A Multifaceted Systems Approach Identifies Polymicrobial Cause of AOD Stem Necrosis S. Denman J. Doonan, M. Broberg, S. Plummer, A. Griffiths, K. Scarlett, M. Kaczmarek, J. Foster, J.E. McDonald Programme 2 Annual Report-back Meeting January 2017, Roslin
Isolation and culture studies • Revealed novel bacterial species • Led to description of 2 new genera and 13 new species • Shift in the microbiome of diseased vs healthy trees • Enterobacteriaceae dominated diseased vs Pseudomonadaceae in healthy trees • 3 spp - high occurrence • G. quercinecans (Gram –ve, Enterobacteriaceae) - closely related to genus Serratia • B. goodwinii (Gram –ve, Enterobacteriaceae) - most closely related to B. rubrifaciens, – pathogen of walnut • Rahnella victoriana • Almost exclusively in symptomatic trees
Many organisms are not tractable to culture Confirmation of detection and abundance of Bg, R. vic, Gq in field sampled oak metagenomes vs extremely low presence or absence in healthy metagenomes Symptomatic trees (AOD lesions) Healthy trees Martin Broberg
Decline Diseases - concept 11/09/2017 7 Decline is the interaction of interchangeable, specifically ordered abiotic and biotic factors to produce a general deterioration, often ending in death of trees (Manion and Lachance, 1991) A Decline is a progressive deterioration, often ending in death, for which single exclusive causes have not been isolated (Sinclair, 1967) A Decline results from an interacting set of factors (Manion, 1981) Concept of Declines is still evolving (Manion, 1991) Decline: Arises from factors that interact sequentially or simultaneously and have a cumulative effect over time. Certain factors may have key roles. Predisposition/altered host condition probably essential.
Test for necrotic capability of organisms associated with the AOD lesions Koch’s Postulates Model: 1 pathogen = 1 disease 1. Isolation of the putative pathogen (the micro-organism occurs in every case of the disease) 2.The putative pathogen is not found in healthy organisms 3.Once in pure culture, the putative pathogen is reintroduced into a healthy host and causes disease anew 4.The micro-organism is re-isolated from the experimentally infected host Problems with the above model in polymicrobial, secondary disease pathosystems (Decline – diseases) 1. More than a single putative pathogen is isolated, microbiome effect not taken into account 2. The putative pathogen(s) may or may not be found in healthy hosts (at very low levels) 3. Re-introducing pathogen into a healthy host to create disease anew – Premise of Decline diseases is that the host is predisposed to infection Cause of stem necrosis
Microbiome Effects - Questions to address in disease development investigations - Gilbert et al. 2016: 1. Taxonomic composition of the microbiome • Have all participants been detected? • Can they be cultured? 2. Selection of species for testing • Consistently occurring micro-organisms on lesion margin (KP1&2) • Micro-organisms that occur in a benign state - transform behaviour under certain circumstances? 3. Testing • Relative abundance of different species • Sequence of presence and function of micro-organisms • Collective gene effect – could the same functional genes be supplied by more than one organism ? • Specific lineages or metabolic pathways 4. Host metabolome – the chemicals produced by the host • What does predisposition look like? • Effects of functional bacterial genes on host – host response? • Host response – what effect on microbial communities?
11/09/2017 10 Sequence whole genomes of isolated key putative pathogens Transcriptomes of field infected material and inoculated material Proteomes of field infected material and inoculated material
Genomics Hypothesis: Gq and Bg are necrotic pathogens of oak Aim: Identify genes or gene families causing pathogenesis and/or necrosis Pathogenicity genes: • Plant Cell-Wall Degrading Enzymes (PCWDEs) • Secretion systems • Oxidative stress • Iron acquisition • Quorum sensing • Effectors Virulence factors: Many eg: Iron acquisition enzyme James Doonan
NB Phloem NB Sapwood NB only Gq only x3 Bg only x3 Gq & Bg x3 Transcriptome analysis of Bg and Gq grown on oak tissue 1. How do Bg and Gq respond to growth on oak tissue? 2. Do Bg and Gq interact? • Growth on nutrient broth + sapwood or phloem • Analysis of gene expression
Axenic Gq Below Co-culture Bg+Gq Growth on oak tissue and in co-culture with Bg alters gene expression of Gq Sapwood Phloem 2h 6h Sapwood Phloem Axenic Bg 2h 6h
Field sampled oak microbiome analysis: functional composition (metagenomics) • 67-95% of predicted genes in AOD lesions were bacterial • 0.6-6% in non-symptomatic samples were bacterial • 627 genes common to all symptomatic samples (not in non-symptomatic trees) • Genes were associated with bacterial virulence and plant defence
Causality: Koch’s Postulates is the standard Overall aims were to test the necrogenic capability of the bacterial species singly and in combination, with or without the addition of Agrilus biguttatus Hypotheses 1. Enterobacterial species consistently isolated from AOD symptomatic oak can cause necrosis of oak stem tissue 2. Combinations of bacterial species cause more severe tissue necrosis (reflected in larger lesions), than individual species only, 3. The interaction between Agrilus larvae (derived from eggs) and bacteria, leads to the development of AOD symptoms. Ojectives 1. To determine necrogenic ability of bacterial species singly and in combination, 2. To determine necrogenic ability of Agrilus eggs on their own and in combination with bacteria, 3. To reproduce on logs, the symptoms and signs that characterise AOD.
Three experiments were carried out over 3 consecutive years as testing could only be done annually when beetle eggs were available. Two of the three trials were carried out in growth chambers and the third trial in the field, where young plantation oak (25 years old) were used instead of logs. Treatment No. Bacterial Treatments – Applied with and without Agrilus biguttatus eggs 1 Brenneria goodwinii 2 Gibbsiella quercinecans 3 Lonsdalea quercina britanica 4 Rahnella victoriana 5 Rahnella variigena 6 Water (control) 7 Egg only 8 Erwinia billingiae 9 B. goodwinii + G. quercinecans 10 B. goodwinii + G. quercinecans + L. quercina britainca 11 B. goodwinii + G. quercinecans + R. victoriana 12 B. goodwinii + G. quercinecans + R. variigena 13 B. goodwinii + G. quercinecans + L. quercina britainca + R. victoriana + R. variigena
11/09/2017 22 0 10 20 30 40 50 60 70 80 90 100 0 50 100 150 200 250 300 Percentage data points out of total uncontaminated inoculation points used (%) Mean Lesion Area mm^2 Treatment Pathogenicity tests: bacteria only and bacteria plus egg treatments lesion area and % usable data points Area % data points used p=0.08 p<0.007 • Agrilus increased lesion size and bacteria were re-isolated along gallery length – spread in tree • Significantly larger lesions with Bg+Gq and with Gq+egg; Bg+egg; Bg=Gq+egg and Lqb+egg • Co-inoculation of Bg+Gq greatly increased back isolation of Bg from 17% to 66% - synergistic effect
Microbiome Effects - Questions to address in disease development investigations - Gilbert et al. 2016: 1. Taxonomic composition of the microbiome • Have all participants been detected? • Can they be cultured? 2. Selection of species for testing • Consistently occurring micro-organisms on lesion margin (KP1&2) • Micro-organisms that occur in a benign state - transform behaviour under certain circumstances ? • Relative abundance of different species * • Sequence of presence and function of micro-organisms * • Collective gene effect – could the same functional genes be supplied by more than one organism ? • Specific lineages or metabolic pathways 3. Host metabolome – the set of chemicals produced by the host • What does predisposition look like? * • Effects of functional bacterial genes on host – host response? * • Host response – what effect on microbial communities? *
Conclusions • Bg, Gq and Rv are consistently detected as abundant members of the cultivable and ‘omics’-based AOD microbiome. • Integrated multi-omics of AOD lesions revealed the dominance of a bacterial community and concomitant detection of host defence responses. • Bg, Gq and Rv have genome-encoded virulence genes found in canonical plant pathogens belonging to the Enterobacteriaceae • Genes and transcripts of Bg, Gq and Rv are abundant in the lesion microbiome and experiments confirm functionality. • Log inoculation tests confirm the ability of Bg and Gq to produce lesions and demonstrate synergistic effects. • Microbiome analysis of AOD lesions and artificial inoculation has provided compelling evidence of a polymicrobial disease with by Bg, Gq and Rv as key agents of disease. • We have adopted a systems approach to analysis of the stem bleeds and have developed a template for analysis of complex disease symptoms, which is still likely to develop especially with regard to taking account of host predisposition.
Thank you to our funders for their generous support and for all the technical support from TSU as well as help from landowners and managers Working hard to future-proof the resilience of oak
forestresearch
matsui_528
muraoka7
yasumuusan
kosuken
smatsumori
miyuoba
kuribayashi4
gmoriki
satotaichi
yosukefk
yukizenimoto
rasmusluckow
reverentgeek
csswizardry
tmm1
bkeepers
eileencodes
revolveconf
aki_iinuma
brettharned
dougneiner
phodgson
jacobian