It was possible to select unique sequences about 16-30 basepair in length that discriminated each species (except F. nucleatum) down to the level of one mismatch when compared to the respective target sequences of all other species described so far. The Fusobacterium nucleatum probe also cross-hybridized with the respective sequence of Fusobacterium periodonticum. These species exhibit a 100% homology in the 528 initial basepairs of the 16S rRNA (178). Besides F. nucleatum the genus Fusobacterium comprises a wide range of species, such as F. necrophorum, F. mortiferum, F. simiae, F. gonidoformans, F. alocis, F. varium, F. russi, F. ulcerans, F. periodontiticum, F. perfoetens, all complementary to the respective probe. Most of the species are considered a part of the indigenous oral flora.
All probes are non-self-complementary and with low G-C content (checked with the program OLIGO).
The oligonucleotide probe sequences are listed in Table 1.

PROBES 5' - 3'

T°C (Hyb)

T°C (Wash)

Washing buffer

Actinobacillus actinomycetemcomitans (A.a.)
TCC-ATA-AGA-CAG-ATT-C

40°

47°

WP 0

Tannerella forsythensis (T.f.)
CGT-ATC-TCA-TTT-TAT-TCC-CCT-GTA

54°

64°

WP 0

Campylobacter rectus (C.r.)
TTA-ACT-TAT-GTA-AAG-AAG

40°

40°

WP 0

Capnocytophaga ochracea (C.o.)
TCG-GGC-TAT-CCC-CCA-GTG-AAA-GGC-AGA-T

54°

61°

WP 2

Eikenella corrodens (E.c.)
AGT-TAT-CGG-CCG-CTC-GAA-TAA-CGC

54°

62°

WP 1+2

Fusobacterium nucleatum/F. periodonticum (F.n.) GCC-TCA-CAG-T(C,G)TA-GGG-ACA-ACA-T

54°

67°

WP 1

Fusobacterium spp. (Fuso)
GAG-AGC-TTT-GCG-TCC-CAT-TAG

54°

62°

WP 1

Porphyromonas gingivalis (P.g.)
CAA-TAC-TCG-TAT-CGC-CCG-TTA-TTC

54°

62°

WP 1+2

Prevotella intermedia (P.i.)
CTT-TAC-TCC-CCA-ACA-AAA-GCA-GTT-TAC-AA

54°

62°

WP 1

Veillonella parvula (V.p.)
TCT-AAC-TGT-TCG-CAA-GAA-GGC-CTT-T

54°

56°

WP 0

Table 1. Probe sequences (5'-3') with respective hybridization conditions. Washing buffers listed in Suppl., 2.5.-2.7.

In order to obtain positive and negative controls for the probe optimization and evaluation, the reference bacterial strains were cultured and their identity was checked by sequencing the PCR products or by biochemical tests (Suppl., Table 1). 16S rDNA of these strains was amplified in the PCR and the products were dotted onto the nylon membranes which were further used for the optimization and evaluation of the constructed probes.

Membranes were initially hybridized with eubacterial probe (EUB 338) to demonstrate successful PCR-amplification and immobilization for all samples investigated.

Oligonucleotide probes were able to distinguish between complementary and nearly complementary sequences on the basis of single mismatch achieving the expected specificity. However, this requires stringent hybridization conditions which often results in a reduction of the quantitative probe binding to the target nucleic acid (179). The intention of this study was to achieve an optimal balance between probe sensitivity and specificity by optimizing hybridization parameters, such as the washing temperature and the ionic strength of the washing buffer.
Each probe was tested against a panel of PCR products isolated from 42 reference microorganisms common to the oral cavity or extra-oral regions (Suppl., Table 1). The analysis indicated that all probes (except for Fusobacterium nucleatum) were specific and did not cross-hybridize with closely related species. The hybridization conditions for each probe are listed in Table 1.
Using the described assays the detection of target bacteria in multiple samples could be performed rapidly (1h DNA isolation, 4 h PCR-amplification, 5 h dot-blot hybridization, 1 h film exposure).

A total of 224 subgingival samples from 45 GAP patients and 84 samples from 21 elderly subjects were analyzed.
All investigated species were detected in the subgingival plaque of GAP patients and well-maintained elderly (for the prevalence calculation one GAP patient was excluded because of a missing control site). All GAP patients and elders were colonized by the members of genus Fusobacterium, however, F. nucleatum/F. periodonticum were detected in 91% of GAP patients but significantly less (57%) in elders (Fig.1). T. forsythensis occurred in 95.5% of GAP patients and in 85.7% of the elders showing no significant difference between both groups. The microorganism appears to be a common colonizer even in well-maintained individuals. The prevalence of P. gingivalis and P. intermedia in the GAP patients was 63.6% and 70.5%, respectively. In this study no significant difference was observed between the groups. From the GAP patients 36.4% harbored A. actinomycetemcomitans. In contrast, only two out of 21 elderly subjects were colonized by this species. As compared to healthy controls, significantly more GAP patients were colonized by E. corrodens (75%) and C. rectus (56.8%) than the elderly. The prevalence of V. parvula was moderate in both groups (GAP 25.4%, elderly 42.9%), without statistically significant difference. A highly significant difference (p<0.0001) was observed for C. ochracea. As much as 95.2% of the elders were positive, while the prevalence in GAP population was only 15.9%. The prevalence calculations suggest the evidence of colonization with target species in both groups (Fig.1).

Fig. 1 Bar chart showing the prevalence of the respective species in 44 GAP patients (one GAP patient was excluded because of a missing control site) and 21 elderly subjects. A patient was regarded positive when at least one site harbored the respective species. The significance of differences between the groups was calculated using the chi-square test.

The detection frequency of the bacteria in periodontal lesions and healthy control sites of 44 GAP patients was studied (Fig. 2). One arbitrarily chosen pocket and one control site per patient was included in the analysis. The patients revealed a high number of pockets colonized by Fusobacterium spp. (97.7%), F. nucleatum/F.p. (70.5%), T. forsythensis (88.6%) and P. gingivalis (59%) (Fig. 2). P. intermedia, E. corrodens and C. rectus could be identified in 30%-40% of the pockets. The comparison between positive pockets and control sites demonstrated a highly significant difference (p<0.001) for Fusobacterium spp., F. nucleatum, T. forsythensis and P. gingivalis. These species, as well as C. rectus and P. intermedia, were more frequently detected in diseased sites as compared to clinically healthy sites. Low detection frequencies in periodontal lesions were observed for A. actinomycetemcomitans (20.5%), V. parvula (6.8%) and C. ochracea (2.4%). However, all the investigated species (except V. parvula) could be identified in the control sites, but less frequently (Fig. 2).
Colonization of sites without clinical signs of a disease is still remarkable. T. forsythensis could be identified in 34%, F. nucleatum in 25%, P. gingivalis and E. corrodens in 22.7%, P. intermedia and A. actinomycetemcomitans in 15.9% of the control sites of GAP patients (Fig. 2).

Fig. 2. Percent of positive periodontal pockets and control sites in 44 GAP patients. The significance of differences was evaluated using the chi-square test.

More detailed information about the differences in the prevalence of species between the two groups could be gained from the comparison of the number of positive pockets of GAP patients and positive sites of the elderly (Fig. 3). From every subject four sites were included (from GAP patients only pockets) into the calculations. Pronounced discrepancies were observed between the groups. All species, except V. parvula, were significantly more frequently detected in the pockets of GAP patients as compared to the elderly (Fig. 3). For most of the species (except P. intermedia) the difference was highly significant (p<0.0001).
Interestingly, T. forsythensis was identified in as much as 48.8%, P. gingivalis in 32.1%, F. nucleatum in 25% and P. intermedia in 28.6% of the sites of elderly. The prevalence of V. parvula was similar in both groups (15% positive sites). In the elders E. corrodens, C. rectus and A. actinomycetemcomitans were detected with lowest frequency (8.3%, 9.5% and 3.6%, respectively). Only C. ochracea was found significantly (p<0.0001) more frequently in the control group (54.8% positive sites) (Fig. 3).

Fig. 3. Site-prevalence of the species in 45 GAP patients and 21 elderly (4 sites per subject were included). The significance of the differences between the groups was determined using the chi-square test.

The comparison of shallow sites (PD 1-3 mm) of GAP patients and the elderly revealed no significant difference for most of the species (Fig. 4).

Fig. 4. Comparison between the colonization of shallow sites (PD 1-3 mm) in 44 GAP patients and 17 elderly. Only one site per subject could be included. The significance of differences was evaluated using the chi-square test.

The exception was C. ochracea, which was more often detected in the elderly subjects (70.6% in elderly versus 6.8% in GAP patients). The difference was highly significant (p<0.0001). V. parvula could be detected more frequently from the shallow sites of the elderly as well (p<0.05). However, the total number of shallow sites in the control group was very small (n=17). As only one control site was sampled from the GAP patients only one arbitrary chosen site from the elderly subjects could be included in the statistics.
Comparison of the presence of the bacteria between sites with PD 1-3 mm and pockets with PD 4-7 mm in the GAP patients demonstrated a significant difference for most of the species. Only 23 patients with available samples from every pocket depth category could be included in the statistics. The 4-7 mm pockets of the GAP patients were colonized significantly more frequently by the investigated bacteria than the control sites, with the exception for A. actinomycetemcomitans, C. ochracea, and V. parvula (Fig. 5).

Fig. 5. Bar chart of the percent of positive sites with probing depth 1-3 mm, 4-7 mm and 8-12 mm in GAP patients (n=23). The significance of differences was determined using the chi-square test.

The presumption that the putative pathogens occur mostly in deep pockets of periodontitis patients could not be confirmed. No significant difference was observed between the number of deep and moderate lesions positive for any species in the GAP patients. C. ochracea was never detected in deep pockets (Fig. 5).
The bacterial load in a patient was determined by the number of positive sites. In general, fewer sites in a patient were colonized by putative periodontal pathogens in the healthy elderly than in the GAP group (Fig. 6). The data for T. forsythensis and F. nucleatum showed highly significant difference (p<0.0001).

Fig.6. The bars depict the percent of subjects with either 0, 1, 2, 3 or 4 positive sites for T. forsythensis, P. gingivalis, F. nucleatum, P. intermedia, E. corrodens and C. ochracea in GAP group (n=45) and elderly (n=21). The significance of differences between the groups were determined using the Mann-Whitney test.

The load of T. forsythensis, P. gingivalis and F. nucleatum was high in the GAP patients. The number of patients with four positive pockets was 73.3%, 53.3% and 51.1%, respectively. The load of these species was significantly lower in the elderly. However, 37.7% of the GAP patients revealed no colonization of P. gingivalis in any sampled pocket. In comparison with the elderly significantly more pockets per patient in the GAP group were positive for E. corrodens, C. rectus and A. actinomycetemcomitans (p<0.05). Interestingly, A. actinomycetemcomitans occurred in all four sampled pockets in 11% of the GAP patients. The load of P. intermedia and V. parvula showed no significant difference between the groups.
C. ochracea consistently was detectable in more than one site in the healthy subjects. The load of this species was significantly (p<0.0001) higher in the elderly subjects than in the GAP patients (Fig. 6).

An attempt to investigate bacterial profiles in subgingival areas and assess how complexes of bacteria relate within an ecosystem was made by Socransky et al. (177). A simplified approach using the 2x2 contingency tables was applied in the current study to quantify strengths of associations between two bacterial species. Odds ratio (OR) >2 shows a positive and <0.5 a negative association.

P.g.

P.i.

F.n.

Fuso

T.f.

C.r.

E.c.

A.a.

V.p.

C.o.

P.g.

--

--

--

--

--

--

--

--

--

--

P.i.

2,7

--

--

--

--

--

--

--

--

--

F.n.

1,5

1

--

--

--

--

--

--

--

--

Fuso

12,9

14,3

--

--

--

--

--

--

--

--

T.f.

23,5

10,7

9,1

0

--

--

--

--

--

--

C.r.

2

2,3

10,5

0

35,6

--

--

--

--

--

E.c.

0,2

0,7

6,3

3,6

1,4

0,2

--

--

--

--

A.a.

0,5

0,9

2

0,8

0,6

0,8

5,8

--

--

--

V.p.

0,8

0,7

0

0

2,3

3,5

5,7

1,6

--

--

C.o.

0,4

0,5

0

0

0,3

1,5

4

2,7

8,9

--

Table. 2. Odds ratios of associations among investigated species. The value 0 is the result of the absence of one species in any site in one group.

A strong positive association exists between T. forsythensis and C. rectus (OR 35.6), P. gingivalis (OR 23.5) or P. intermedia (OR 10.7). Fusobacterium spp. occurred frequently together with P. gingivalis (OR 12.9) and P. intermedia (OR 14.3). F. nucleatum was observed frequently with C. rectus (OR 10.5). E. corrodens and A. actinomycetemcomitans occurred together relatively often (OR 5.8).
A negative association was observed between E. corrodens and C. rectus (OR 0.2) or P. gingivalis (OR 0.2).