3.1 Sample collection, environmental data, and assessment of condition index
C. gallina samples were collected from the 4 different production sites along the Western coast of the North Adriatic Sea (Fig. 1). Sampling was carried out during the summer season (September 2022) on the same day at all sites, to avoid temporal variations in environmental conditions during the sampling campaign. For each site, 65 clams and a corresponding seawater and sediment sample per site were collected, for a total of 260 clam individuals, 4 seawater samples and 4 sediment samples. Environmental parameters on the day of sampling are reported in Table 1. A total of 15 clams per site were used for the assessment of the Condition Index (CI) as a proxy for the general nutritional and health state of the animal [43]. The CI values showed a decreasing trend from the Southern to the Northern sites, with Rimini having a CI score significantly higher than the other sites and Marina di Ravenna, Lido di Savio, and Cesenatico having comparable values (two-way ANOVA) (Table 2). According to these data, Rimini was considered the only high-CI site, while Marina di Ravenna, Lido di Savio, and Cesenatico were considered low-CI sites.
Table 1 - Environmental parameters on the day of sampling at the 4 sites.
Site
Parameter
|
Marina di
Ravenna
|
Lido di
Savio
|
Cesenatico
|
Rimini
|
Latitude
|
44°28’38’’N
|
44°18’44’’N
|
44°11’56’’N
|
44°03’48’’N
|
Longitude
|
12°17’09’’E
|
12°20’44’’E
|
12°23’43’’E
|
12°34’51’’E
|
pH
|
8.05
|
8.06
|
8.05
|
8.06
|
Salinity (psu)
|
37.45
|
37.61
|
37.85
|
38.05
|
OD (mmol/m3)
|
213.98
|
213.94
|
214.02
|
214.11
|
Temperature (°C)
|
22.56
|
22.60
|
22.76
|
22.81
|
Chlorophyll a (mg/m3)
|
0.31
|
0.30
|
0.27
|
0.26
|
Abbreviations: psu = practical salinity unit; OD = dissolved molecular oxygen. Data were retrieved from Copernicus website (https://data.marine.copernicus.eu/product/MEDSEA_ANALYSISFORECAST_BGC_006_014/description and https://data.marine.copernicus.eu/product/MEDSEA_ANALYSISFORECAST_PHY_006_013/description)
Table 2 – Condition index values for all sampling locations.
Site
|
Marina di Ravenna
|
Lido di Savio
|
Cesenatico
|
Rimini
|
CI (mean ± SD)
|
18.1 ± 3.0
|
20.3 ± 2.7
|
20.0 ± 2.4
|
25.9 ± 4.1
|
ANOVA
|
|
Marina di Ravenna
|
Lido di Savio
|
Cesenatico
|
Rimini
|
Marina di Ravenna
|
1
|
> 0.05
|
> 0.05
|
< 0.001
|
Lido di Savio
|
|
1
|
> 0.05
|
< 0.001
|
Cesenatico
|
|
|
1
|
< 0.01
|
Rimini
|
|
|
|
1
|
Condition index (CI) was calculated for clams collected at all sampling sites. Two-way ANOVA test was used to assess significant differences in CI values among sites.
3.2 Core and variable fractions of the clam digestive gland-associated and sediment microbiomes
For each site, 36 out of the 50 collected clams were pooled on base 3, for a total of 12 pools per site, which were used to profile the DG-associated microbiome by 16S rRNA amplicon sequencing. The microbial communities of corresponding sediments and seawater were also characterized. Sequencing was performed on a total of 48 clam pools, 4 seawater and 4 sediment samples, resulting in 6,392.9 ± 4,117.3 mean high-quality reads per sample and 1,183 ASVs. The microbial compositional structure at phylum and family level is shown in Suppl. Fig. S1. Overall, the clam DG-associated microbiome was dominated by Firmicutes (42%) and Planctomycetota (18%), with Proteobacteria (11%) and Spirochaetota (11%) as subdominant phyla. The seawater microbiome was dominated by Proteobacteria (43%), Actinobacteriota (22%), Firmicutes (17%), and Bacteroidota (14%), whereas the sediment microbiome was dominated by Firmicutes (33%), Proteobacteria (24%), and Actinobacteriota (18%). Alpha and beta diversity measurements of the described microbial datasets (Fig. 2) showed a clear segregation of clam-associated and environmental microbiomes (p-value ≤ 0.001 for beta diversity), with the former showing lower levels of alpha diversity (p-value ≤ 0.05 for the alpha diversity comparison between clam and environmental microbiomes in all metrics; see Fig. 2 for further details). According to the PCoA plot, the clam DG-associated microbiome was closer to the sediment microbiome compared to seawater samples. These results confirm previous observations [44, 45] and emphasize the important connection between the clam-associated and sediment microbiomes. The separation in the PCoA also highlighted a certain heterogeneity of the clam DG-associated microbiome according to collection sites (p-value ≤ 0.05, data not shown), confirming the observed site-specific features of the clam DG-associated microbiome [19, 46], even at a local scale. In Suppl. Fig. S2, we provide the compositional profile of the DG-associated microbiome at each sampling site. Despite a certain degree of site specificity, a core DG-associated microbiome of C. gallina was detectable, defined as the taxa with a prevalence higher than 70% in our sample set. This core included the phyla Firmicutes, Planctomycetota, Proteobacteria, and Verrucomicrobiota, and the families Mycoplasmataceae, Peptostreptococcaceae and Pirellulaceae as major components. On the other hand, no significant differences in the alfa diversity of the DG-associated microbiome were observed between sites (Suppl. Fig. S3). Finally, focusing on the sediment microbiome variation among sampling sites, we found that sediments from the high-CI site were characterized by several microorganisms of environmental origin, both marine and terrestrial, such as members of the Ilumatobacteraceae, Flavobacteriaceae, Hungateiclostridiaceae, Rhizobiaceae, Xanthobacteraceae and Rubritaleaceae. Conversely, sediments from low-CI sites were enriched in host-associated or opportunistic microorganisms, such as Lactobacillaceae, Streptococcaceae, Paenibacillaceae, Staphylococcaceae and Pseudomonadaceae (Suppl. Table S1).
3.3 Variations in the clam digestive gland-associated microbiome according to condition index
To identify the compositional specificities of the DG-associated microbiome at high- and low-CI sites, we first applied LEfSe [40] at the ASV level (Fig. 3), and then used BLAST [41] to assign the corresponding bacterial species to the discriminant ASVs (Suppl. Table S2). According to our findings, the high-CI site was characterized by 7 discriminant taxa with a best hit corresponding to the families Lachnospiraceae (percentage identity, 93.10%), Prevotellaceae (94.88%), Odoribacteraceae (93.27%), Muribaculaceae (86.67%), and Simkaniaceae (90.32%), and the genera Bacteroides (97.12%) and Mariniblastus (97.77%). Low-CI sites were characterized by 4 discriminant taxa with a best hit corresponding to the families Mycoplasmataceae (genera Mycoplasmopsis, 90.61%, and Mycoplasma, 88.73%), Verrucomicrobiaceae (91.67%) and Pirellulaceae (94.74%).
To highlight the possible connections between these DG-associated microbiome components and the respective environmental ecosystem, we investigated their presence in the corresponding water and sediment microbial ecosystems. The heatmap of the relative abundance of the abovementioned discriminant taxa in the sediment and water microbiomes from high- and low-CI sites is shown in Fig. 4. Overall, the presence of DG-associated taxa characterizing the high-CI site was sporadic in the environmental microbiomes, with only Lachnospiraceae present in the Rimini seawater, Lachnospiraceae and Prevotellaceae in the Ravenna and Cesenatico sediments, and Mariniblastus in the respective seawater. Conversely, the DG-associated microbiome components characterizing low-CI sites were most pervasive in the environmental microbiomes. Both ASVs assigned to Mycoplasmataceae were quite pervasive in seawater and sediments from Savio, Cesenatico and Rimini. Verrucomicrobiaceae and Pirellulaceae were detected in Ravenna seawater, with Pirellulaceae also detected in Cesenatico seawater.
The functional features of the discriminant taxa identified for high- and low-CI sites were inferred using PICRUSt2 [42]. To emphasize the respective specificities while excluding core functionalities, only exclusive functions were considered, defined as KO_ASVs exclusive for at least 2 taxa for each discriminant group. According to our findings, DG-associated microbiome taxa discriminating high- and low-CI sites showed different metabolic propensities, especially for pathways involved in carbohydrate, lipid, amino acid, nucleotide, and energy metabolism (Suppl. Fig. S4). Interestingly, only DG-associated microbial taxa characterizing the high-CI site were endowed with pathways involved in the metabolism of cofactors and vitamins, including, among others, the biosynthesis of pyridoxine (vitamin B6), folate, riboflavin and terpenoids, suggesting their possible role as health-promoting bacteria (HPB) (Fig. 5).
3.4 Proof-of-concept experiment in controlled environment
In order to further explore possible connections between the clam DG-associated microbiome components, the surrounding sediment microbial communities and animal health, experiments were conducted under controlled conditions. We tested the hypothesis that sediments from the high-CI site (i.e., Rimini) would favor the increase of health-promoting microorganisms in clams from a low-CI site (i.e., Ravenna), possibly resulting in an improved physiological status of the animal. To this end, clams collected in Ravenna were reared in aquaria on both Ravenna and Rimini sediments. The DG-associated and sediment microbiomes were assessed at 3 time points: T0 (right after the experimental set-up), T1 (after 7 days of incubation) and T2 (after 21 days of incubation). For each aquarium, 15 clams were collected and pooled on base 3 at T0 and T1, and 9 clams were collected and pooled at T2, for a total of 52 clam pools, resulting in 5,364.2 ± 2,615.7 mean high-quality reads per pool and 2,078 ASVs. The remaining clams were used to assess CI values at T1 and T2. According to our observations, Ravenna clams reared on Rimini sediments showed a better performance in terms of ΔCI when compared to their initial condition (Ravenna clams on Ravenna sediments). More specifically, after the first week of acclimation, the ΔCI between T2 and T1 for Ravenna clams reared on Rimini sediments was > 1 (ΔCIRa-Ri = 1.37), whereas for Ravenna clams reared on Ravenna sediments it was close to 0 (ΔCIRa-Ra = -0.15). Interestingly, when we assessed the total relative abundance of previously identified HPB (namely, Lachnospiraceae, Prevotellaceae, Odoribacteraceae, Muribaculaceae, Simkaniaceae, Bacteroides and Mariniblastus) in the clam DG-associated microbiomes in controlled environment, we observed that clams reared on Rimini sediments maintained a certain proportion of these microorganisms up to T2 (570 RPKM, where Reads Per Kilobase per Million reads mapped were calculated by dividing the number of reads mapped to each reference sequence by the mean kilobase length of that sequence and the total number of reads in that sample times 1 million). Conversely, the identified HPB progressively disappeared in the clams reared on Ravenna sediments during the observation time (0 RPKM at T2). No relevant difference in the reads count of these microorganisms was observed between Ravenna and Rimini sediments. Taken together, our data suggest the possibility that sediments from Rimini may favor the physiological status of the clam, by promoting the acquisition of certain microorganisms, resulting in an overall improvement in animal health.