Household arthropod- associated microbiota
We studied the microbiota composition associated with the arthropods
captures and measured how they varied by arthropod’s family, season
(month of sampling) and environments (urban/suburban).
The microbiota identified with the culture-based approach significantly
differed among the arthropod families, and between environments, whereas
it did not vary by season (Table 1A; Fig. 3A). We cultured and isolated
bacteria from 84 pools (72.6%), whereas the remaining 31 pools (27.4%)
did not harbour any culturable bacteria. Among the bacteria that could
be cultured and taxonomically identified, Gram-positive cocci (including
those strains not identifiable at family level) and Gram-positive rods
were the most abundant (Fig. 3A, Table S2), followed byStaphylococcaceae (30.36%), Bacillaceae (10.71%) andMicrococcaceae (10.12%) (Fig. 3A, Table S2).Streptococcaceae and Enterococcaceae accounted for 9.52%
and 8.33% of the taxonomically assigned sequences, respectively (Fig.
3A; Table S2). The remainder of the taxonomically assigned reads
belonged to: Moraxellaceae (4.76%), Dysgonomonandaceae(4.17%), Rhodobacteraceae (4.17%), Yersiniaceae(4.17%), Erwiniaceae (2.98%), Carnobacteriaceae(1.79%), Clostridiaceae (1.79%), Brevibacteraceae(1.19%) and Microbacteriaceae (1.19%) (Fig. 3A, Table S2). Rare
families accounting for 0.6% each of the total identified sequences
were: Peptostreptococcaceae , Caryophanaceae ,Corynebacteriaceae , Dermabacteraceae ,Enterobacteriaceae , Neisseriaceae ,Pseudomonandaceae and Veillonellaceae (Table S2). A total
of 192 isolates could not be identified at family level; they were
Gram-positive cocci (49.48% of the unassigned sequences), Gram-positive
rods (27.6%), Gram-negative rods (15.10%), Gram-negative cocci
(6.77%), Gram-negative coccobacilli (0.52%) and Gram-negative
filamentous bacteria (0.52%) (Table S2).
Gram-positive cocci and Gram-positive rods were found in all arthropods,
except for Vespidae (Gram-positive cocci) and Blattidae (Gram-positive
rods) (Fig. S1A). Even if few bacteria families were shared across the
artropod’s families, each arthropod had a unique microbiota composition
(Fig. S1A).
The microbiota identified with the metabarcoding approach significantly
differed among arthropod families, season (sampling months) and
environments (urban/suburban) (Fig. 3B; Table 1). The interaction term
Arthropod Family*Season was also significant (Table 1). The
metabarcoding approach identified 771 ASVs, taxonomically assigned to 15
phyla, 111 families and 249 genera of bacteria. Among the 10 most
abundant families, Rickettsiaceae comprised 67.3 % of the
identified families, followed by Anaplasmataceae (10.59%),
unidentified family of the class Bacilli (3.59%),Proprionibacteriaceae (3.71%), Porphyromonadaceae(3.29%), Micrococcaceae (2.89%), Enterobacteriaceae(2.89%), Enterococcaceae (2.12%), Methylobacteriaceae(1.83%) and an unidentified family of the order Burkholderiales(1.78%) (Fig. 3B; Table S2). The metabarcoding approach revealed thatRickettsiaceae were common and abundant in the microbiota of all
arthropod families, followed by Anaplasmataceae (Fig. S1B; Table
S2). Bacilli were abundant in Tipulidae ; they were present
in low abundance in 50% of the captured arthropods (Fig. S1B). Although
the microbiota of the different arthropod families was dominated byRickettsiaceae , the relative abundance of other bacterial
families was distinct among the arthropods (Fig. S1B, Table S2). For
example, Enterobacteriaceae and Micrococcaceae made up
15.2 and 13.9%, respectively, of the Agelenidae microbiota, whereasEnterococcaceae (14.4%) were the second most abundant bacteria
family in Blattidae (Fig. S1B; Table S2). The microbiota ofCalliphoridae was dominated by Porphyromonandaceae(70.7%), whereas Rickettsiaceae (8.3%) accounted for a small
proportion of this microbiota (Fig. S1B).
We studied the change in community composition of the exogenous and
endogenous microbiota separately in the arthropod captures between July
and October 2019. We asked if changes in the endogenous and exogenous
microbiota were explained by the arthropod family, the month of
collection and their interaction terms. We also studied the correlation
between the exogenous and endogenous microbiota communities with
environmental variables. In the culture-based approach, the endogenous
and exogenous microbial communities varied significantly by the
arthropod families, whereas they did not differ significantly among
seasons (month of collection; Fig. 4A; Table 2). Gram-positive cocci
were abundant in both the endogenous and exogenous microbiota (Fig. 4A).
Some bacteria families were only found in the exogenous microbiota
(Staphylococcaceae , Peptostreptococcaceae ,Bacillaceae , Enterococcaceae , Rhodobacteraceae andErwiniaceae were isolated from the exoskeleton of the
arthropods), whereas others were unique to the endogenous microbiota
(Enterococcaceae , Erwiniaeae , Rhodobacteraceae ,Micrococcaceae , Carnobacteriaceae andStreptococcaceae ) (Fig. 4A). Some bacterial families were
arthropod-specific: Staphylococcaceae were only found in the
exogenous microbiota of Coccinellidae and Blattidae ;Rhodobacteraceae were only found in the endogenous microbiota of
Blattidae and the exogenous microbiota of Gnaphosidae (Fig. 4A).Carnobacteriaceae and Streptococcaceae were only found in
the Coccinellidae ’s endogenous microbiota (Fig. 4A).
The endogenous and exogenous microbiota identified with the
metabarcoding approach were overall divergent and varied significantly
by the arthropod family, whereas they did not differ significantly by
the sampling month (Table 2). Rickettsiaceae were both common and
abundant across the arthropods’ families (Fig. 4B). The exogenous
microbiota was dominated by Rickettsiaceae, comprising 64.1% of the
microbiota in Blattidae , 30.2% in Coccinellidae , 55.8%
in Gnaphosidae spiders, 76.3% in Pholcidae and 34.4% of inTipulidae . The exogenous microbiota of Coccinellidae andTipulidae were the most diverse among the arthropods studied
here. In addition to Rickettsiaceae , Coccinellidaeexogenous microbiota included Propionibacteriaceae (18.1%),Methylobacteriaceae (13.7%), Micrococcaceae (11.4%), andStreptococcaceae (9.4%) in relatively high abundance (Fig. 4B).
The exogenous microbiota of Tipulidae comprisedEnterobacteriaceae (23.7%), Bacilli (16.9%),Planococcaceae (6.1%), Streptococcaceae (5%), in
addition to Rickettsiaceae (Fig. 4B). The endogenous microbiota
composition was arthropod-specific. For example, Enterococcaceaemade up 94% of the endogenous microbiota of Blattidae , whereas
they were in low abundance (<1%) in other arthropods (Fig.
4B). Bacilli comprised 93% of the endogenous microbiota inTipulidae and 42% in Coccinellidae (Fig. 4B).Anaplasmataceae were present in low abundance in the endogenous
microbiota of all arthropods and abundant (35%) in Gnaphosidae(Fig. 4B).
We used the Spearman correlation combined with the Shannon-Wiener index,
which combines species richness and relative abundances, to establish
associations between endogenous/exogenous bacterial communities and
environmental variables: temperature, precipitation, humidity, and wind
speed. Neither the endogenous nor the exogenous microbiota compositional
changes were explained by significant covariation with the environmental
variables (Table S3).