We identified a total pool of 18,816 sequence variants (SVs) among the total 8,235,928 reads encompassed by the data set, with an average of 59,251 reads per sample. SVs distributed among 33 unique eukaryotic phyla throughout the study, where Streptophyta dominated the communities, accounting, on average, for 25.35% of the SVs found throughout the study. Sequence variants richness in individual samples ranged from 2 to 2,265, with an average (± SE) of 572 ± 43.46 SVs per sample based on the Chao1 index at the phylum level. A total of 5562 SVs were identified more than 2 times among the total 4,161,129 reads were identified in the onshore air samples. In the offshore air samples, 2444 SVs with ≥ 2 counts among 2,843,534 total read counts. However, the Red Sea surface water contains 1831 SVs with 2 or more counts among the total 1,095,969 reads.
Reads that were retrieved from airborne dust samples were comprised mainly by Streptophyta (29.26% represented mainly by Rosids and Liliopsida), Apicomplexa (12.93% represented only by Conoidasida), Ascomycota (6.20% represented mainly by Sordariomycetes and Saccharomycetes), Basidiomycota (6.14% affiliated mainly by Malasseziomycetes and Tremellomycetes), Chordata (4.23% represented mostly by Mammalia (Cetacea) and Aves), Annelida (1.85% dominated by Polychaeta), Chlorophyta (1.23% represented mainly by Chlorophyceae), and Chytridiomycota (1.54% represented mainly by Chytridiomycetes).
From offshore air, reads of 18S rRNA genes mostly belonged to the following phyla: Streptophyta (39.64% represented mainly by Liliopsida), Apicomplexa (10.15% represented only by Conoidasida), Dinoflagellates (6.90% represented by Dinophyceae), Basidiomycota (6.44% represented mainly by Malasseziomycetes and Tremellomycetes), Chordata (4.94% represented mostly by Mammalia (Primates and Cetacea) and Aves), Ascomycota (4.16% represented mainly by Sordariomycetes and Saccharomycetes), Chlorophyta (0.97% represented mainly by Chlorophyceae), Annelida (0.95%% dominated by Polychaeta), and Chytridiomycota (0.13% represented mainly by Chytridiomycetes).
Sequences identified from onshore air samples predominantly belong to the following phyla: Streptophyta (22.04% represented mainly by Liliopsida), Apicomplexa (14.87% represented only by Conoidasida), Ascomycota (7.61% represented mainly by Sordariomycetes and Saccharomycetes), Basidiomycota (5.94% represented mainly by Malasseziomycetes and Tremellomycetes), Chordata (3.74% represented mostly by Mammalia and Aves), Dinoflagellates (2.70% represented by Dinophyceae), Annelida (2.48% dominated by Polychaeta), Chytridiomycota (2.39% represented mainly by Chytridiomycetes), and Chlorophyta (1.41% represented mainly by Chlorophyceae).
Reads of 18S rRNA gene region retrieved from surface waters (one m depth) were associated to the following phyla: Dinoflagellates (50.95% represented by Dinophyceae), Arthropoda (10.16% mainly belong to Hexanauplia), Chlorophyta (6.65% affiliated mostly by Mamiellophyceae and Chloropicophyceae), Chordata (2.79% represented only by Appendicularia), and Cnidaria (2.09% affiliated by Hydrozoa) (Fig. 1a, S2A).
The number of SVs shared by the three sampling domains (onshore air, offshore air, and surface water) was surprisingly small, only 14 SVs, whereas 348 SVs were shared between onshore air and offshore air (Fig. 1b). Surprisingly, 11,568 SVs were unique to onshore air samples, while 5,224 and 3,183 SVs were unique to offshore air and surface water samples, respectively.
Significant differences (ANOVA, p < 0.001) of AEC SVs were observed among onshore air, offshore air, and surface water communities (Fig. S3A) based on Non-metric multidimensional scaling (NMDS) and principal coordinates analysis (PCoA) (Fig. 1d). The community structure of airborne eukaryotes was also different between onshore and offshore air based on PCoA and NMDS (p < 0.002). Shannon diversity of AEC SVs differed significantly among sampling locations (ANOVA, p = 0.017), which were higher in the airborne samples collected at offshore and onshore regions than in Red Sea surface seawater samples (Table S2, Fig. 1c). The average alpha diversity explained 8% of the variance in community structure (ANOVA, p = 0.001).
The contribution of different taxa to AECs communities sampled along 14 months showed slight variations in the relative abundances among the four seasons. The most detected eukaryotic phyla in winter were Streptophyta (22.53%), Apicomplexa (20.86%), Basidiomycota (4.07%), Chlorophyta (2.76%), Chordata (2.42%), Ascomycota (1.85%), and Annelida (1.61%). In addition, the main classes driving those phyla were Conoidasida, Liliopsida, Spirotrichea, Colpodea, Chlorophyceae, Tremellomycetes, Dinophyceae, Mammalia (Cetacea), Polychaeta, Saccharomycetes, and Florideophyceae.
The spring airborne dust samples harbored high numbers of Streptophyta (19.23%), Apicomplexa (17.25%), Basidiomycota (13.07%), Chordata (6.50%), Ascomycota (5.66%), Arthropoda (1.47%), and Annelida (1.21%) represented by Conoidasida, Liliopsida, Spirotrichea, Colpodea, Chlorophyceae, Tremellomycetes, Dinophyceae, and Mammalia (Cetacea).
During the summer sampling period, the AEC was dominated by Streptophyta (48.87%), Annelida (3.28%), Ascomycota (2.07%), and Chordata (1.76%), Apicomplexa (1.21%) mainly belong to Conoidasida, Malasseziomycetes, Mammalia, Sordariomycetes, Liliopsida, Spirotrichea, Dinophyceae, Colpodea, Tremellomycetes, Saccharomycetes, and Polychaeta (Fig. 2a, S2B).
There was significant clustering (ANOVA, p < 0.001) differences among AEC SVS among seasons based on NMDS and PCoA (Fig. 2c, S3B) of Bray–Curtis distances among samples. The average alpha diversity of the AECs has significantly differed among seasons, which explained 8% of the variance in community structure (ANOVA, p = 0.001) (Fig. 2b, Table S2, ANOVA, p < 0.001). The richest and most diverse samples were those sampled in Spring 2016, with diversity being lowest in Fall 2016 (Fig. S4).
Most of the dust samples (69%) originated from air masses from the northwest, the dominant wind direction in the Red Sea (Bower and Farrar 2015; Eladawy et al. 2017). The contribution of different phyla to AECs suspended in air masses sampled from different trajectories was rather conserved among air masses. The most frequently detected taxa were Streptophyta in the samples that came from SE and E, Apicomplexa in N, NE, and S backward trajectories, Ascomycota in the N samples, Basidiomycota in the W atmospheric source, Chordata in the S, E, and NW backward trajectory, Chytridiomycota in S origins, Annelida in SW and W air sources, Bacillariophyta in N, W, and NW airborne dust samples, Chlorophyta in W and S air sources, and Arthropoda and Nematoda in NE and NW samples (Fig. 3a, S2C).
No significant clustering (ANOVA, p < 0.83) of AEC SVs between different backward air mass trajectories (Fig. 3c, S3C) based on NMDS and PCoA of Bray–Curtis distances among samples. The history of the atmospheric air transportation has no significant effect on the AEC diversity and richness (Fig. 3b, ANOVA, p = 0.76).
The relationship between dust load and AEC diversity indexes was tested using linear regression analysis (Fig. 4). There was no relationship between dust concentration and the AEC diversity and richness (R2 = 0.014, 0.001, p = 0.18 and 0.67, for Shannon and Chao1 index, respectively).
Characterization of atmospheric eukaryotic communities in the Red Sea region provide a representation of the taxa in the air and their relative abundance in regard to sampling domain, dust origins, and sampling seasons. The most abundant micro and macro eukaryotes in the Red Sea atmosphere are similar to those reported elsewhere, such as those typically reported in the atmosphere of Spain (Caliz et al. 2018), China (Liu et al. 2019), Japan (Tanaka et al. 2019) and Brazil (Womack et al. 2015), which may be due to the influence of marine and terrestrial environments as point sources for aerosolized micro and macro eukaryotes to the atmosphere.
In general, the AEC over the Red Sea was dominated by Streptophyta (Liliopsida (Poaceae)), a plants clade that release large amounts of pollen into the atmosphere transporting from arid terrestrial habitats (Bouchenak-Khelladi et al. 2014) that might cause allergic respiratory disease (García-Mozo 2017) in different regions and seasons based on changes in land uses and climate (García-Mozo et al. 2016). Streptophyta bioaerosols were mostly transported to the Red Sea region from the Arabian Peninsula in all seasons. However, Streptophyta was not in detectable quantity in the Red Sea surface water. Dust particles in the global dust belt originate from arid regions, and is, therefore, expected to contain significant loads of arid plant material, as demonstrated here for dust over the Red Sea, receiving inputs from the Sahara Desert and deserts in the Arabian Peninsula.
The interaction between surface waters and airborne dust in the Red Sea (Parajuli et al. 2020) could be the main reason for Dinophyceae being aerosolized into the atmosphere since this group dominates the phytoplankton community in the Red Sea (Touliabah et al. 2010). In addition, the aerosolized biological materials of Annelida (Polychaeta) over the Red Sea might also result from air-sea interactions (Reuscher 2016).
In the atmospheric dust, the high number of SVs belonging to Apicomplexa (Conoidasida), a parasitic alveolate that infects all mammals, some birds, some fish, some reptiles, and some amphibians in their gastrointestinal tract, might be of concern in regard to human and animal health exposed to these air masses (Oborník 2020). The Apicomplexa group has been recently reported in suspended dust in airplane cabins (Sun et al. 2020) and reported in the southeast Mediterranean atmosphere (Katra et al. 2014). Ascomycota (Sordariomycetes and Saccharomycetes) are large classes comprise some plant and mammalian pathogens and many members with a role in nutrient cycling (Llopis et al. 2014; Li et al. 2016). They are found in terrestrial, freshwater and marine habitats globally (Oborník 2020) and commonly identified in indoor and outdoor atmospheres (Adams et al. 2013). Transported mostly from the west, Basidiomycota (Malasseziomycetes and Tremellomycetes) consists of commensal and pathogenic organisms and spores that are associated with allergies in humans (Moelling and Broecker 2020). This fungal phylum is associated with humid ecosystems and releases the spores in the night when atmospheric humidity is relatively high (Elbert et al. 2007; Gusareva et al. 2019). Hence, the entire globe and especially the Red Sea atmosphere, characterized by high humidity, would be a suitable environment for Basidiomycota, particularly as atmospheric humidity increases as a result of the global warming (Raymond et al. 2020).
A number of Chordata-related SVs identified in the airborne dust represented by Primates, Cetacea and Aves, while represented by Appendicularia in the Red Sea surface water. Identifying aerosolized biological materials from Primates, Cetacea and Aves support the fact that micro-sized tissue fragments and debris originating from humans and animals could be airborne and transported through the atmospheric (Després et al. 2012). Detection of Appendicularia SVs is expected, as it is one of the most abundant zooplankton in the upper layers of the Red Sea water (Rasul and Stewart 2019).
Remarkably, a high number of eukaryotic SVs were unique to the onshore air samples compared to offshore air and surface water, which suggests a combination of both soil and marine-derived micro and macro eukaryotes. While AEC diversity is significantly differed per sample location and type (onshore air, offshore air, and surface water), the three sample types share similar phylotypes as a consequence of air-sea exchange processes (Mayol et al. 2014; Tesson et al. 2016). However, the higher diversity of terrestrial eukaryotes in air samples which can be transported over long distances (Mayol et al. 2017), indicates that some species have spores adapted to the harsh conditions characteristic of the airborne habitat (Aalismail et al. 2019).
A recent study reported that AEC is influenced more by environmental conditions compared to airborne prokaryote communities (Tanaka et al. 2019). Hence, the airborne eukaryotic richness and diversity in the Red Sea region also showed variability over seasonal conditions, consistent with the eukaryotes in the Red Sea (Pearman et al. 2017). The AEC was most diverse in spring, due to the increased loads resulting from pollen released by flowering plants (Ruiz-Valenzuela and Aguilera 2018), as the AEC is dominated by plant-related DNA. However, the independence of diversity and richness of the origin of the air masses sampled and dust concentrations may relate to the suspension period of AEC in the global dust belt, as eukaryotic cells can be suspended and transported freely through the atmosphere without attachment to dust particles.