from BLIS-negative to BLIS-positive may have had small populations (<10% of total streptococci) of SalA-positive S. salivarius just prior to taking the BLIS milk. To test for these putative SalA producers, we used PCR to detect salA (the SalA structural gene) in the precolonization samples from each of the 20 subjects (convertors) showing marked increase in their BLIS-producing status and also from 47 Trial 1 subjects who were consistently BLIS-negative (nonconvertors) (Table 5). The precolonization samples from 16 (80%) of the convertors were salA+ by PCR. In addition, single colony picks of their S. salivarius populations (from MSA cultures) detected BLIS-positive (inhibitory to Indicator I1) isolates from each of these subjects, and in every case the BLIS-positive colonies amounted to fewer than 5% of the total S. salivarius colonies tested. Moreover, for only one of the 16 PCR-positive convertor subjects was the level of colonization achieved on day 1 (viz 15%) by the BLIS-milk strain considered likely to be sufficiently high to account for the production of detectable BLIS in the P-type test. In every other case the level of colonization was <1%. Hence it seemed that most of the observed increase in total BLIS activity in the convertor children resulted from an increase in the relative numbers and/or inhibitory activity of BLIS-positive members of their indigenous streptococcal tongue populations. Similarly, just one of the four PCR salA-negative convertor subjects had sufficiently high levels (25% in this case) of the colonizing strain to account for their increased tongue microbiota BLIS activity. For the three remaining salA PCR-negative convertor subjects the colonizing strain accounted for only 0.6% or less of their total streptococcal tongue population. In summary, it appears that in these subjects the increase of total BLIS activity probably arises from the expansion/stimulation of the production of heterologous (non-SalA) BLIS, even if the mechanism for this is at present unclear. Only three of the 47 tested nonconvertors were salA PCR-positive, and in each case the levels of colonization achieved at day 1 were 0.14% or less. Failure to induce inhibitor production in some salA+ clones is understandable in view of our previous finding that the SalA locus is not always intact, deletions sometimes being evident in the processing and transporter genes (Wescombe et al., 2006).

View this table:

Relationship of preexisting PCR-positive salA status in 67 subjects to the development of increased tongue microbiota BLIS activity following exposure to BLIS-milk for 2 days

StrainNumber (%) of 19 salA+ and 48 salA subjects whose tongue microbiotas displayed
Increase in BLIS*No change in BLIS
Total16 (84)4 (8) §3 (16)44 (92)
  • * Increase in BLIS activity was monitored by change in P-type. Fourteen BLIS-negative children and six with substantial BLIS P-type increase were tested along with 47 randomly selected BLIS-negative subjects.

  • Precolonization samples were tested for the salA structural gene by PCR.

  • Numbers in brackets indicate the percentage of each group (salA+ or salA−) that showed an increase in BLIS P-type.

  • § §One child was colonized at 25%, so their increase in BLIS activity may be caused by the colonizing strain.

  • Numbers in brackets indicate the percentage of each group (salA+ or salA−) that showed no change in BLIS P-type.

On the basis of these observations we conclude that for individuals who have SalA+S. salivarius as a component of their indigenous oral microbiota the consumption of BLIS-milk (containing SalA+S. salivarius and some associated SalA peptide) may cause the level of salivary SalA to increase sufficiently to be sensed by the indigenous SalA+ population, thereby upregulating the expression of SalA and in turn increasing the proportion of SalA+S. salivarius within the tongue microbiota. In follow-up tests, three subjects having streptococcal tongue populations PCR-negative for salA and three subjects salA PCR-positive were given 2-day courses of (a) skim milk+chocolate powder (control) and then 2 weeks later (b) 20P3 BLIS milk. No substantial differences were detected between the BLIS P-type activities of the streptococcal populations on the subjects' tongues in samples obtained at day 0 and day 1 either following the taking of the control milk or for the subjects having salA PCR-negative precolonization tongue populations. By comparison, two of the three salA PCR-positive subjects showed a marked increase in the BLIS P-type activity of their day 1 tongue samples and this was accompanied by an associated increase in the proportion of indigenous SalA+S. salivarius. In one subject the proportion changed from 5% to 40% and in the second the change was from <1% to 45%. The third subject (also <1% precolonization) showed no apparent increase in SalA+ proportion. As previously, the level of colonization at day 1 by strain 20P3 (streptomycin resistant) was <5% in all six subjects.

Deferred antagonism testing of S. salivarius strain 20P3 against 18 indicator strains of S. salivarius (11 were salA+ and seven salA−) showed that all seven salA− strains and only two of the salA+ strains were inhibited (Table 6). Interestingly, however, neither of these two salA+ strains appeared capable of expressing SalA. In strain K33 a transposon had inserted within the SalA locus, but the reason for lack of expression in strain Pirie is at present not known. This illustration of the in vitro anti-S. salivarius activity of strain 20P3 provides further support for the hypothesis that enhanced salivary levels of SalA, either transiently acquired from the BLIS milk or from that produced (possibly at induced levels) by the indigenous or newly introduced SalA+S. salivarius, can lead to a clonal expansion of S. salivarius. These expanded S. salivarius populations may in turn, by producing and exhibiting immunity to SalA, outgrow non-SalA+ members of the subject's S. salivarius indigenous microbiota and thus potentially increase local protection of the host against S. pyogenes infection.

View this table:

Correlation of the SalA sensitivity of Streptococcus salivarius strains and their producer and salA status

Streptococcus salivarius strainsalA*P-typeSensitivity to strain 20P3 (SalA+)
K8P ori226+
K19 P1777+
  • * PCR primers can detect any of the five reported salA variants (Wescombe et al., 2006).

  • Does not express SalA.

  • Tn916-derived SalA- (i.e. does not express SalA or immunity to SalA) variant of S. salivarius strain 20P3.

  • § §P-type is variable for this strain, sometimes being 677.


This work was supported by grants from the Thrasher Research Fund, the Health Research Council of New Zealand, and the National Heart Foundation of New Zealand.


  • Present address: Karen P. Dierksen, Department of Microbiology, Oregon State University, Corvallis, OR, USA.

  • Present address: Philip A. Wescombe, Chris J. Moore, BLIS Technologies Ltd, Centre for Innovation, Dunedin, New Zealand.

  • Editor: Julian Marchesi


View Abstract