Endosymbionts associated with the whitefly Bemisia tabaci cryptic species are known to contribute to host fitness and environmental adaptation. The genetic diversity and population complexity were investigated for endosymbiont communities of B. tabaci occupying different micro-environments in Pakistan. Mitotypes of B. tabaci were identified by comparative sequence analysis of the mitochondria cytochrome oxidase I (mtCOI) gene sequence. Whitefly mitotypes belonged to the Asia II- 1, -5, and -7 mitotypes of the Asia II major clade. The whitefly-endosymbiont communities were characterized based on 16S ribosomal RNA Operational Taxonomic Unit (OTU) assignments, resulting in 43 OTUs. Most of the OTUs occurred in the Asia II-1 and II-7 mitotypes (r2=0.9, p<0.005) while the Asia II-5 microbiome was less complex. The microbiome OTU groups were mitotype-specific, clustering with a basis in phylogeographical distribution and the corresponding ecological niche of their whitefly host, suggesting mitotype-microbiome co-adaptation. The primary endosymbiont Portiera was represented by a single, highly homologous OTU (0-0.67% divergence). Two of six Arsenophonus OTUs were uniquely associated with Asia II-5 and -7, one occurred exclusively in Asia II-1, two only in Asia II-5, and one in both Asia II-1 and -7. Four other secondary endosymbionts, Cardinium, Hemipteriphilus, Rickettsia, and Wolbachia OTUs were found at ≤29% frequencies. The most prevalent Arsenophonus OTU was found in all three Asia II mitotypes (55% frequency), whereas the same strain of Cardinium and Wolbachia were found in both Asia II-1 and -5, and a single Hemipteriphilus OTU occurred in Asia II-1 and -7. This pattern is indicative of horizontal transfer, suggestive of a proximity between mitotypes sufficient for gene flow at overlapping mitotype ecological niches.

The 16S rDNA (~1500 bp) sequence was PCR-amplified using ‘universal’ primers 27F (5'-AGAGTTTGATCMTGGCTCAG) and 1513R (5'-ACGGYTACCTTGTTACGACTT) (0.4 μM) [64, 65]. The PCRs were set as explained above using 20ng of whitefly DNA. Cycling parameters were an initial denaturation at 94°C for 2 min, and 35 cycles of 94°C for 30 s, 55°C for 30 s, and 72°C for 2 min, with a final extension at 72°C for 10 min. Amplicon sizes were verified by agarose gel electrophoresis and cloned. Twelve clones per whitefly were selected by colony PCR amplification, and sequenced. 

Figure S1. Bayesian phylogeny of the Bemisia tabaci mitotypes from Pakistan based on the 3’-end fragment of the mitochondrial cytochrome oxidase I gene (COI) (725 bp). Acronyms are lineages designations by De Moya et al. (2019). Numbers at nodes are posterior probabilities.

Figure S2. Distribution of a Portiera, b Arsenophonus, c Cardinium, d Hemipteriphilus, e Rickettsia, and f Wolbachia in twelve districts of Pakistan. Bars represent the relative abundance of endosymbionts per district. Bemisia tabaci mitotypes Asia II-5 and -7 were only recorded from Lahore, whereas mitotype Asia II-1 was widely distributed across cotton growing districts. Blue lines in map represent rivers, dotted lines indicate provincial limits and the polygon shows the cotton growing area of Pakistan.

Figure S3. Rarefaction curves for individual whitefly samples. Red lines are whitefly samples belonging to Asia II-7, black lines are Asia II-1 individuals and blue are whitefly samples from the Asia II-5 group. The curves were generated in mothur v1.43.0 based on the observed richness with 1000 iterations.

Table S1. Collection information, phylogenetic affiliation and endosymbiotic composition of individual whitefly samples from Pakistan. The presence and absence of the endosymbiont is given by the + and – signs, respectively. Capital letters were used to denote the endosymbiont genus as follows:  P= Portiera, A= Arsenophonus, C= Cardinium, H= Hemipteriphilus, R= Rickettsia, W= Wolbachia

Table S2. Taxonomic classification, accession numbers of 16S ribosomal RNA (16S rRNA) sequences and host mitochondrial cytochrome oxidase I (mtCOI), sample location, whitefly host plant, and relative frequency per mitotype of operational taxonomic units (OTUs) detected in whitefly samples from Pakistan.

Table S3. Absolute frequencies of bacterial classes and families based on 1084 bp the 16S ribosomal RNA gene (16S rRNA), for the three Bemisia tabaci mitotypes identified in Pakistan, Asia II-1, -5, and -7. The 16S rDNA sequences were taxonomically classified using the SILVA taxonomy database. Approximately 90% of the sequences clustered within the Bacteroidaceae, Enterobacteriaceae, and Halomonadaceae families. 

Table S4. The OTU abundance table per whitefly sample, taxonomic classification and whitefly mitotypic group designation.