Genetic Diversity, Host Relationships, and Bioclimatic Modeling to Predict Potential Global Distribution of the Myrtle Rust Pathogen Ned B

Genetic diversity, host relationships, and bioclimatic modeling to predict potential global distribution of the myrtle rust pathogen Ned B. Klopfenstein USDA Forest Service, Rocky Mountain Research Station, Moscow, Idaho U.S.A J.E. Stewart, M.-S. Kim, P.G. Cannon, A.L. Ross-Davis, J.W. Hanna, E.W.I. Pitman, R.N. Graҫa, A.C. Alfenas, T.L. Peever, J.Y. Uchida, R.D. Hauff, C.Y. Kadooka, S. Namba, S. Simeto, C.A. Pérez, M.B. Rayamajhi, D.J. Lodge, M. Arguedas, R. Medel-Ortiz, M.A. López-Ramirez, P. Tennant, M. Glen, P. da S. Machado, A.R. McTaggart, and A.J. Carnegie

Myrtle Rust Symposium, New Zealand Invasive Species Working Group, Better Border Biosecurity (B3), New Zealand Institute for Plant and Food Research Ltd., 28 August 2017, Auckland, New Zealand Austropuccina psidii

• Primary example of an emerging forest disease caused by rust pathogen • Biotrophic rust fungus • Infects young, actively growing foliage, floral buds, and young fruits of host species in the Myrtaceae • Unusually wide host range (450+ species; 33+ genera) • Brown lesions with masses of yellow or orange urediniospores; dark brown teliospores; purpling with age • Decreased growth; loss of apical dominance • Native to South and Central America? • Several races or biotypes Life Cycle

Glen et al., 2007 Austropuccinia psidii - Host range examples Angophora Callistemon Corymbia Eucalyptus Eugenia Heteropixis Marlierea Melaleuca Metrosideros Myrcia Myrciaria Pimenta Psidium

Syzigium From: A.C. Alfenas Threats of myrtle rust

• Susceptible species are often dominant components of flora in Oceania, Southeast Asia, South and Central America, and southern Africa • outbreaks  change structure, composition, and function of http://www.hear.org/species/puccinia_psidii/ forests

• >700 species of Eucalyptus, mostly native to Australia • most widely planted genus

http://en.wikipedia.org/wiki/Eucalyptus Reports of Austropuccinia psidii occurrence

Brazil 1884: guava First report Reports of Austropuccinia psidii occurrence

Jamaica 1934: Puerto allspice, Rico rose apple, 1912: rose (did not infect apple guava) Colombia 1926: rose Brazil apple 1884: guava 1902: rose apple 1912: eucalypt Reports of Austropuccinia psidii occurrence

Japan 2007: Ohi’a (nursery) Florida, USA 1977: Allspice Hainan, China 2009: Rose apple Taiwan 1992: eucalypt Jamaica Puerto Rico (not established) 1934: Rose 1912: Rose Indonesia 2015: New Caledonia apple, Allspice apple 2013: Rose apple Hawaii, USA Eucalypt, Melaleuca 2005: Ohi’a Colombia Brazil 1926: rose 1884: Guava apple 1912: Eucalypt South Africa 1929-32: Rose apple, 2013: Myrtus Australia 1973: Serious communis 2010: New Zealand outbreak in diverse 2017: eucalypt Myrtaceae Metrosideros plantations Several examples of new reports of myrtle rust around the world What are the genetic similarities and differences among Austropuccinia psidii genotypes that infect diverse hosts in widely ranging global areas?

= Austropuccinia psidii occurrence From: A.C. Alfenas Characterization of Austropuccina psidii populations in Brazil

HYPOTHESIS: A. psidii jumped from guava to eucalypts shortly after their introduction to Brazil

Guava (Psidium guajava) Eucalypt (Eucalyptus spp.)

Graça et al. 2013 Mol Ecol 22: 6033-6047. Population genetics approaches to understand Austropuccinia psidii ecology and improve threat assessments.

• (Allele 2)---CACACACACACACACA---- = (CA)8 • (Allele 1)---CACACACACACACACACA--- = (CA)9

Precise collection information (e.g., host, GPS coordinates, date, etc.)

Host range tests Microsatellite (SSR) sequencing and analyses DNA extraction and PCR Austropuccinia psidii collections in Brazil

Brazil # samples Host 70 Eucalypt 63 Guava (Psidium guajava) 4 Rose apple (Syzygium jambos) 2 Brazilian guava (P. guineense) 4 Java plum (S. cumini) 3 Jabuticaba (Myrciaria cauliflora) 2 Pitanga (Eugenia uniflora)

Single pustule isolation

Graça et al. 2013 Mol Ecol 22: 6033-6047. Microsatellite genotyping of Austropuccinia psidii isolates

• (Allele 2)---CACACACACACACACA---- = (CA)8 • (Allele 1)---CACACACACACACACACA--- = (CA)9

DNA extraction and PCR

10 microsatellite loci were scored for 148 A. psidii isolates, which revealed 25 unique multilocus genotypes Principal coordinates analysis – 25 unique Austropuccinia psidii genotypes

Guava and Brazilian guava Eucalypt and rose apple

Graça et al. 2013 Mol Ecol 22: 6033-6047. Population structure of 148 Austropuccinia psidii isolates from seven myrtaceous hosts in Brazil

STRUCTURE v2.3.4 Graça et al. 2013 Mol Ecol 22: 6033-6047. Six evolutionary 1-100,000 years scenarios 1-1000 years modeled in

DIYABC 1-100,000 years

1-1000 years

Assuming two populations, an older divergence event between guava- and eucalypt- associated populations (Scenario 2) had a significantly higher posterior probability (0.9932) than divergence between guava- and eucalypt-associated populations within the last 1000 years (Scenario 1; 0.0068)

Assuming three populations, a more recent divergence of eucalypt-associated population from the ‘other’ population (Scenario 4) had significantly higher posterior probability (0.9836) than a more recent divergence event between guava- and eucalypt-associated populations (Scenario 3; 0.0000) Conclusions - Brazil study

• Austropuccinia psidii infections of eucalypt in Brazil did not originate via a host shift from guava;

• A. psidii is differentiated by host in Brazil, with at least two genotypically distinct biotypes; and

• Divergence between the two biotypes within the past 1000 years is highly unlikely How does genetic diversity of Austropuccinia psidii in Central/North America and Hawaii compare with that found in Brazil and Uruguay? Origin Host N BAPS cluster1 Geographic origin, host, Brazil Eucalyptus spp. 70 C2 and C3 Eugenia uniflora 2 C7 Myrciaria cauliflora 3 C9 and genetic cluster of Psidium guajava 63 C6 Psidium guineenese 2 C6 Austropuccinia psidii Syzygium cumini 4 C5 Syzygium jambos 3 C2 samples. Costa Rica Callistemon lanceolatus 2 C1 Jamaica Pimenta dioica 6 C8 Syzygium jambos 4 C1 Mexico Syzygium jambos 1 C1 Puerto Rico Syzygium jambos 1 C1 Uruguay Eucalyptus grandis 1 C2 Eucalyptus globulus 3 C2 Myrrhinium atropurpurea 1 C8 USA - Florida Melaleuca quinquenervia 5 C4 Myrcianthes fragrans 1 C4 Rhodomyrtus tomentosa 2 C4 Syzygium jambos 2 C4 USA - Hawaii Eugenia koolauensis 3 C1 Melaleuca quinquenervia 4 C1 1Bayesian analysis of Metrosideros excelsa 1 C1 population structure (BAPS) Metrosideros polymorpha 9 C1 Myrtus communis 1 C1 identified nine genetic Rhodomyrtus tomentosa 2 C1 clusters (C1 – C9) among Syzygium cumini 1 C1 226 Austropuccinia psidii Syzygium jambos 28 C1 Syzygium malaccense 1 C1 isolates. 226 Population structure of 226 Austropuccinia psidii samples

Population structure of 226 Austropuccinia psidii samples inferred using a Bayesian clustering algorithm implemented in BAPS with each individual represented by a vertical line partitioned into shaded segments corresponding to the isolate’s estimated mean membership coefficient for K = 9 genetic clusters; mean LnP(K) = −1782.56. Principal coordinates analysis of 226 Austropuccinia psidii samples

Principal coordinates analysis of the 23 mutlilocus genotypes of 226 Austropuccinia psidii isolates among nine clusters (C1–C9) as identified by BAPS based on a covariance matrix with data standardization. The first two axes explain 79% of the observed variation. Minimum-spanning network of Austropuccinia psidii microsatellite multilocus genotypes (MLGs) samples from Brazil (BR) Costa Rica (CR) Jamaica (JM) Mexico (MX) Puerto Rico (PR) Uruguay (UR) Florida (FL) USA Hawaii (HI) USA on 18 hosts.

MLGs are represented by BAPS genetic clusters: C1 represents MLGs from Costa Rica on crimson bottlebrush (Callistemon lanceolatus), Jamaica, Mexico, Puerto Rico on rose apple (Syzygium jambos) and Hawaii, USA on koʻolau eugenia (Eugenia koolauensis), broad-leaved paperbark (Melaleuca quinquenervia), pōhutukawa (Metrosideros excelsa), ʻōhiʻa lehua (M. polymorpha), common myrtle (Myrtus communis), rose myrtle (Rhodomyrtus tomentosa), Java plum (S. cumini), rose apple and Malay rose apple (S. malaccense); C2 represents MLGs collected from Brazil on eucalypts (Eucalyptus spp.) and rose apple and from Uruguay on eucalypts (Eucalyptus grandis and E. globulus); C3 represents one MLG collected from Brazil on eucalypts; C4 represents MLGs collected from Florida, USA on broad-leaved paperbark, twin berry (Myrcianthes fragrans), rose myrtle and rose apple); C5 represents one MLG collected in Brazil on Java plum; C6 represents one MLG collected in Brazil on guava (Psidium guajava)and Brazilian guava (P. guineense); C7 represents one MLG collected in Brazil on pitanga (Eugenia uniflora); C8 represents MLGs collected from Jamaica on allspice (Pimenta dioica) and Uruguay on sweet flower (Myrrhinium atropurpureum); C9 represents one MLG collected from Brazil on jabuticaba (Myrciaria cauliflora). Sizes of circles are proportional to MLG frequency. Connections are labelled with Bruvo genetic distances if different from 0.04, which corresponds to 1 mutational step at one locus. Broken lines connect MLGs that are separated by distances >0.20. Loops with dotted lines in the network (i.e., with C1-, C5-and C8-associated MLGs) indicate multiple, tied minimum-spanning trees At least three biotypes are represented within the minimum-spanning network

C6: Guava/Brazilian C1 and C4: Guava - Brazil “Pandemic” diverse hosts in diverse regions

Allspice – Jamaica Biotype ?? C2 and C3: eucalypt/rose apple – Brazil/Uruguay Others have confirmed that the “Pandemic” biotype of Austropuccina psidii occurs in Australia, China (Hainan), New Caledonia, Indonesia, and Colombia

Machado et al. 2015. McTaggart et al. 2016. Granados et al. 2017. Austalasian Pl. Pathol. 44: 455-462 Austalasian Pl. Pathol. 45: 83-89 Austalasian Pl. Pathol. 46: 267-275 Bioclimatic (A) All genotypes (N = 403) modeling of Austropuccinia (B) C1/C4 biotype (Pandemic: psidii occurrence points from diverse hosts in Costa Rica, Jamaica, MaxEnt models of suitable Mexico, Puerto Rico, USA- climate space (potential Hawaii/Florida and Australia; N = 137) distribution) for Austropuccinia psidii based on 19 global (C) C2/C3 biotype (eucalypt/rose apple– bioclimatic variables Brazil/Uruguay; N = 80) derived from the WorldClim (worldclim.org) database (D) C6 biotype (guava/Brazilian guava–Brazil; N = 60) MaxEnt models of suitable climate space (potential distribution) for Austropuccinia psidii in Hawaii, USA

(B) C1/C4 biotype (Pandemic: occurrence points from diverse hosts in (D) C6 biotype Costa Rica, Jamaica, Mexico, Puerto Rico, (guava/Brazilian guava USA-Hawaii/Florida and Australia; N = 137) –Brazil; N = 60) Assessment of Brazilian Austropuccinia psidii biotype virulence to Hawaiian ohi’a (Metrosideros polymorpha)

Mean severity of ohi’a leaf damage 20 15

% UFV2 Euba1 10 Brazilian M. caulif. A. psidii strains

Severity 5 P. guajava P. araca 0 UFV2 Control vs inoculated Euba1 P. araca M. caulif. P. guajava The Brazilian A. psidii eucalypt/rose apple biotype was highly virulent on Hawaiian ohi’a plants

Silva et al. 2014. Pacific Science 68: 47-56 Threat assessment of Austropuccinia psidii eucalypt/rose apple biotype to native ohi’a in Hawaii

Hawaiian ohi’a plants are very Hawaii possesses suitable climate space for susceptible to A. psidii the A. psidii eucalypt/rose apple biotype eucalypt/rose apple biotype from from Brazil Brazil (Silva et al. 2014)

The A. psidii eucalypt/rose apple biotype from Brazil likely poses a major threat to native ohi’a in Hawaii, should this biotype ever be introduced. MaxEnt models of suitable climate space (potential distribution) for Austropuccinia psidii in Australia, New Zealand, and adjacent regions

(B) C1/C4 biotype (Pandemic: occurrence points from diverse hosts in (D) C6 biotype Costa Rica, Jamaica, Mexico, Puerto Rico, (guava/Brazilian guava USA-Hawaii/Florida and Australia; N = 137) –Brazil; N = 60) Example assessment of potential threats posed by eucalypt/rose apple (Brazil/Uruguay) biotype of Austropuccinia psidii if introduced to Oceania

A previous study by Zauza et al. (2010) Bioclimatic modelling indicates that eastern showed that several myrtaceous Australia and part of New Zealand have species from Australia were suitable climate space for A. psidii eucalypt susceptible to A. psidii eucalypt/rose biotype apple biotype from Brazil

Australia

The “eucalypt/rose apple” biotype of A. psidii from Brazil/Uruguay likely represents a threat to several myrtaceous species in regions of Oceania, including New Zealand Climate change will influence the distribution of Austropuccinia psidii An example of climate-change predictions for South America

Current 2050 Predicted current (based on years 1950-2000) Model projection of suitable climate space for suitable climate space for A. psidii in A. psidii in South America for the 2050s (years 2040-2069) South America based on 169 occurrences. using CCCMA-CGCM global circulation model and A1B SRES emissions scenario. These predictions used MaxEnt and 19 bioclimatic surfaces from WorldClim. The dark green represents areas with suitable climate for A. psidii, with light green, yellow, orange, and red indicating increased suitability, respectively (From Klopfenstein et al. 2011) Summary

• Austropuccinia psidii is an invasive rust pathogen that appears to comprise at least 9 genetic clusters (C1-C9).

• The C1/C4 clusters (“Pandemic biotype”) associated with diverse hosts in many geographic regions.

• Each A. psidii biotype (e.g., C1/C4, C2/C3, and C6) has different ecological behavior, in terms of associated hosts and suitable climate space (potential distribution).

• An understanding of A. psidii biotypes/genotypes is essential for tracking pathogen spread, conducting threat assessments developing regulatory measures, and implementing management strategies for this invasive pathogen. Many questions remain

• What is the role of climate change in the emergence of myrtle rust?

• What is the role of sexual reproduction in the adaptation of A. psidii?

• What is the adaptive capacity of asexually reproduced A. psidii?

• Do different genetic groups of A. psidii represent cryptic species?

• What is the source of the Eucalypt/rose apple-Brazil/Uruguay biotype of A. psidii?

• What is the source of the Pandemic biotype of A. psidii? Thank you!

From: A.C. Alfenas Acknowledgments