Looking for Pathogens in Dust from North Africa Arriving in the French West Indies Using Metabarcoding and Cultivable Analysis

Abstract

Periodically, the French West Indies receive dust originating from North Africa (NA). Microorganisms associated with desert dust can be transported over long distances through the atmosphere and could represent a means for the remote colonization of new habitats by putatively pathogenic microorganisms. The aim of this study was to determine the diversity and frequency of microbial agents (bacteria, eukaryotes) in NA dusts and the potential threat toward human and/or animal health by comparing microbial air composition during dust events and in control samples. In 2017 and 2018, 16 samples were collected during seven NA dust episodes and there were 9 controls. The microbial composition of the samples was characterized using a cultivable approach and by metabarcoding analyses (16S and 18S). A greater bacterial load and greater diversity were observed during the dust events, and some genera were significantly associated with the events. Some, such as Geodermatophilus, can be considered signature species of NA dust. No pathogenic species were found with the cultivable approach, whereas the metabarcoding analyses highlighted the presence of several potentially pathogenic species or known human pathogens such as Naegleria fowleri.

Keywords: French West Indies; dust; metabarcoding; microbial pathogens.

Figure 1

Graphical abstract of the study. Desert dust from North Africa regularly crosses the Atlantic Ocean to settle on the West Indies; increasing particle concentration of dust event is indicated on the map from yellow to orange (Event on 18 July 2018, source: https://www.aoml.noaa.gov/our-research/hurricane-research-division/saharan-air-layer/ (accessed on 10 March 2020)). These dusts are able to transport viable microbes that can have an effect on human and animal health but also on the environment. We undertook this work to characterize these microbes and search for pathogens. After dust sampling on Petri dishes and filters, microorganisms were identified by two complementary approaches, a cultivable method and by 16S and 18S rDNA sequencing. No pathogenic microorganisms were found with the cultivable method, but we could evidence the presence of the genetic material of several putative pathogenic species.

Figure 2

Number of CFU/m³ of bacteria in air samples on different media (TSA, trypticase soy agar; PCA, plate count agar; CO, chromagar orientation).

Figure 3

Frequency of the main bacterial species isolated according to sampling condition (during dust events and in control samples).

Figure 4

Beta and alpha diversity analyses of bacterial communities at genus taxonomic level. PCoA plots according to sampling condition (during dust events or in control samples, CSs) based on a Bray–Curtis dissimilarity matrix for (a) the February 2018 dataset (indicated by Feb. 2018) and (b) the November 2018 dataset (indicated by Nov. 2018). PERMANOVA test based on sample type yielded p = 0.498 and 0.443, respectively; 58.3% and 47.9% of variations were explained by the first two PC1 and PC2 axes for the February and November datasets, respectively. (c) Alpha diversity analysis based on alpha, inverted Simpson, Shannon and Simpson indices during dust events and in control conditions; error bars represent 95% confidence intervals.

Figure 5

Dominant bacterial genera according to sample type (during dust events or control samples, CSs) for the February 2018 dataset (indicated by Feb. 2018) and November 2018 dataset (indicated by Nov. 2018). (a) Bar plot of relative abundance. (b) Box plot of log2 abundance.

Figure 6

Beta and alpha diversity analyses for eukaryotic communities at genus taxonomic level. PCoA plots according to sampling condition (during dust events or in control conditions, CSs) based on a Bray–Curtis dissimilarity matrix for (a) the February 2018 dataset (indicated by Feb. 2018) and (b) the November 2018 dataset (indicated by Nov. 2018); a PERMANOVA test based on sample type yielded p = 0.603 and 0.156, respectively; 40.7% and 39.6% of variations were explained by the first two PC1 and PC2 axes for the February 2018 and November 2018 datasets, respectively. (c) Alpha diversity analysis based on alpha, inverted Simpson, Shannon and Simpson indexes during dust events and in control regular conditions; error bars represent 95% confidence intervals.

Figure 7

Dominant eukaryotic genera according to sample type (during dust events or control samples, CSs) for the February 2018 dataset (indicated by Feb. 2018) and November 2018 dataset (indicated by Nov. 2018). (a) Bar plot of relative abundance. (b) Box plot of log2 abundance.