Page 95 - Read Online

Page 10 of 15 Longhi et al. Microbiome Res Rep 2024;3:4 https://dx.doi.org/10.20517/mrr.2023.02

[33]
their superficial biological structures . Compared to Gram-positives, Gram-negative bacteria possess a
much thinner cell wall covered by an external membrane. This has been reported to increase susceptibility
to specific cell lysis methods, thus explaining the drastic decrease in the number of Gram-negative
associated DNA after saponin/DNase treatment [31,32] . Regarding this, it has been widely demonstrated that
the antibacterial activity of saponins depends on their chemical structures; indeed, several natural
compounds with surfactant properties extracted from plants are able to break down the bacterial cell wall,
resulting in the leakage of the cell contents and affecting bacterial metabolism adversely [34,35] .

Supporting this hypothesis, we observed that the sum of the relative abundance of Gram-negative bacteria
in the antrum and the gastric body biopsies was completely depleted after saponin treatment from 91.6%
and 79.7%, respectively, to 0%. Furthermore, in saliva and sputum samples where Gram-negative bacteria
represented 74.1% and 48.8%, respectively, of the total amount of bacterial content in the control samples,
their presence was completely depleted after treatment with saponin due to a dominance of Gram-positive
bacteria [Figure 2A and Supplementary Table 4]. In contrast, the biological samples where the microbial
populations were mainly represented by Gram-negative bacteria, such as skin and nasopharyngeal swabs,
were least affected due to a reduced shift in the Gram-positive/Gram-negative ratio, thus explaining the
limited impact of saponin treatment on the taxonomic profiles obtained for these biological matrices
[Figure 2A and Supplementary Table 4].

Overall, these data underline that saponin can markedly lower the Gram-positive/Gram-negative ratio in
the recovered bacterial DNA, thus altering the post-sequencing microbial profiles. In this regard, a
preliminary taxonomic investigation of the microbial population is important to evaluate whether saponin
will be acceptable in altering the taxonomic profiles. In fact, in the case of samples with a bacterial
population dominated by Gram-positive taxa, alterations by saponin treatment will be more contained and
acceptable in terms of the accuracy of data analysis.

Comparison of the effect of different saponin percentages on the microbial profile
Although 2.5% wt/vol is the generally used amount of saponin in most protocols used for eukaryotic DNA
removal, we decided to evaluate whether a lower amount of this detergent could avoid the bias in the
estimation of Gram-negative/Gram-positive bacteria ratio in the biological samples assayed. To do this,
DNA isolated from the same sputum sample was processed with variable concentrations of saponin, i.e.,
0.0125%, 0.05%, 0.1%, 0.5%, 1.5% and 2% wt/vol. DNA obtained from microbial DNA extraction was
subjected to shallow shotgun metagenomic analysis.

Notably, the analysis of the achieved metagenomic data revealed that saponin had an impact on both
human DNA abundance and Gram-positive/Gram-negative ratio directly proportional to the amount used.
In fact, host DNA content ranged from 36.4% in the 0.0125% to 3.2% in the 2% wt/vol saponin-treated
saliva, compared to the control sample in which the host DNA content was 53.2% [Table 2].

To better detail the impact of saponin concentration, we focused our interest on the key Gram-negative
bacterial taxa previously identified as dominant in the untreated sputum sample. The Gram-negative
Prevotella histicola was present in the control sample with a relative abundance of 19% and decreased from
9.1% to 5%, respectively, in the sample treated with 0.0125% wt/vol and 2% wt/vol of saponin
[Supplementary Table 5 and Supplementary Figure 3]. Similarly, the Gram-negative Veillonella atypica
decreased from an initial relative abundance of 8.1% in the untreated to 4.5% and 2.6%, respectively, in the
0.0125% and 2% wt/vol treated samples [Supplementary Table 5 and Supplementary Figure 3]. In contrast,
the Gram-positive Streptococcus salivarius was detected as a dominant bacterial taxon in the sputum treated