Molecular Ecology (2014) 23, 1433–1444

doi: 10.1111/mec.12637

SPECIAL ISSUE: NATURE’S MICROBIOME

The bacterial communities in plant phloem-sap-feeding insects XIANGFENG JING,* ADAM C-N WONG,*1 JOHN M. CHASTON,* JOHN COLVIN,† C I N D Y L . M C K E N Z I E ‡ and A N G E L A E . D O U G L A S * § *Department of Entomology, Comstock Hall, Cornell University, Ithaca, NY 14853, USA, †Agriculture, Health and Environment Group, Natural Resources Institute, University of Greenwich, Kent, UK, ‡ASDA-ARS, U.S. Horticultural Research Laboratory, 2001 South Rock Road, Fort Pierce, FL 34945, USA, §Department of Molecular Biology and Genetics, 526 Campus Road, Cornell University, Ithaca, NY 14853, USA

Abstract The resident microbiota of animals represents an important contribution to the global microbial diversity, but it is poorly known in many animals. This study investigated the bacterial diversity in plant phloem-sap-feeding whiteflies, aphids and psyllids by pyrosequencing bacterial 16S rRNA gene amplicons. After correction for sequencing error, just 3–7 bacterial operational taxonomic units were recovered from each insect sample sequenced to sufficient depth for saturation of rarefaction curves. Most samples were dominated by primary and secondary symbionts, which are localized to insect cells or the body cavity, indicative of a dearth of bacterial colonists of the gut lumen. Diversity indices of the bacterial communities (Shannon’s index: 0.40–1.46, Simpson’s index: 0.15–0.74) did not differ significantly between laboratory and field samples of the phloem-feeding insects, but were significantly lower than in drosophilid flies quantified by the same methods. Both the low bacterial content of the phloem sap diet and biological processes in the insect may contribute to the apparently low bacterial diversity in these phloem-feeding insects. Keywords: Acyrthosiphon pisum, bacterial diversity, Bactericera cockerelli, Bemisia tabaci, Diaphorina citri, microbiota, phloem sap, secondary symbiont, symbiosis Received 22 April 2013; revision received 6 December 2013; accepted 13 December 2013

Introduction Animals provide multiple habitats for microorganisms. Microorganisms regularly colonize animal surfaces (skin, cuticle, gut, etc.), and the body cavity, body fluids and cells of some animals bear dense microbial populations without apparent ill-effect (Buchner 1965; Douglas 2010; Human Microbiome Project 2012a; Findley et al. 2013). The diversity of the microbiota varies widely with both location, for example generally lower diversity in cells than in the gut lumen, and animal phylogeny, for example generally higher diversity in vertebrates than invertebrates (McFall-Ngai 2007; Bright & Bulgheresi 2010). Some animal-associated communiCorrespondence: Angela E. Douglas, Fax: 1 607 255 0939; E-mail: [email protected] 1 Present address: School of Biological Sciences, University of Sydney, Sydney, NSW 2006, Australia © 2013 John Wiley & Sons Ltd

ties are of exceptionally high species-richness, for example often exceeding 1000 taxa in the gut of mammals and some termites (Warnecke et al. 2007; Human Microbiome Project 2012a). It can also vary with the age, sex, physiological condition, and genotype of the animal host, as well as environmental circumstance, including temperature and diet (Turnbaugh et al. 2009; Claesson et al. 2012; Wernegreen 2012; Yatsunenko et al. 2012; Franzenburg et al. 2013; Hildebrand et al. 2013). These broad conclusions come from studies of many animalassociated microbiota, with the most systematic analyses conducted on humans and laboratory mice (Faith et al. 2010; Human Microbiome Project 2012b). This study concerns the bacterial diversity in plant phloem sap feeders. The only animals that feed on plant sap through the life cycle are insects of the order Hemiptera (Dolling 1991; Douglas 2003, 2006), and this habit is correlated absolutely with the possession of microorganisms, usually bacteria that are restricted to

1434 X . J I N G E T A L . specialized insect cells and are obligately vertically transmitted (Buchner 1965). These bacteria are known as primary symbionts. Many phloem-sap-feeding insects also bear one to several other bacteria, informally called secondary symbionts, which may be localized to the bacteriocytes, other insect cells or the body cavity, but not the gut lumen, and are capable of both vertical and horizontal transmission (Buchner 1965; Degnan et al. 2010; Ferrari & Vavre 2011). The primary and secondary symbionts account for most or all of the 16S rRNA gene amplicons obtained by low-resolution methods, for example sequencing of cloned amplicons, TRFLP (Haynes et al. 2003; Ferrari et al. 2012; Singh et al. 2012), and these symbionts also dominated the sequence reads in recent pyrosequencing analyses of aphids (Jones et al. 2011; Russell et al. 2013). The specific purpose of this study was to quantify the bacterial diversity in phloem-feeding insects, focusing on the whitefly Bemisia tabaci, the pea aphid Acyrthosiphon pisum and two psyllid species Diaphorina citri and Bactericera cockerelli. The B. tabaci samples comprised representatives of seven species. [Bemisia tabaci is a morphological species that comprises >30 partially or completely reproductively isolated candidate species (De Barro et al. 2011; Liu et al. 2012).] Following evidence that the microbial diversity associated with some insects is lower in laboratory samples than field samples (Kuzina et al. 2001; Behar et al. 2008; Kounatidis et al. 2009; Chandler et al. 2011; Staubach et al. 2013; Wong et al. 2013), both laboratory and field samples of the phloemfeeding insects were included in the analysis. Our study used pyrosequencing of 16S rRNA gene fragments amplified with general bacterial primers, with complementary quantification by qPCR with taxon-specific primers, and was facilitated by extensive published data sets on the primary and secondary symbionts in these taxa (e.g. Gottlieb et al. 2006; Nakabachi et al. 2006, 2013; Oliver et al. 2010; Sloan & Moran 2012; Bing et al. 2013). Our specific questions were as follows: first, are the bacterial communities in phloem-feeding insects of low diversity, dominated by the primary and secondary symbionts? And, second, does the diversity of the bacterial communities differ between laboratory stocks and field populations of phloem-feeding insects?

Methods The experimental material The insect samples comprised multiple laboratory stocks and field populations of Bemisia tabaci species (Liu et al. 2012), two laboratory stocks of the potato psyllid Bactericera cockerelli, one of which was infected by the plant pathogen Candidatus Liberibacter

solanacearum (Liefting et al. 2009), and a laboratory stock and field population of each of the psyllid Diaphorina citri and the pea aphid Acyrthosiphon pisum (Table 1). The laboratory stocks of B. tabaci were reared at 27  1 °C with a 14L:10D photoperiod at The Natural Resources Institute, UK. The laboratory stocks of B. cockerelli were reared on eggplant or tomato at 26.8 °C, D. citri on Citrus macrophylla at 27 °C, and A. pisum on Vicia faba cv. Windsor at 20 °C with 16L:8D. All the samples were preserved in 90% ethanol prior to molecular analysis. The names of the B. tabaci species used in this study are trinomials, for example B. tabaci MEAM1, following the recommendation of Tay et al. (2012). The identity of every B. tabaci sample was confirmed to species level (De Barro et al. 2011; Liu et al. 2012) by sequencing the mitochondria cytochrome c oxidase I (mtCOI) gene (Shatters et al. 2009).

DNA isolation The insect samples, each comprising 30–40 adult whiteflies, 10 adult psyllids or six adult aphids, were rinsed three times in extraction buffer (20 mM Tris–HCl pH 8.0, 2 mM sodium EDTA, 1.2% Tritonâ X-100), then hand-homogenized in extraction buffer containing 20 mg lysozyme m/L and incubated at 37 °C for 1.5 h to achieve DNA extraction from both Gram-positive and Gram-negative bacteria (Wong et al. 2011 and A. C-N. Wong, unpublished data). The DNA in the samples was then extracted using the DNeasy Blood and Tissue Kit (Qiagen, Valencia, CA, USA) following the manufacturer’s instructions for Gram-positive bacteria. The quantity and quality of the DNA were measured with a Nanodrop 2000 (Thermo Scientific).

Multiplex 454 pyrosequencing of 16S rRNA gene sequences 16S ribosomal RNA amplicons of the V6–V7 region were prepared using general 16S rRNA gene primers 907Fmod (5′-AAACTCAAADGAATTGACGG-3′) modified from Sundquist et al. (2007) and 1237R (5′-GTAGYACGYGTGTWGCCC-3′; Turner et al. 1999). The 907Fmod primer was designed, following in silico analysis indicating poor predicted amplification of the 16S rDNA of Portiera, the primary symbiont of whiteflies, with the 907F primer, and also with a widely used general bacterial primer pair 27F-338R (V2 region). A pilot analysis confirmed that the 907Fmod-1237R primers and 27F-338R primers yielded comparable results in a test insect-microbiota sample that does not include Portiera (Table S1, Supporting information). Each sample-specific 907Fmod primer was designed to bear a multiplex identifier sequence (Table S2, Supporting © 2013 John Wiley & Sons Ltd

L O W D I V E R S I T Y M I C R O B I O T A I N P H L O E M - F E E D I N G I N S E C T S 1435 Table 1 Insect species and experimental samples Insect

Host plant

Insect original site (collecting date)

(A) Laboratory cultures of Bemisia tabaci species reared at Natural Resources Institute, University of Greenwich, UK Asia1 Egg plant India (2005) AsiaII-5 Cassava Kerala, India (2005) AsiaII-7 Cotton Hangzhou, China (2010) Australia Wild poinsettia Australia (2010) China1 Cotton Hangzhou, China (2010) MEAM1 Cotton/Brassica Peru (2009) Mediterranean Pepper Montpellier, France (2009) (B) Field populations of B. tabaci species collected at different countries Asia1* Eggplant cv. India Brinjal Raichur, India (9 February 2011) MEAM1_1* Eggplant cv. Hangqie 1 Hangzhou, China (9 June 2013) MEAM1_2* Tomato cv. BHN110 Hangzhou, China (9 June 2013) MEAM1_3* Cotton Pisco valley, Peru (4 March 2010) MEAM1_4* Paprika Pisco valley, Peru (4 March 2010) MEAM1_5* Tomato Pisco valley, Peru (4 March 2010) MEAM1_6* Hibiscus Florida, USA (18 May 2013) MEAM1_7* Tomato Florida, USA (25 June 2012) Mediterranean_1* Poinsettia Florida, USA (25 June 2012) Mediterranean_2* Sweet potato Namulonge and Mukono, Uganda (11 June 2012) (C) Psyllids Bactericera cockerelli† Eggplant cv. special Hibush Texas, USA (2010) B. cockerelli Tomato cv. Florida Lana Texas, USA (2010) Diaphorina citri Orange Jasmine Florida, USA (2000) D. citri* Orange Jasmine Florida, USA (2000) (D) Aphids Acyrthosiphon pisum Faba bean cv. Windsor New York, USA (June 2009) A. pisum* Pea cv. Sweet Ann New York, USA (13 June 2013) *Insects were collected in the field; otherwise, they were laboratory stocks. † Potato psyllid culture containing Candidatus Liberibacter solanaceum plant pathogen.

information). PCRs were conducted as previously described (Wong et al. 2011). Briefly, equal amounts (ng) of the products of three PCRs per sample were mixed, purified using the QIAquick PCR purification kit (QIAGEN) and quantified by the Quant-iTTM PicoGreenâ Kit. Emulsion PCR was conducted at 1.5 copies per bead using only ‘A’ beads for unidirectional 454 GS-FLX pyrosequencing with standard Titanium chemistry. All analyses included the pooled ethanol in which samples had been stored and two negative controls, specifically, extraction buffer as a control for contamination during sample preparation for sequencing, and the pooled extraction buffer used to wash the samples for contaminants associated with the insect samples. Pyrosequencing flowgrams were converted to sequence reads using 454 LIFE SCIENCE software (www. 454.com). Reads with ambiguous nucleotides (N) and 97% sequence identity to Rickettsia sp. described previously in B. tabaci MEAM1 (Gottlieb et al. 2006). The second major Rickettsia OTU, OTU11 (with related minor OTUs 3, 44 and 49), was detected in B. tabaci China1, Asia1 and AsiaII7, originating from China and India (Fig. 1A). This Rickettsia is allied to Rickettsia strain RI1 identified in B. tabaci Asia II-India, collected from India (Singh et al. 2012). The probability that sequencing error accounts for the sequence difference between the two major Rickettsia OTUs (Rickettsia1 OTU26 and Rickettsia2 OTU11) is 4.7 9 10 10, which is 4–5 orders of magnitude lower than the critical probability for this analysis. Parallel analysis of the 17 field samples from three species of B. tabaci identified 29 OTUs (Table S3B, Supporting information). Most of the bacterial taxa comprised a single major OTU and one or more minor OTUs (Table S4B, Supporting information), and sequencing error could account for all the minor OTUs. The top BLAST © 2013 John Wiley & Sons Ltd

L O W D I V E R S I T Y M I C R O B I O T A I N P H L O E M - F E E D I N G I N S E C T S 1437 Table 2 Pyrosequencing analysis of 16S rRNA gene amplicons of Bemisia tabaci with diversity indices Males

Females Diversity indices

Sample

No. reads

(A) Laboratory samples Australia 2568 Asia1 5564 China1 2120 AsiaII-5 2822 AsiaII-7 4869 MEAM1 6700 Mediterranean 5382 (B) Field populations Asia1 6006 MEAM1_1 15 137 MEAM1_2 13 152 MEAM1_3 — MEAM1_4 — MEAM1_5 — MEAM1_6 13 554 MEAM1_7 9463 Mediterranean_1 19 748 Mediterranean_2 16 035

Diversity indices

No. OTUs*

Shannon index

Simpson index

No. reads

3 4 3 3 6 3 3

0.65 1.14 0.97 0.36 1.40 0.54 1.02

0.36 0.65 0.58 0.17 0.73 0.28 0.61

5328 8835 4418 3200 11 150 13 982 15 196

6 3 3 — — — 3 3 5 5

1.20 0.84 0.92 — — — 1.07 1.01 0.70 1.21

0.65 0.53 0.55 — — — 0.65 0.61 0.49 0.65

19 286 26 350 26 449 9069 11 119 3780 16 568 14 011 21 396 23 651

No. OTUs*

3 4 3 3 5 3 3 5 3 3 5 4 4 3 3 5 16/4

Shannon index

Simpson index

0.61 0.92 0.90 0.49 1.46 0.81 0.83

0.36 0.53 0.53 0.25 0.74 0.46 0.48

0.73 0.77 0.51 1.06 1.36 0.68 0.99 0.68 0.67 0.96

0.37 0.43 0.27 0.60 0.74 0.37 0.60 0.43 0.45 0.50

*Operational taxonomic units (OTUs) were defined with pairwise 97% sequence identity after correction for sequencing error.

hit for the various symbionts was the same for all laboratory and field samples: NCBI Accession noCP003868 for Portiera, JN896335 for Arsenophonus, JQ009300 for Hamiltonella, DQ077707 for Rickettsia1, JN204501 for Rickettsia2 and DQ402518 for Wolbachia.

Bacterial communities in different species of Bemisia tabaci The bacterial communities in B. tabaci appeared to be of low diversity, as indicated by the few OTUs (3–6 per sample) and low values for the Shannon index (0.36– 1.46) and Simpson index (0.15–0.74; Table 2). The laboratory and field samples did not differ significantly for number of OTUs (Mann–Whitney U-test for males: W = 56.5; and for females: W = 101, both P > 0.05) or diversity indices (ANOVA of Shannon index: sample type: F1,27 = 0.28, sex F1,27 = 0.62, interaction F1,27 = 0.49; Simpson index: sample type: F1,27 = 0.92, sex F1,27 = 1.18, interaction F1,27 = 1.01; all P > 0.05). The primary symbiont Portiera was detected in every sample of B. tabaci tested and accounted for between 5% (in AsiaII-7 males) and 86% (in AsiaII-5 females) of the total reads per sample for the laboratory stocks, and 17.2% (MEAM1_5 females) and 84.7% (MEAM1_2 females) for the field samples (Fig. 1, Table S3A,B, Supporting information). To assess whether this variation © 2013 John Wiley & Sons Ltd

reflects biological variation in the abundance of Portiera or is merely a consequence of variation in the abundance of other bacteria in the pyrosequenced samples, the total abundance of Portiera in the laboratory samples was quantified by qPCR of 16S rRNA gene amplicons using Portiera-specific primers (Table S5, Supporting information with raw data in Table S6, Supporting information). The two indices of Portiera abundance are significantly positively correlated (Spearman’s rank correlation: rs = 0.679, P = 0.008), confirming the reliability of % reads as an index of Portiera abundance. For 20 of the 22 B. tabaci that included samples of both sexes, Portiera contributed a higher percentage of reads in the sample of female than male insects, and overall, the difference between the two sexes was statistically significant (paired t = 3.42, P = 0.005). The male and female samples of all seven species of B. tabaci in the laboratory bore between two and four secondary symbionts, in addition to Portiera. The laboratory and field samples of B. tabaci Asia1 bore the same symbionts (Fig. 1A,B), but the bacteria in B. tabaci MEAM1 and Mediterranean varied with source. The field samples of MEAM1 on cotton and paprika crops from Peru yielded 16S sequences with high identity to previously described c-proteobacterial secondary symbionts in lachnid aphids (Burke et al. 2009), in addition to Hamiltonella and Rickettsia1 present in other MEAM1 isolates. The secondary

1438 X . J I N G E T A L . A Australia

Asia1

China1

AsiaII-7

AsiaII-5

MEAM1 Mediterranean

Arsenophonus Cardinium Hamiltonella Portiera Rickettsia 1 Rickettsia 2 Wolbachia

B

Asia1

MEAM1

Mediterranean

Arsenophonus Cluster L symbiont Hamiltonella

Fig. 1 Relative abundance of pyrosequence reads assigned to bacterial symbionts of Bemisia tabaci. (A) Laboratory samples with B. tabaci species (Australia, Asia1, etc.) shown. (B) Field samples, with B. tabaci species (Asia1, etc.), and country of origin (India, etc.) and collection plant (E: eggplant, T: tomato, C: cotton, P: paprika, H: hibiscus, Po: poinsettia, SP: sweet potato) shown. The Rickettsia-1 and Rickettsia-2 have high sequence ID with Rickettsia described by Gottlieb et al. (2006) and Singh et al. (2012), respectively. Circle area corresponds to relative abundance of bacteria in each sample (see scale), with alternate black and gray colours to facilitate discrimination between rows; operational taxonomic units (OTUs) accounting for

The bacterial communities in plant phloem-sap-feeding insects.

The resident microbiota of animals represents an important contribution to the global microbial diversity, but it is poorly known in many animals. Thi...
291KB Sizes 0 Downloads 0 Views