The wheat blast pathogen Pyricularia graminis-tritici has complex origins and a disease cycle spanning multiple grass hosts

Abstract

The wheat blast disease has been a serious constraint for wheat production in Latin America since the late 1980s. We used a population genomics analysis including 95 genome sequences of the wheat blast pathogen Pyricularia graminis-tritici ( Pygt ) and other Pyricularia species to show that Pygt is a distinct, highly diverse pathogen species with a broad host range. We assayed 11 neutral SSR loci in 526 Pygt isolates sampled from wheat and other grasses distributed across the wheat-growing region of Brazil to estimate gene flow, assess the importance of sexual reproduction, and compare the genetic structures of Pygt populations infecting wheat and nearby grasses. Our results suggest a mixed reproductive system that includes sexual recombination as well as high levels of gene flow among regions, including evidence for higher gene flow from grass-infecting populations and into wheat-infecting populations than vice versa. The most common virulence groups were shared between the grass- and wheat-infecting Pygt populations, providing additional evidence for movement of Pygt between wheat fields and nearby grasses. Analyses of fruiting body formation found that proto-perithecia and perithecia developed on senescing stems of wheat and other grass hosts, suggesting that sexual reproduction occurs mainly during the saprotrophic phase of the disease cycle on dead residues. Phalaris canariensis (canarygrass) supported the fullest development of perithecia, suggesting it is a promising candidate for identifying the teleomorph in the field. Based on these findings, we formulated a more detailed disease cycle for wheat blast that includes an important role for grasses growing near wheat fields. Our findings strongly suggest that widely grown pasture grasses function as a major reservoir of wheat blast inoculum and provide a temporal and spatial bridge that connects wheat fields across Brazil.

Author summary
After the first wheat blast epidemic occurred in 1985 in Paraná, Brazil, the disease spread to Bolivia, Argentina, and Paraguay, and was introduced into Bangladesh in 2016 followed by India in 2017. Wheat blast is caused by Pyricularia graminis-tritici ( Pygt ), a highly diverse pathogen species related to the rice blast fungus P. oryzae , but with an independent origin and a broader host range. We conducted a large scale contemporary sampling of Pygt from symptomatic wheat and other grass species across Brazil and analyzed the genetic structure of Pygt populations. Pygt populations on both wheat and other grasses had high genotypic and virulence diversity, a genetic structure consistent with a mixed reproductive system that includes regular cycles of recombination. The pathogen formed sexual fruiting structures (perithecia) on senescing stems of wheat and other grasses. Historical migration analyses indicated that the majority of gene flow has been from Pygt populations on other grasses and into the Pygt population infecting wheat, consistent with the hypothesis that Pygt originated on other grasses before becoming a wheat pathogen. We found that the Pygt populations infecting wheat were indistinguishable from the Pygt populations infecting other grass species, including signal grass ( Urochloa brizantha ). Because U. brizantha is a widely grown grass pasture often found next to wheat fields, we propose that it functions as reservoir of Pygt inoculum that provides a temporal and spatial bridge that connects wheat fields in Brazil.

Full text

1
The wheat blast pathogen Pyricularia graminis-tritici has complex origins and a disease 1
cycle spanning multiple grass hosts 2
3
Vanina L. Castroagudín1¶, Anderson L. D. Danelli2¶; Silvino I. Moreira3¶; Juliana T. A. 4
Reges1¶, Giselle de Carvalho1, João L.N. Maciel4, Ana L. V. Bonato4, Carlos A. Forcelini5, 5
Eduardo Alves3, Bruce A. McDonald6, Daniel Croll6,7 , Paulo C. Ceresini1* 6
7
1 Department of Crop Protection, Agricultural Engineering, and Soils, UNESP University of 8
São Paulo State, Ilha Solteira Campus, São Paulo, Brazil.
9
2 Faculdades Integradas do Vale do Iguaçu, Uniguaçu, União da Vitória, Paraná, Brazil. 10
3 Department of Plant Pathology, Federal University of Lavras, Lavras, Minas 11
Gerais, Brazil. 12
4Brazilian Agriculture Research Corporation – Embrapa Wheat (EMBRAPA – Trigo), Passo 13
Fundo, Rio Grande do Sul, Brazil. 14
5Faculdade de Agronomia e Medicina Veterinária, UPF, Passo Fundo, Rio Grande do Sul, 15
Brazil 16
6Plant Pathology Group, Institute of Integrative Biology, ETH Zürich, Zürich, Switzerland. 17
7Laboratory of Evolutionary Genetics, Institute of Biology, University of Neuchâtel, 18
Neuchâtel, Switzerland. 19
20
Short title: 21
Pyricularia graminis-tritici on wheat and other poaceous hosts 22
*Corresponding author. E-mail: paulo.ceresini@bio.feis.unesp.br 23
¶These authors contributed equally to this manuscript. 24
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

2
Abstract
25
The wheat blast disease has been a serious constraint for wheat production in Latin America
26
since the late 1980s. We used a population genomics analysis including 95 genome
27
sequences of the wheat blast pathogen Pyricularia graminis-tritici (Pygt) and other
28
Pyricularia species to show that Pygt is a distinct, highly diverse pathogen species with a
29
broad host range. We assayed 11 neutral SSR loci in 526 Pygt isolates sampled from wheat
30
and other grasses distributed across the wheat-growing region of Brazil to estimate gene
31
flow, assess the importance of sexual reproduction, and compare the genetic structures of
32
Pygt populations infecting wheat and nearby grasses. Our results suggest a mixed
33
reproductive system that includes sexual recombination as well as high levels of gene flow
34
among regions, including evidence for higher gene flow from grass-infecting populations and
35
into wheat-infecting populations than vice versa. The most common virulence groups were
36
shared between the grass- and wheat-infecting Pygt populations, providing additional
37
evidence for movement of Pygt between wheat fields and nearby grasses. Analyses of
38
fruiting body formation found that proto-perithecia and perithecia developed on senescing
39
stems of wheat and other grass hosts, suggesting that sexual reproduction occurs mainly
40
during the saprotrophic phase of the disease cycle on dead residues. Phalaris canariensis
41
(canarygrass) supported the fullest development of perithecia, suggesting it is a promising
42
candidate for identifying the teleomorph in the field. Based on these findings, we formulated
43
a more detailed disease cycle for wheat blast that includes an important role for grasses
44
growing near wheat fields. Our findings strongly suggest that widely grown pasture grasses
45
function as a major reservoir of wheat blast inoculum and provide a temporal and spatial
46
bridge that connects wheat fields across Brazil.
47
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

3
Author summary (200 words)
48
After the first wheat blast epidemic occurred in 1985 in Paraná, Brazil, the disease
49
spread to Bolivia, Argentina, and Paraguay, and was introduced into Bangladesh in 2016
50
followed by India in 2017. Wheat blast is caused by Pyricularia graminis-tritici (Pygt), a
51
highly diverse pathogen species related to the rice blast fungus P. oryzae, but with an
52
independent origin and a broader host range. We conducted a large scale contemporary
53
sampling of Pygt from symptomatic wheat and other grass species across Brazil and analyzed
54
the genetic structure of Pygt populations. Pygt populations on both wheat and other grasses
55
had high genotypic and virulence diversity, a genetic structure consistent with a mixed
56
reproductive system that includes regular cycles of recombination. The pathogen formed
57
sexual fruiting structures (perithecia) on senescing stems of wheat and other grasses.
58
Historical migration analyses indicated that the majority of gene flow has been from Pygt
59
populations on other grasses and into the Pygt population infecting wheat, consistent with the
60
hypothesis that Pygt originated on other grasses before becoming a wheat pathogen. We
61
found that the Pygt populations infecting wheat were indistinguishable from the Pygt
62
populations infecting other grass species, including signal grass (Urochloa brizantha).
63
Because U. brizantha is a widely grown grass pasture often found next to wheat fields, we
64
propose that it functions as reservoir of Pygt inoculum that provides a temporal and spatial
65
bridge that connects wheat fields in Brazil.
66
67
Introduction
68
69
Pyricularia is a species-rich genus including many fungal pathogens that show specialization
70
towards different host species in the Poaceae family, including rice (Oryza sativa), wheat
71
(Triticum aestivum), oat (Avena sativa), barley (Hordeum vulgare), and millets (Eleusine
72
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

4
coracana, Pennisetum glaucum, Setaria italica), as well as more than 50 other species of
73
grasses [1-5]. Several studies indicated that distinct Pyricularia species emerged through
74
repeated radiation events from a common ancestor [6, 7]. Such radiation events often result
75
from ecological adaptations that include host jumps or shifts and changes in pathogenicity [4,
76
8]. These ecological adaptations may lead to the emergence of new species of "domesticated"
77
host-specialized fungal pathogens infecting agricultural crops from "wild" ancestral source
78
populations found on undomesticated plants [4, 8]. Examples of speciation following host
79
specialization are common in cereal agro-ecosystems and were already described for several
80
plant pathogenic fungi, including Pyricularia oryzae on rice and P. grisea on Digitaria spp.
81
[5], Zymoseptoria tritici on wheat [9], Rhynchosporium commune on barley [10],
82
Ceratocystis fimbriata on cacao (Theobroma cacao), sweet potato (Ipomoea batatas) and
83
sycamore (Platanus spp.) [11], and Microbotryum violaceum on Silene spp. [12]. For P.
84
oryzae, causal agent of rice blast [5, 13], strains that infect rice are thought to have emerged
85
by ecological adaptation via host shifts from millet (Setaria spp.) to rice and to have co-
86
evolved with their respective hosts during the domestication of rice and millet in China about
87
7000 BC [14].
88
A previous study indicated that a new Pyricularia species, named Pyricularia
89
graminis-tritici (Pygt), emerged in southern Brazil during the last century as the pathogen
90
causing wheat blast [15]. Pygt is closely related to P. oryzae [15]. Wheat blast was first
91
reported in Paraná State, Brazil in 1985 [16, 17] and since then has become an increasingly
92
important disease, causing crop losses ranging from 40% to 100% [18]. Blast disease has also
93
been reported in other important crops growing in the same agro-ecosystems in Latin
94
America, including pastures of signal grass (Urochloa brizantha, ex Brachiaria brizantha),
95
barley, oats, rye (Secale cereale), and triticale (x Triticosecale). Although other Pyricularia
96
species can cause blast symptoms on wheat, we focused this study on Pygt, which is the
97
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

5
major species associated with wheat blast [15, 17, 19-24]. Since its discovery, Pygt has
98
spread across all wheat-cropping areas in Brazil [17, 18, 25-27] and is now found in Bolivia,
99
Argentina and Paraguay [28]. Its first report outside South America was an outbreak in
100
Bangladesh in 2016 [29-31] followed by its spread to India in 2017 [32, 33]. Wheat blast is a
101
major quarantine disease in the United States [27] and it is considered a threat to wheat
102
cultivation in disease-free areas across Asia, Europe, and North America [34].
103
Pygt can be dispersed over short and long distances by aerial inoculum (conidia) [35]
104
and also on infected seeds [36]. Unlike most Pyricularia species, Pygt isolates recovered
105
from wheat can infect a wide range of hosts, including the tribes Hordeae, Festuceae,
106
Avenae, Chlorideae, Agrosteae and Paniceae [37]. Under natural field conditions, close
107
physical proximity between cultivated plants and other poaceous hosts (i.e., weeds or
108
invasive grass species) could enable genetic exchange among Pyricularia populations on
109
different hosts and facilitate host shifts. Cross-infection and inter-fertility between fungal
110
strains from different grass hosts were hypothesized to play a role in the emergence of wheat
111
blast [38, 39]. Evidence to support this hypothesis was presented in a recent study that
112
analyzed variation in the avirulence genes PWT3 and PWT4 [40]. This study proposed that
113
wheat blast emerged via a host shift from a Pyricularia population infecting Lolium. In their
114
model, a Lolium-derived isolate carrying the Ao avirulence allele at the PWT3 locus infected
115
a susceptible wheat cultivar carrying the rwt3 susceptibility allele. The model further
116
proposes that the spread of wheat blast in the 1980s was enabled by the widespread
117
cultivation in Brazil of susceptible wheat cultivars carrying rwt3. Selection on less common
118
Rwt3 wheat cultivars favored the emergence of pathogen strains with non-functional PWT3
119
alleles, and the authors proposed that it was these pwt3 strains that eventually became the
120
epidemic wheat blast population found in South America.
121
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

6
Pyricularia is considered a genus of pathogens with high evolutionary potential [39,
122
41, 42]. The evolutionary potential of a pathogen population reflects its ecology and biology,
123
and its population genetic structure [41, 42]. Pioneering studies on the genetic structure of
124
Pygt indicated a highly variable population distributed across different Brazilian states [43,
125
44]. Analyses of three regional populations sampled in Brazil between 2005 and 2008
126
suggested long distance gene flow and a mixed reproductive system [39]. These findings
127
indicated that Pygt is a pathogen with high evolutionary potential, according to the risk model
128
proposed by McDonald and Linde [41, 42].
129
Knowledge about the evolutionary potential of Pygt populations is needed to predict
130
the durability of genetic resistance to wheat blast. An intense search for blast resistance began
131
with the first report of the disease more than 30 years ago, but breeding success has been
132
erratic and inconsistent [45-48]. The average durability of resistant wheat varieties has been
133
only two to three years [49]. Furthermore, wheat genotypes behaved differently in different
134
regions, indicating genotype-by-environment interactions or a region-specific distribution of
135
virulence groups [50]. Given that Pygt is now present in all Brazilian wheat growing areas
136
[15, 28], it is likely that both the incidence and severity of wheat blast are affected by the
137
virulence groups that predominate in each region [39]. In fact, the occurrence of virulence
138
groups in Pygt populations was already described [39, 43, 50, 51], but information about the
139
virulence composition and genetic structure of contemporary populations of the wheat blast
140
pathogen remains limited.
141
Several lines of evidence indicate that Pygt populations recombine regularly in Brazil:
142
both mating types and fertile strains were present in wheat fields, field populations contain
143
high genetic diversity, and gametic equilibrium is found among neutral marker loci [26, 39,
144
52]. Under laboratory conditions, Pygt isolates showed the capacity for sexual reproduction
145
[37] and were shown to be sexually compatible with Pyricularia isolates from other poaceous
146
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

7
hosts including plantain signalgrass (Urochloa plantaginea, ex Brachiaria plantaginea),
147
goosegrass (Eleusine indica), finger-millet (Setaria italica), rescuegrass (Bromus
148
catharticus), canary grass (Phalaris canariensis) and triticale (x Triticosecale) [52, 53].
149
Crosses between isolates recovered from wheat and Urochloa plantaginea produced
150
perithecia with asci and ascospores, a clear indicator of sexual reproduction [54], but
151
perithecia have not yet been found in blasted wheat fields and it remains unclear where and
152
when the sexual stage occurs.
153
Here we bring together findings from a series of experiments conducted to better
154
understand the origins of wheat blast and formulate an improved disease cycle. We first used
155
population genomic analyses including 36 Pygt strains originating from many different hosts
156
and 59 strains of other Pyricularia species to infer the genealogical relationships among
157
Pyricularia species and better define the phylogenetic boundaries of Pygt. We next generated
158
and analyzed a microsatellite dataset from 526 contemporary Brazilian isolates of Pygt
159
sampled from wheat fields and invasive grasses across Brazil to compare the genetic
160
structures of Pygt populations found on wheat and other grasses. We then compared the
161
distribution of Pygt virulence groups found in wheat fields with the distribution of virulence
162
groups found on invasive grasses growing in or near those wheat fields. Finally, we
163
conducted experiments to identify grass hosts and tissues where sexual perithecia are most
164
likely to form to better understand the importance of sexual recombination in Pygt population
165
biology and identify the hosts most likely to support formation of the teleomorph. This
166
combination of experiments provided novel insights into the origins and epidemiology of
167
wheat blast.
168
169
Results
170
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

8
Several Pyricularia species were recovered from blast lesions on wheat and invasive
171
grasses
172
We sampled Pyricularia spp. from wheat and other poaceous hosts in naturally infected
173
wheat fields distributed across the seven states where wheat is grown in Brazil. Amongst the
174
556 Pyricularia spp. isolates included in our analyses, 30 isolates were not Pygt (Table 1,
175
Supplementary Table 2). Based on the sequence of the hydrophobin MPG1, an isolate from
176
DF-GOW was classified as P. urashimae. This was the only isolate recovered from a wheat
177
head that was not Pygt. The 23 isolates from MSP included two isolates of P. grisea
178
(recovered from Digitaria sanguinalis), 13 isolates of P. pennisetigena (from Cenchrus
179
echinatus, Eragrostis plana, Panicum maximum and Urochloa brizantha), five isolates of
180
P. urashimae (from Avena sativa, Echinochloa crusgalli, P. maximum, and U. brizantha),
181
and three Pyricularia isolates that could not be identified at the species level (from P.
182
maximum and U. brizantha). The five isolates found in PRP included two isolates of P. grisea
183
(from D. sanguinalis), one of P. pennisetigena (from U. brizantha), and two of P. urashimae
184
(from Chloris distichophylla and P. maximum). Isolate 363 came from a rice field, probably
185
from a Digitaria spp., and was classified as P. grisea.
186
187
Population genomic analyses reveal that Pyricularia graminis-tritici comprises a single
188
highly diverse species
189
Our first goal in this study was to infer the genealogical relationships among the Pyricularia
190
species found in Brazil and to determine if the Pygt strains associated with blast on wheat and
191
other grasses comprise a single species. We extended the analysis from Islam et al. [31] by
192
adding into the genealogy Pyricularia isolates from 10 non-wheat hosts sampled in sympatry
193
with 22 wheat blast isolates. The 47 P. oryzae strains associated with rice blast grouped
194
together as a near-clonal genotype that was distinct from the group of 32 Pygt strains found
195
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

9
on wheat and other grasses in Brazil and Bangladesh (Fig 1). The inferred genealogical
196
relationships indicated that the Pygt strains sampled mainly from wheat comprise a single
197
highly diverse species. The formerly described P. oryzae pathotype Triticum clade (indicated
198
as PoT in the genealogy) [15] was not distinct from the P. graminis-tritici (Pygt) clade (Fig
199
1). The clade formed by Pygt strains sampled from infected wheat ears and other grass hosts
200
contained much more polymorphism than the rice-infecting P. oryzae strains available in
201
public genome databases. Despite the higher overall diversity, several of the Brazilian Pygt
202
strains formed sub-clades that may represent expanded clonal lineages (Fig 1). In two of
203
these sub-clades, closely related strains from the same sub-clade were found infecting
204
different hosts.
205
206
Populations of Pygt from wheat and other grasses share genotypes
207
To explore the possibility of gene and genotype flow among the Pygt populations infecting
208
wheat and other grasses, we conducted population genetic analyses using 11 neutral
209
microsatellite (SSR) markers in an expanded dataset including 526 Brazilian Pygt isolates. A
210
total of 198 different multilocus microsatellite genotypes (MLMGs) were found among the
211
526 isolates (Table 2, Fig 2). Of these MLMGs, 165 (83%) were found in only one
212
population (Tables 2-4), but 33 MLMGs (17%) were shared by sympatric (from the same
213
region) or allopatric (from different regions) populations of Pygt. These 33 MLMGs
214
encompassed 257 isolates (224 from wheat, and 33 from other grasses), with 20 of these
215
MLMGs (corresponding to 176 isolates) found exclusively on wheat. The number of
216
MLMGs within a population that were shared across populations ranged from four (7
217
isolates) in SPW to 15 (46 isolates) in MS. No MLMGs were shared between the isolates
218
collected in 2005 and those collected in 2012 (Tables 3 and 4), indicating that Pygt clones do
219
not persist over time.
220
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

10
Table 1. Populations of the blast pathogen Pyricularia graminis-tritici from wheat and other poaceous hosts characterized in this study.
221
Species, host, cultivar
Population
Sampling Location, State
Coordinates
Sampling year
N
Triticum aestivum
Several cultivars
2005W
Central-southern Brazil
...
2005
79
BRS 254, BR18
DF-GOW
Brasília, DF; and Rio Verde, GO
17°19'46.8"S, 50°06'17.5"W
2012-2013
86
BRS 264, BR18
MGW
Patrocínio, and Perdizes, MG
19°09'10.1"S, 47°16'5.7"W
2012-2013
62
BRS Guamirim
MSW
Amambaí, and Aral Moreira, MS
22°59'33.1"S, 55°20'11.1"W
2012-2013
82
CD 104
PRW
Londrina, Jandaia do Sul, PR
23°14'16.0"S, 51°10'10.5"W
2012-2013
74
...
RSW
Passo Fundo, São Luiz Gonzaga, São
Borja and Três de Maio, RS
28°33'16"S, 55°21'52.5"W
2012-2013
52
CD 116
SPW
Itaí, SP
23°33'8"S, 49°3'24.9"W
2012-2013
31
Total (T. aestivum)
466
Other poaceous hosts
Avena sativa, Cenchrus
echinatus, Chloris
distichophylla, Cynodon spp.,
Digitaria insularis, Digitaria
sanguinalis, Echinochloa
crusgalli, Eleusine indica,
Eragrostis plana, Panicum
maximum, Rhynchelytrum
repens, Sorghum sudanense,
and Urochloa brizantha.
MSP
Amambaí and Aral Moreira, MS
22°59'33.1"S, 55°20'11.1"W
2012-2013
29
PRP
Londrina, PR
23°14'16.0"S, 51°10'10.5"W
2012-2013
31
Total (Other poaceous hosts)
60
Total
526
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

11
The MSP and PRP populations sampled from other grass hosts shared 11 MLMGs
222
with the populations from wheat. These 11 shared MLMGs were found in 48 strains
223
originating from wheat and 33 strains recovered from other grass species, including Avena
224
sativa, Chloris distichophylla, Cynodon spp., Digitaria insularis, Digitaria sanguinalis,
225
Echinochloa crusgalli, Eleusine indica, Eragrostis plana, Panicum maximum, Rhynchelytrum
226
repens, Sorghum sudanense and Urochloa brizantha (Table 4). The genetic similarity among
227
all MLMGs and their geographical and host distributions are displayed as a minimal spanning
228
network in Fig 3, with the 11 shared MLMGs indicated in red text. The probability that any
229
two isolates drawn at random from the pool of 526 isolates would share one of these 11
230
MLMGs by chance in a recombining population ranged from 6.82-6 to 1.28-10 [55, 56] (Table
231
4), hence it is highly likely that isolates with the same MLMG represent the same clone or
232
clonal lineage. These 11 MLMGs found on both wheat and other grasses provide compelling
233
evidence for the existence of Pygt clones with a broad host range, with transmission among
234
hosts growing in the same region likely occurring via dispersal of asexual spores, and
235
transmission among distant geographical regions likely occurring via movement on infected
236
seeds.
237
The clonal fraction inferred in each geographical population ranged from 0.13 in SPW
238
to 0.72 in DF-GOW, whereas the evenness ranged from 0.19 in the DF-GOW population to ~
239
0.90 in SPW. Overall, we found that the MLMGs were not uniformly distributed in the
240
majority of the populations (Table 2). The effective number of genotypes (GO) ranged from
241
4.5 to 23.4 and was highest in Pygt populations from SPW (GO = 23.4), MSW (GO = 21.8) and
242
PRW (GO = 18.3) and lowest in DF-GOW (GO = 4.5) (Table 2). The allelic richness averaged
243
across ten populations was 2.75. The MSP population from other grasses had the highest
244
allelic richness (3.18) (Table 2).
245
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

12
Table 2. Measures of gene and genotypic and clonal diversity in populations of Pyricularia graminis-tritici from wheat and other
246
poaceous hosts in Brazila
247
Species, host,
population
Nb
No. genotypesc
No. population
specific and
shared
genotypes d
Clonal fractione
Eve.f
Ef. No.
genotypesg,h
Allelic
Richnessi
Triticum aestivum
2005W
79
26
26 (0)
0.67
0.18
4.80 c
3.06 ab
DF-GOW
86
23
14 (9)
0.73
0.20
4.50 c
2.40 b
MGW
62
27
15 (12)
0.56
0.27
7.42 bc
2.29 b
MSW
82
45
26 (19)
0.45
0.48
21.83 a
3.05 ab
PRW
74
45
34 (11)
0.39
0.41
18.38 a
3.08 ab
RSW
52
19
9 (10)
0.63
0.44
8.40 b
2.53 ab
SPW
31
27
22 (5)
0.13
0.87
23.44 a
2.74 ab
Other Poaceae
MSP
29
16
9 (7)
0.43
0.64
10.32 b
3.18 a
PRP
31
17
10 (7)
0.45
0.47
7.94 bc
2.45 b
Total
526
198
165 (33)
0.49 j
2.75 j
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

13
a The measures of genotypic/clonal diversity were calculated with GenoDive ver. 2.0b.17
248
[57].
249
b N = sample size.
250
c Number of genotypes identified with the different markers in each population.
251
d Number of specific genotypes per population; the number of genotypes shared with
252
other populations is shown in brackets.
253
e Clonal fraction is the proportion of fungal isolates originating from asexual
254
reproduction. The clonal fraction was calculated as 1 – [number of different
255
genotypes/total number of isolates].
256
f Eve, the evenness calculated as the ratio between the effective number of genotypes
257
and the number of genotypes. An evenness value of 1 indicates that all genotypes have
258
equal frequencies.
259
g Effective number of genotypes = Stoddart and Taylor’s genotypic diversity (GO).
260
h Means followed by the same letter are not significantly different (p ≤ 0.05) based on
261
pairwise bootstrap tests, based on 1,000 permutations with subsampling to match the
262
size of the smallest population, calculated with GenoDive ver. 2.0b.17 [57].
263
i Average allelic richness based on minimum sample size of 16 individuals calculated
264
according to El Mousadik and Petit [58].
265
j Averaged over the nine populations examined.
266
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

14
Table 3. Number of multilocus microsatellite genotypes shared between Brazilian population of Pyricularia graminis-tritici from wheat and
267
other poaceous hosts.
268
269
270
Species, populations
Wheat (Triticum aestivum)
Other Poaceae
2005w
DF-GOw
MGw
MSw
PRw
RSw
SPw
MSp
PRp
2005W
-
0
0
0
0
0
0
0
0
DF-GOW
-
9
1
0
1
0
0
0
MGW
-
3
1
4
0
0
0
MSW
-
8
8
2
5
4
PRW
-
1
4
0
3
RSW
-
0
3
0
SPW
-
0
1
MSP
-
3
PRP
-
Total number of shared genotypes in each population
0
9
11
14
10
10
5
7
7
Total number of isolates with shared genotypes in each population
0
68
45
43
19
41
8
12
21
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

15
Table 4. Number of isolates showing each of the eleven multilocus microsatellite genotypes shared among sympatric populations of Pyricularia
271
graminis-tritici sampled from wheat and other poaceous hosts from Central-southern Brazil
272
Genotype
Number of isolates with shared genotype in each population
Total
Pgena
Hosts
2005W
DF-GOW
MGW
MSW
PRW
RSW
SPW
MSp
PRp
68
5
5
1
11
1.48-9
Triticum aestivum, Urochloa brizantha
69
10
5
2
17
2.10-7
T. aestivum, Echinochloa crusgalli, U.
brizantha
75
1
4
5
1.97-9
T. aestivum, Avena sativa, Digitaria
sanguinalis, Rhynchelytrum repens
76
2
1
2
5
1.01-7
T. aestivum, A. sativa, Eragrostis plana, U.
brizantha
79
2
3
9
14
1.39-8
T. aestivum, E. crusgalli, D. sanguinalis,
Eleusine indica, Panicum maximum, R.
repens, Sorghum sudanense, U. brizantha
80
8
2
1
11
2.32-8
T. aestivum, Chloris distichophylla
83
2
3
5
1.65-9
T. aestivum, D. sanguinalis, E. crusgalli
99
2
1
2
5
1.27-7
T. aestivum, D. sanguinalis, R. repens
122
1
1
1
3
1.28-10
T. aestivum, A. sativa, Cynodon spp.
156
1
1
2
6.82-6
T. aestivum, Digitaria insularis
170
1
2
3
2.46-9
T. aestivum, C. distichophylla
Total
0
0
0
32
4
11
1
12
21
81
13 different hosts
a Pgen provides an estimate of the probability of identical genotypes arising from sexual reproduction and random mating and it is identical to
273
the genotype probability [55, 56].
274
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

16
Pygt populations on wheat and other grasses are connected by gene flow
275
The overall fixation index indicated a weak but significant differentiation (RST = 0.07, p ≤
276
0.001) among all populations. When Pygt populations from wheat were analyzed separately,
277
AMOVA showed a low but still significant level of population differentiation (RST = 0.07, p ≤
278
0.001), with 93% of the genetic variation distributed within populations. In contrast, when the
279
two Pygt populations from other grasses (separated by ~ 430 km) were compared, AMOVA
280
indicated an absence of population differentiation (RST = 0.02, p = 0.29), with 98% of genetic
281
variation distributed within grass-infecting populations. The orthogonal contrast of Pygt
282
populations from wheat with Pygt populations from other poaceous hosts was significant but
283
the level of differentiation was very low (RCT = 0.04), with the majority of genetic variation
284
distributed within populations (93%) (Table 5). It is notable that no subdivision was found for
285
12 of the 15 pairwise comparisons between the two Pygt populations obtained from other
286
grass hosts (MSP and PRP) and the Pygt populations from wheat (Table 6).
287
288
Historical gene flow was detected among Pygt populations from wheat and other
289
grasses.
290
The unidirectional migration models gave a better fit to the data than the panmictic or
291
bidirectional models (Table 7). Historical migration analyses support unidirectional gene
292
flow into the Pygt population infecting wheat from the Pygt population infecting other
293
grasses (contributing 4.3 migrants per generation in average) (Table 8), suggesting that the
294
Pygt population infecting wheat is composed of immigrants from the Pygt population
295
infecting other grasses. There were no significant differences between Θ values (Table 8).
296
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

17
Table 5. Hierarchical distribution of gene diversity among populations of Pyricularia
297
graminis-tritici from wheat and other poaceous hosts and P. oryzae from rice in Central-
298
southern Brazil a
299
Source of variation
d.f.
Variance
components
% of
variance
Fixation
Index
p
Among populations from wheat
Among populations
6
1.71
7.1
RST = 0.07
< 0.0001
Within populations
205
22.54
92.9
Total
211
24.25
Among populations from other poaceous host
Among populations
1
0.71
2.0
RST = 0.02
0.2092
Within populations
31
35.82
98.0
Total
32
36.53
Populations from wheat blast vs. other poaceous hosts
Between groups
1
1.12
3.85
RSC = 0.04
< 0.0001
Among populations within groups
26
0.84
2.90
Within populations
627
27.13
93.25
Total
654
29.09
a. The analysis of molecular of variance (AMOVA) was performed using Arlequin version 3.1
300
[59]. The distance method is based on the sum of squared size differences among alleles
301
between two haplotypes for microsatellite data according Slatkin [60]; Significance values
302
were obtained using a non-parametric approach (1023 permutations) [61].
303
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

18
Table 6. Pairwise differentiation among populations of Pyricularia graminis-tritici from wheat and other poaceous hosts in Central-southern
304
Brazil
305
RST
Wheat (T. aestivum)
Other Poaceae
2005W
DF-GOW
MGW
MSW
PRW
RSW
SPW
MSP
PRP
2005W
-
DF-GOW
0.0066
-
MGW
0.0078
0.0340
-
MSW
0.0561
0.1567*
0.0403
-
PRW
0.0944*
0.2387*
0.1213*
0.0092
-
RSW
0.0610
0.1583*
0.0400
0.0298
0.0830
-
SPW
0.0281
0.1734*
0.0969
0.0480
0.0660
0.1499
-
MSP
-0.0099
0.0750
0.0203
0.0316
0.0690
0.0733
-0.0050
-
PRP
0.0380
0.1794*
0.1274*
0.0927
0.0674
0.1661*
0.0044
0.0196
-
306
a. Fixation index among the nine populations of P. graminis-tritici: RST = 0.07 (p ≤ 0.001)
307
b. Calculation were conducted on clone corrected data and based on the sum of squared size differences among alleles (RST) between two
308
haplotypes for microsatellite data according Slatkin [60]. The test was performed using Arlequin version 3.1[59]. Significance values have
309
been tested using a non-parametric approach (1023 permutations) [61], at  = 0.05 after Bonferroni correction for multiple comparisons [62].
310
RST values followed by ‘*’ showed statistically significant p-values at corrected  = 0.005).
311
312
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

19
Table 7. Comparison of models of historical migration between pairs of Brazilian
313
populations of Pyricularia graminis-tritici grouped by original hosts (wheat and other
314
Poaceae) based on Bezier approximation scores to the marginal likelihooda
315
a Migration analyses were implemented in MIGRATE-n v. 3.6.11 at the CIPRES Science
316
Gateway [63], using a maximum likelihood test based on the Markov chain Monte Carlo
317
(MCMC) method [64-68]. Each of the five runs had ten short initial chains, one long final
318
chain, a static heating scheme (temperatures: 1.0, 100, 100, 1,0000 and 100,000), and
319
swapping interval of 1. The initial chains had 500-recorded steps, a sampling increment of
320
100, with 2,500 trees recorded per short sample. The long chain had 8,334-recorded steps, a
321
sampling increment of 500, six concurrent replicates, and 500 trees as burn-in. The final
322
number of sampled parameter values was 25,002,000 iterations.
323
b The likelihood values of the four migration models were compared to select the model that
324
best fitted the data based on the Log of the Bayes Factor (LBF). LBF was calculated as:
325
2[ln(Prob(Data | ModelX)) – ln (Prob(Data | best of the four models))]. The highest the LBF
326
values, the better the fit of the migration model to the data [64].
327
c, d. The run with the highest likelihood chosen to represent a given model is in bold, and the
328
model that best fit the migration between a given pair of populations is shaded.
329
Populations, run
Migration model (Bezier approximation score)
Panmictic
Bidirectional
From 1 to 2
From 2 to 1
Wheat (1) and other Poaceae (2)
1
-15811.0
-13724.8
-13759.2
-13127.7
2
-15821.7
-13650.8
-13767.6
-13123.6
3
-15831.5
-13761.6
-13769.7
-13112.1
4
-15828.0
-13801.1
-13760.5
-13104.3
5
-15823.1
-13651.3
-13751.6
-13117.0
LBF
-5397.8
-1077.4
-1279.0
0.0
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

20
Table 8. Migration parameter between pairs of Brazilian populations of Pyricularia graminis-tritici from wheat and other Poaceae, under
330
the best fit migration model and obtained by Bayesian inference.
331
Population pairs
Migration model that best fit
the data
Migration parameters estimates a
ΘWheatb
Θ Other Poaceae
xNmDonorRecipientc
Wheat and other
Poaceae
Directional (from other
Poaceae to wheat)
2.97
(0.0001 – 3.48)
0.63
(0.0001 – 0.84)
xNmPoaceaeWheat = 4.26
(0.0001 - 13.00)
a Bayesian estimates of the migration parameters were calculated with MIGRATE-n v. 3.6.11 at the CIPRES Science Gateway [63],
332
using a maximum likelihood test based on the Markov chain Monte Carlo (MCMC) method [64-68]. Values represent the mean
333
Bayesian estimate, and are followed by the 95% credibility intervals for each parameter given by the 0.025 and 0.975 quantiles of its a
334
posteriori distribution in parenthesis.
335
b Theta (Θ) values provide a measure of the effective population size; for haploids, Θ = 2Neμ, where Ne = effective population size and μ
336
= mutation rate for each locus [68].
337
c xNmDonorRecipient is the number of immigrants per generation; where N is the real population size, m is the fraction of the new
338
immigrants of the population per generation, and x is an inheritance scalar and x= 1 for haploids. The number of immigrants per
339
generation can also be expressed as the product ΘM (xNmDonorRecipient = Θrecipient MDonorRecipient); where M is m divided by the mutation
340
rate μ (M = m/μ), and it represents the importance of variability brought into the receiving population by immigration from the donor
341
population compared with the variability created by mutation [68].
342
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

21
Most of the Pygt populations were sexually recombining.
343
We consider a population to be sexually recombined when the majority of locus pairs are at
344
gametic equilibrium and/or IA or 𝑟̅
𝐷 are not significant (p > 0.05). Under these assumptions, 7
345
of the 9 populations had signatures consistent with sexual recombination. Only 2005W and
346
MSw showed evidence for significant clonal reproduction, with six and five pairs of loci
347
showing significant GD, respectively, and significant IA and 𝑟̅
𝐷 (p < 0.001) (Table 9).
348
Because MSw possessed the highest number of shared MLMGs among populations (N=32,
349
Table 4), we believe that the GD detected in this case was generated by the large influx of
350
immigrants into this population.
351
Perithecia of Pygt develop on senescing stems of wheat and other grasses
352
To better understand the role of sexual reproduction in the Pygt life cycle and determine
353
whether the sexual cycle was more likely to occur on wheat or other grass hosts we
354
performed a fruiting experiment and measured the production of proto-perithecia (the
355
primordium that when fertilized develops into a perithecium) and perithecia on different host
356
substrates. The ascocarps formed on autoclaved pieces of wheat stem were indistinguishable
357
from those observed on naturally senescing pieces of stems of wheat and other Poaceae. The
358
proto-perithecia and perithecia developed on the epidermal plant surface and within stems,
359
where they were partially immersed in the internode culm. Proto-perithecia were black or
360
very dark brown and sub-globose shaped. The mature perithecia were black and generally
361
formed long beaks that often came from perithecia that were immersed in the plant tissue (Fig
362
4 and 5). Perithecia showed a mean size of 196 µm in length and 128 µm in width, with
363
average neck size of 243 µm in length and 27 µm in width. Only the proto-perithecia formed
364
on Phalaris canariensis reached a mature size consistent with complete development (Fig 5),
365
suggesting that the sexual cycle was more likely to be completed on P. canariensis than on
366
wheat.
367
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

22
Table 9. Estimates of gametic disequilibrium in populations of Pyricularia graminis-tritici from wheat and other poaceous hosts in Brazil.
368
Population
Number of
monomorphic loci
Locus pairs at significant
disequilibriuma
Bonferroni
correction b
Mat1-1
(%)
Mat1-2
(%)
Ratio
Mat1-1:
Mat1-2
IAc
𝑟̅
𝐷c
pc
2005W
3
5 of 28
0.00183
81.0
19.0
4:1
1.818
0.263
< 0.001
DF-GOW
1
0 of 45
0.00114
100.0
0.0
1:0
-0.191
-0.022
0.850
MGW
2
1 of 36
0.00142
100.0
0.0
1:0
0.177
0.022
0.097
MSW
1
6 of 45
0.00114
77.8
22.2
4:1
0.648
0.073
< 0.001
PRW
1
0 of 45
0.00114
90.9
9.1
10:1
-0.093
-0.011
0.777
RSW
2
1 of 36
0.00142
73.7
26.3
3:1
0.041
0.005
0.365
SPW
3
0 of 28
0.00183
96.3
3.7
26:1
0.039
0.006
0.347
MSP
1
0 of 45
0.00114
23.1
76.9
1:4
0.694
0.079
0.001
PRP
3
0 of 28
0.00183
66.7
33.3
2:1
-0.079
-0.012
0.608
a Pairs of loci at significant disequilibrium according to the Fisher exact test (probability test) implemented by GENEPOP 3.4 [69] at p ≤ 0,05
369
after Bonferroni correction for multiple comparisons [62].
370
b Value of  after Bonferroni correction used for multiple comparisons in the calculation of locus pairs at significant disequilibrium. Initial
371
significance  = 0.05.
372
c IA and 𝑟̅
𝐷 are indexes of multilocus gametic disequilibrium (for the random association of alleles among distinct locus pairs). 𝑟̅
𝐷 is adjusted for
373
the number of loci. The calculation of IA and 𝑟̅
𝐷 and their significance was performed using Multilocus software according Agapow and Burt,
374
[70]. We tested H0 = complete panmixia based on 1,000 randomizations; if p ≤ 0.05 the population is under significant disequilibrium.
375
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

23
The virulence spectra of Pygt populations varied across geographical regions.
376
We examined the virulence spectra for 173 Pygt isolates on both seedlings and detached
377
heads of ten differential wheat cultivars and one barley cultivar. These differentials were
378
chosen based on previous experiments which suggested a gene-for-gene interaction that
379
would allow us to distinguish Pygt pathotypes [39]. Our aim in this analysis was to assess the
380
geographical distribution of virulence groups of Pygt and determine if virulence groups were
381
shared between strains infecting wheat and other grasses. The 173 assessed Pygt isolates,
382
encompassing 80 unique MLMGs, produced typical leaf blast lesions (Fig. 6) and could be
383
grouped into 25 seedling virulence groups (SVGs) (Table 10). These SVGs were named A to
384
Y. SVG L was the predominant group, comprising 47% of the tested isolates. SVG A was the
385
second most frequent group, found in 13% of tested isolates. The 23 remaining SVGs were
386
relatively infrequent (Tables 10 and 11, Fig 7). SVG L was the most widely distributed
387
virulence group across Brazil. The MSP population had the highest number of SVGs (11
388
groups), whereas the PRW and SPW populations had the lowest number of SVGs (1 and 2
389
groups, respectively). Nine SVGs (A, F to I, and K to N) were shared among Pygt isolates
390
originating from wheat and other grasses (Tables 10 and 11).
391
The same isolates fell into nine different head virulence groups (HVGs) when
392
virulence spectra were assessed on detached, mature wheat heads. Five of these HVGs (A to
393
D, and T) had virulence spectra that were identical to the five SVGs (A to D, and T), so we
394
used the same nomenclature for these SVGs and HVGs. The remaining HVGs were
395
designated AA to DD. HVG A was the predominant virulence group, found in 138 isolates,
396
followed by HVG B found in 25 isolates (Table 12). Both of these virulence groups were
397
found in all Pygt populations (Table 13), including the grass-infecting populations. The
398
remaining seven virulence groups were found in only 1 or 2 isolates. As found for the
399
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

24
seedling assay, MSP was the population with the highest number of HVGs (6), and PRW was
400
the population with the lowest number of HVGs (1) (Table 14, Fig 8 and 9).
401
402
Discussion
403
The phylogenetic analyses based on entire genome sequences did not support the
404
earlier hypothesis that two distinct species (named P. graminis-tritici (Pygt) and P. oryzae
405
pathotype Triticum (PoT) in Fig 1) cause wheat blast [15]. Instead, our phylogenetic analyses
406
indicate that Pygt is a single, highly diverse pathogen species with a broad host range that
407
encompasses many grasses that were either native (e.g. Chloris distichophylla, Cynodon spp.,
408
Digitaria insularis) or introduced into Brazil for food production during the last 200 years.
409
Our current phylogenetic analyses do not allow us to determine whether there were multiple
410
origins or a single origin for the wheat blast pathogen, but the absence of strict host
411
specialization among the major sub-clades suggests that the ability to infect wheat may have
412
originated multiple times. All of our findings are consistent with the hypothesis that wheat
413
blast emerged in Brazil through a host shift from the Pygt population infecting other grasses
414
growing near wheat fields, with strong evidence that gene flow still occurs between the Pygt
415
population infecting wheat and the Pygt population infecting other grasses. We hypothesize
416
that this recurring gene flow enables Pygt populations to maintain significant genetic
417
variation on multiple hosts, a finding that stands in stark contrast to what is found for
418
populations of P. oryzae causing rice blast.
419
The microsatellite and virulence datasets revealed that the contemporary Pygt
420
population of Brazil possesses a high degree of genetic and phenotypic diversity. We
421
identified 198 MLMGs and 25 virulence groups among 526 Pygt isolates.
422
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

25
Table 10. Pathogenicity spectra of seedling virulence groups (SVGs) of isolates of Pyricularia graminis-tritici a
423
SVGsb
N
%
Wheat
Barley
Total of
resistant (R)
reactions
Anahuac
75
BR 18
Terena
BR
24
BRS
220
BRS
229
BRS
234
BRS
Buriti
CNT
8
MGS 3
Brilhante
Renan
PFC
2010123
A
22
12.7
S
S
S
S
S
S
S
S
S
S
S
0
B
3
1.7
S
S
S
S
S
S
S
S
S
R
S
1
C
1
0.6
S
S
S
S
S
S
S
S
S
S
R
1
D
1
0.6
S
S
S
S
S
S
S
S
S
R
R
2
E
3
1.7
S
S
S
S
S
S
R
S
S
S
S
1
F
5
2.9
S
S
S
S
S
R
S
S
S
R
S
2
G
13
7.5
S
S
S
S
S
R
R
S
S
R
S
3
H
9
5.2
S
S
S
S
R
R
R
S
S
R
R
5
I
2
1.2
S
S
S
S
S
R
S
S
S
S
S
1
J
3
1.7
S
S
S
S
S
R
R
S
S
S
S
2
K
2
1.2
S
S
S
S
S
S
R
S
S
R
S
2
L
82
47.4
S
S
S
S
R
R
R
S
S
R
S
4
M
2
1.2
S
S
S
S
R
S
R
S
S
R
S
3
N
2
1.2
S
S
S
S
R
R
R
S
S
S
S
3
O
1
0.6
S
S
S
S
R
S
R
S
S
S
S
2
P
1
0.6
S
S
S
S
S
R
S
S
S
S
R
2
Q
2
1.2
S
S
S
S
S
R
R
S
S
R
R
4
R
1
0.6
S
R
S
S
R
R
R
R
S
R
S
6
S
1
0.6
S
R
S
S
R
R
R
S
S
R
S
4
T
5
2.9
S
R
R
R
R
R
R
R
R
R
S
9
U
1
0.6
S
S
S
S
R
S
S
S
S
S
S
1
V
4
2.3
S
R
R
S
R
R
R
R
R
R
S
8
W
4
2.3
S
R
R
S
R
R
R
S
R
R
S
7
X
1
0.6
S
S
S
S
R
R
R
S
R
R
S
5
Y
2
1.2
R
R
R
R
R
R
R
R
R
R
S
10
Total
173
100.0
Total of
R reactions
1
6
4
2
13
16
17
4
5
16
5
.
424
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

26
a Procedures for inoculation assays were previously described [15, 39]. Briefly, spore suspensions (1 × 105 spores ml-1) were uniformly
425
sprayed onto the adaxial leaf surfaces of 15-day-old seedlings at 4-leaf stage (Zadoks' growth stage 14, [71]) until run off. Two pots
426
containing ten plants in the seedling test were inoculated with each of the 173 isolates. Inoculated pots were placed onto plastic trays
427
and incubated in a plant growth chamber at 25 °C. Plants were kept in the dark for the first 24 h, followed by a 12 h dark /12 h
428
fluorescent light regime. Five days after inoculation, disease reactions in response to individual isolates were visually scored based on
429
the percentage of leaf surface showing typical blast symptoms. Upon this, cultivars were classified in qualitative terms either as
430
resistant (R) or susceptible (S). Cultivars were considered R when they showed ≤ 10% of affected areas consistently across inoculation
431
tests’ repetitions and replicates [39]. Experiments were carried out using a two-factor completely randomized balanced design, and the
432
inoculation test was conducted twice.
433
b According to their pathogenicity spectra on the set of cultivars, isolates of P. graminis-tritici were grouped in seedling virulence groups
434
(SVGs), and SVGs were named with letters.
435
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

27
Table 11. Isolates of Pyricularia graminis-tritici assigned to each seedling virulence group (SVG) per population
436
SVG
Population
Total
DF_GOW
MGW
MSW
PRW
RSW
SPW
MSP
PRP
N
%
N
%
N
%
N
%
N
%
N
%
N
%
N
%
N
%
A
1
2.3
3
12.5
5
20.0
8
100.0
1
6.7
.
.
.
.
4
15.4
22
12.7
B
.
.
1
4.2
2
8.0
.
.
.
.
.
.
.
.
.
.
3
1.7
C
.
.
.
.
.
.
.
.
1
6.7
.
.
.
.
.
.
1
0.6
D
.
.
.
.
.
.
.
.
1
6.7
.
.
.
.
.
.
1
0.6
E
.
.
.
.
1
4.0
.
.
.
.
2
33.3
.
.
.
.
3
1.7
F
1
2.3
1
4.2
1
4.0
.
.
2
13.3
.
.
.
.
.
.
5
2.9
G
3
6.8
3
12.5
.
.
.
.
3
20.0
.
.
2
8.0
2
7.7
13
7.5
H
5
11.4
2
8.3
1
4.0
.
.
.
.
.
.
.
.
1
3.8
9
5.2
I
.
.
.
.
.
.
.
.
1
6.7
.
.
1
4.0
.
.
2
1.2
J
2
4.5
.
.
.
.
.
.
1
6.7
.
.
.
.
.
.
3
1.7
K
.
.
.
.
2
8.0
.
.
.
.
.
.
.
.
.
.
2
1.2
L
31
70.5
12
50.0
10
40.0
.
.
5
33.3
4
66.7
5
20.0
15
57.7
82
47.4
M
.
.
.
.
1
4.0
.
.
.
.
.
.
.
.
1
3.8
2
1.2
N
1
2.3
.
.
.
.
.
.
.
.
.
.
.
.
1
3.8
2
1.2
O
.
.
.
.
1
4.0
.
.
.
.
.
.
.
.
.
.
1
0.6
P
.
.
1
4.2
.
.
.
.
.
.
.
.
.
.
.
.
1
0.6
Q
.
.
1
4.2
1
4.0
.
.
.
.
.
.
.
.
.
.
2
1.2
R
.
.
.
.
.
.
.
.
.
.
.
.
1
4.0
.
.
1
0.6
S
.
.
.
.
.
.
.
.
.
.
.
.
1
4.0
.
.
1
0.6
T
.
.
.
.
.
.
.
.
.
.
.
.
5
20.0
.
.
5
2.9
U
.
.
.
.
.
.
.
.
.
.
.
.
1
4.0
.
.
1
0.6
V
.
.
.
.
.
.
.
.
.
.
.
.
4
16.0
.
.
4
2.3
W
.
.
.
.
.
.
.
.
.
.
.
.
3
12.0
1
3.8
4
2.3
X
.
.
.
.
.
.
.
.
.
.
.
.
1
4.0
.
.
1
0.6
Y
.
.
.
.
.
.
.
.
.
.
.
.
1
4.0
1
3.8
2
1.2
Total
44
100.0
24
100.0
25
100.0
8
100.0
15
100.0
6
100.0
25
100.0
26
100.0
173
100.0
Total SVGs
7
.
8
.
10
.
1
.
8
.
2
.
11
.
8
.
25
.
437
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

28
Table 12. Pathogenicity spectra of head virulence groups (HVGs) of isolates of P. graminis-triticia
438
HVGs
N
%
Wheat
Barley
Total of
resistant (R)
reactions
Anahuac
75
BR 18
Terena
BR
24
BRS
220
BRS
229
BRS
234
BRS
Buriti
CNT
8
MGS 3
Brilhante
Renan
PFC
2010123
A
138
79.8
S
S
S
S
S
S
S
S
S
S
S
0
B
25
14.5
S
S
S
S
S
S
S
S
S
R
S
1
C
2
1.2
S
S
S
S
S
S
S
S
S
S
R
1
D
2
1.2
S
S
S
S
S
S
S
S
S
R
R
2
T
1
0.6
S
R
R
R
R
R
R
R
R
R
S
9
AA
2
1.2
S
S
R
S
S
S
S
S
S
S
S
1
BB
1
0.6
S
S
R
S
S
S
R
S
S
R
S
3
CC
1
0.6
S
S
S
S
S
R
S
S
S
R
R
2
DD
1
0.6
S
R
S
R
R
R
R
S
R
R
S
7
Total
173
100.0
Total of R reactions
0
2
3
2
2
3
3
1
2
6
3
-
439
a. Procedures for inoculation assays have been previously described [15, 39]. In short, spore suspensions (1 × 105 spores ml-1) were
440
uniformly sprayed onto the head detached heads harvested from plants between anthesis and initial grain-3 to grain-filling stage
441
(Zadoks' growth stages 63 to 71, [71]) until run off. Three polyurethane foam blocks with ten detached heads apice were inoculated
442
with each of the 173 isolates. Inoculated heads were placed in plastic boxes and incubated in a plant growth chamber at 25 °C. Plants
443
were kept in the dark for the first 24 h, followed by a 12 h dark /12 h fluorescent light regime. Five days after inoculation, disease
444
reactions in response to individual isolates were visually scored based on the percentage of detached head showing typical blast
445
symptoms. Upon this, cultivars were classified in qualitative terms either as resistant (R) or susceptible (S). Cultivars were considered R
446
when they showed ≤ 10% of affected areas consistently across inoculation tests’ repetitions and replicates [39]. Experiments were
447
carried out using a two-factor completely randomized balanced design, and the inoculation test was conducted twice.
448
449
b. According to their pathogenicity spectra on the set cultivars, isolates of P. graminis-tritici were grouped in head virulence groups
450
(HVGs). Groups were named with letters, and the name was maintained when the pathogenicity spectra on seedling and on detached
451
heads coincided.
452
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

29
Table 13. Isolates of Pyricularia graminis-tritici assigned to each head virulence group (HVG) per population
453
HVG
Population
Total
DF_GOW
MGW
MSW
PRW
RSW
SPW
MSP
PRP
N
%
N
%
N
%
N
%
N
%
N
%
N
%
N
%
N
%
A
33
75.0
17
70.8
23
92
8
100
10
66.6
5
83.3
19
76.0
23
88.3
138
79.8
B
10
22.7
6
25.0
2
8
.
.
3
20
1
16.7
2
8.0
1
3.9
25
14.5
C
.
.
.
.
.
.
.
.
1
6.7
.
.
.
.
.
.
2
1.2
D
1
2.3
1
4.2
.
.
.
.
1
6.7
.
.
.
.
.
.
2
1.2
T
.
.
.
.
.
.
.
.
.
.
.
.
.
.
1
3.9
1
0.6
AA
.
.
.
.
.
.
.
.
.
.
.
.
1
4.0
1
3.9
2
1.2
BB
.
.
.
.
.
.
.
.
.
.
.
.
1
4.0
.
.
1
0.6
CC
.
.
.
.
.
.
.
.
.
.
.
.
1
4.0
.
.
1
0.6
DD
.
.
.
.
.
.
.
.
.
.
.
.
1
4.0
.
.
1
0.6
Total
44
100.0
24
100.0
25
100.0
8
100.0
15
100.0
6
100.0
25
100.0
26
100.0
173
100.0
Total HVGs
3
.
3
.
2
.
1
.
4
.
2
.
6
.
4
.
9
.
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

30
We found little differentiation among populations infecting wheat and other grasses,
454
indicating that Pygt is not a wheat-specialized pathogen. Populations separated by more than
455
2000 km were very similar, indicating a high degree of gene flow across large spatial scales
456
and/or high levels of genetic diversity, which would reduce the impact of genetic drift and
457
maintain similar allele frequencies over longer periods. The high gene flow may reflect
458
efficient wind-dispersal of conidia and/or ascospores as well as long distance dispersal on
459
infected seed of wheat and Urochloa [72]. Gametic equilibrium was found among SSR
460
markers in most populations, with both mating types present, though with a predominance of
461
the Mat1-1 idiomorph. These findings, coupled with both high genotype diversity (198
462
MLMGs out of 526 total strains analyzed) and evidence for some clonality, indicate that Pygt
463
has a mixed reproductive system in which cycles of sexual reproduction are followed by the
464
dispersal of locally-adapted clones. The absence of shared MLMGs between populations
465
sampled in 2005 and 2012 suggest that clones do not persist for long periods of time, unlike
466
what has been reported for P. oryzae [73]. Alternatively, very high genetic diversity would
467
make it less likely to find the same MLMGs among populations.
468
Historical analyses of gene flow indicated significant genetic exchange between Pygt
469
populations on wheat and other grasses, with the direction of gene flow predominantly from
470
the population infecting other grasses and into the populations infecting wheat. We
471
hypothesize that the fungal strains capable of infecting both wheat and other grasses can
472
move back and forth between hosts, with recombination occurring mainly on the other
473
grasses and giving rise to the highly diverse Pygt population we observe today. Support for
474
this scenario can be found in previous reports of cross infection and inter-fertility between
475
isolates from wheat and other poaceous hosts [52-54], as well as in the lack of differentiation
476
among wheat- and other Poaceae-adapted populations, the sharing of genotypes and virulence
477
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

31
groups between the two host groups, and the finding of gametic equilibrium consistent with
478
sexual recombination in most populations.
479
The finding that populations of Pygt from wheat and other grasses were not
480
genetically subdivided suggests that several grass species can be hosts for the wheat blast
481
pathogen, unlike the strict host specialization observed for the rice blast pathogen. We
482
hypothesize that Pygt spends most of its life cycle colonizing grass species neighboring or
483
invading the wheat fields affected by wheat blast. We further postulate that sexual
484
recombination takes place mainly or exclusively in these other poaceous hosts, generating
485
most of the genetic diversity observed in the Pygt populations infecting wheat. Other crop
486
pathogens, especially rusts, are also known to undergo sexual recombination on a non-crop
487
host. These hypotheses are consistent with earlier observations that the forage species signal
488
grass (U. brizantha) plays a major role in the genetic variation of the wheat blast pathogen by
489
providing a niche for the fungus to sexually reproduce [15, 54]. Because U. brizantha is a
490
widely grown forage grass occupying more than 90 million ha in Brazil [74], and is often
491
found growing next to wheat fields, we propose that U. brizantha constitutes a major
492
reservoir of wheat blast inoculum and provides a temporal and spatial bridge that connects
493
wheat crops between growing seasons and across the wheat growing areas of Brazil.
494
Virulence phenotyping of 173 Pygt strains differentiated 25 seedling- (SVG) and nine
495
head-virulence groups (HVG). Many wheat cultivars that are resistant to leaf infections are
496
susceptible to head infections, in agreement with the earlier findings [1]. SVG A and HVG A
497
were capable of causing blast on the entire set of tested cultivars. The isolates in these
498
virulence groups form a “super race” that occurs at a relatively high frequency on Brazilian
499
wheat and are also found on Avena sativa (N = 10), U. brizantha (8), Chloris distichophylla
500
(4), Echinochloa crusgalli (4), Rhynchelytrum repens (4), Digitaria sanguinalis (3), Eleusine
501
indica (2), Eragrostis plana (2), Cenchrus echinatus, Cynodon spp., D. insularis, Panicum
502
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

32
maximum, and S. sudanense.
503
The closely related rice blast pathogen P. oryzae is often presented as a model for
504
understanding wheat blast. P. oryzae populations are considered strictly asexual [75], except
505
for rare sexual populations of P. oryzae associated with rice in South-eastern Asia (the origin
506
of rice domestication, and the proposed center of origin for rice blast) [73, 76], and the
507
population associated with finger millet (Eleusine coracana) in West Africa. The Pyricularia
508
population adapted to finger millet is probably a new Pyricularia species distinct from P.
509
oryzae, with a center of origin in western Kenya and north-eastern Uganda [77]. However, it
510
is yet to be reclassified. Remarkably, sexual perithecia have not been found in the field for
511
either of these sexual populations, illustrating the challenge of proving a population is sexual
512
even when it exhibits the population genetic "signature of sex" composed of gametic
513
equilibrium among neutral markers, low clonality and mating types at equal frequencies. As
514
was the case for the sexual Pyricularia populations on rice in Southeast Asia and on finger
515
millet in West Africa, we have not yet found natural perithecia of Pygt in Brazilian wheat
516
fields, but we have abundant population genetic and biological evidence that strongly indicate
517
the occurrence of sexual Pygt populations in Brazil.
518
Our biological evidence for sexual reproduction is the formation of proto-perithecia
519
and perithecia of Pygt on autoclaved wheat stems and on senescing stems of wheat and other
520
grasses. Moreira [78] conducted similar experiments by injecting stems of living wheat plants
521
with the same sexually compatible isolates. In that experiment, no sexual structures were
522
produced in living plant tissues [78]. These contrasting results suggest that senescent plant
523
tissues are necessary to stimulate sexual reproduction in Pygt. The same pattern emerged
524
when sexually compatible isolates of P. oryzae were placed on living rice plants: perithecia
525
formation occurred only in senescent or detached leaf sheaths [79].
526
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

33
While perithecia produced in our assays did not harbor detectable asci and ascospores,
527
the induction of sexual structures in Ascomycetes is known to be affected by many factors
528
including substrate, light, temperature, and the availability of female fertile strains [80]. We
529
hypothesize that the lack of ascospore production in our assays results from one or more of
530
these factors. We suggest that future studies aiming to identify perithecia of Pygt in the field
531
should focus on poaceous hosts such as Phalaris canariensis that support the development of
532
fully formed perithecia.
533
Based on all of the existing knowledge of Pygt biology and epidemiology, we propose
534
a provisional disease cycle for wheat blast (Fig 10). At the end of a cropping season (Ae), ear
535
infections lead to infected seed (B, C), providing inoculum for both local and long distance
536
dispersal of the pathogen [72]. Crop residues left in the field after harvest provide a niche for
537
Pygt sexual reproduction (D, 1-4); the resulting perithecia release airborne ascospores (D1)
538
that create new genotypes that can cause new infections locally or in distant host populations
539
by the germination of terminal cells (D2), which is followed by fungal vegetative growth and
540
subsequent conidiogenesis (D3) [81]. The asexual conidia produced in the resulting infection
541
are released (D4) and provide airborne inoculum for leaf infection on other grasses located
542
within or next to wheat fields (E, F) [1, 53, 82]. Perithecia can also form in other infected
543
poaceous hosts and on major pasture grasses, with the resulting ascospores falling onto
544
nearby wheat crops (E). Seedborne inoculum (B, C) results in primary infections in newly
545
established wheat crops. (F4) conidia released from leaf blast lesions on other poaceous hosts
546
growing near wheat fields can also contribute inoculum leading to blast on wheat ears [1, 53].
547
Conidia production on leaves (Af) in the lower canopy of some wheat cultivars can coincide
548
with spike emergence in the field and provide an important source of inoculum for wheat
549
blast epidemics on ears (Ae) [83].
550
In summary, our experiments showed that Brazilian Pygt populations maintain very
551
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

34
high levels of genetic diversity and are able to infect a surprisingly wide array of grass hosts.
552
Pygt populations exhibit a mixed reproductive system and are characterized by high levels of
553
gene flow over long distances. There is evidence for substantial genetic exchange between
554
Pygt populations infecting wheat and Pygt populations infecting nearby grasses. This
555
combination of properties is likely to make wheat blast a particularly difficult disease to
556
control. We hypothesize that the majority of sexual recombination is occurring on nearby
557
poaceous hosts and that Urochloa brizantha, as the major pasture grass in Brazil, plays an
558
important role as a host that provides a steady source of inoculum that connects wheat crops
559
across Brazil.
560
561
Material and methods
562
Population sampling. A total of 556 isolates of Pyricularia spp. were characterized in this
563
study, comprising ten regional populations sampled from wheat or other poaceous hosts. 526
564
of these isolates were found to be Pygt while 30 isolates were found to be different
565
Pyricularia species. Six populations of Pygt (387 isolates) were collected from symptomatic
566
heads during the 2012 and 2013 cropping seasons in naturally infected wheat fields in Rio
567
Grande do Sul (RSW), Paraná (PRW), Mato Grosso do Sul (MSW), São Paulo (SPW), Minas
568
Gerais (MGW), Goiás and the Federal District (DF-GOW). The isolates from Distrito Federal
569
and Goiás were grouped into a single population because these locations comprise a single
570
cropping region. Pygt strains from wheat fields were sampled along transects as described
571
previously [26]. A seventh Pygt population was composed of 79 isolates with distinct
572
multilocus SSR genotypes representing the Pygt diversity found in the major Brazilian
573
wheat-growing areas in 2005 [39] (Table 1, Supplementary Table 1).
574
Two additional Pygt populations comprised isolates sampled from other poaceous hosts
575
commonly growing as invasive grasses or weeds located within or nearby wheat fields. The
576
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

35
two populations from other poaceous hosts (60 isolates) were sampled from within or nearby
577
three wheat fields in Londrina County, Paraná state (PRP), and six wheat fields in Dourados
578
County in Mato Grosso do Sul state (MSP). For each field, infected leaves were sampled from
579
invasive grass species exhibiting typical blast symptoms located either within the wheat field
580
or less than 100 m from the edge of the wheat field. The Poaceae species sampled included:
581
Avena sativa, Cenchrus echinatus, Chloris distichophylla, Cynodon spp., Digitaria insularis,
582
Digitaria sanguinalis, Echinochloa crusgalli, Eleusine indica, Eragrostis plana, Panicum
583
maximum, Rhynchelytrum repens, Sorghum sudanense, and Urochloa brizantha.
584
585
Inference of genealogical relationships among Pyricularia graminis-tritici and other
586
Pyricularia species.
587
We performed population genomics analyses using single nucleotide polymorphisms (SNPs)
588
across the genome. For the population genomic analyses, the samples included 47 rice blast-
589
associated P. oryzae strains with publically available genome sequences, 32 Brazilian strains
590
of P. graminis-tritici sampled from wheat and other poaceous hosts, two isolates of P. oryzae
591
from Hordeum vulgare, two isolates of P. grisea from Digitaria sanguinalis, two isolates of
592
Pyricularia spp. from Setaria italica and Eleusine indica, one isolate resulting from a cross
593
between K76-79 (from weeping lovegrass, Eragrostis curvula) and WGG-FA40 (from finger
594
millet, Eleusine coracana) and four wheat blast transcriptome samples collected in
595
Bangladesh in spring 2016 [31]. Among the 32 Brazilian Pygt strains sampled between 2005
596
and 2013, 22 were wheat-infecting strains included in an earlier analysis to infer the origin of
597
wheat blast in Bangladesh [31] and 10 were new blast strains sampled from other grasses and
598
included in this paper. Transcriptomic (RNA) SNPs were identified based on short read
599
alignments against the P. oryzae reference genome 70-15, available at Ensembl Fungi
600
(http://fungi.ensembl.org/Magnaporthe_oryzae/Info/Index). For all the completely sequenced
601
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

36
genomes, we used Bowtie version 2.2.6 [84] to align quality-trimmed Illumina short read
602
data against the reference genome 70-15. Quality-trimmed Illumina short read data generated
603
from RNA from the Bangladesh transcriptomic samples were mapped using TopHat version
604
2.0.14 [85]. The variants in the genomes of the different strains were identified using the
605
Genome Analysis Toolkit (GATK) version 3.5 available at the Broad Institute
606
(https://software.broadinstitute.org/gatk/) [86]. A two-step variant calling was used following
607
the GATK best practice guidelines. Firstly, raw variants with local reassembly of read data
608
were called using Haplotype Caller. All the raw variant calls and filtration were jointly
609
genotyped using the GATK Genotype GVCFs. Secondly, SelectVariants was used to subset
610
the variant calls to contain only SNPs. Finally, we applied SNPs hard-filters to remove low-
611
quality SNPs using the following criteria: QUAL ≥ 5000.0, QD ≥ 5.0, MQ ≥ 20.0, – 2.0 ≤
612
ReadPosRankSum ≤ 2.0, –2.0 ≤ MQRankSum_upper ≤ 2.0, –2.0 ≤ BaseQRankSum ≤ 2.0.
613
Furthermore, we used vcftools (https://vcftools.github.io) to generate a SNP dataset for
614
phylogenomic analyses. To avoid biases in the phylogenetic reconstruction, we only retained
615
SNPs that were called in at least 90% of all analyzed strains. Furthermore, we retained a SNP
616
only if the SNP was called in the best-sequenced Bangladesh sample 12, as described
617
previously [31] (Supplementary Table S1). We retained 55,041 informative SNPs. A
618
maximum likelihood phylogeny was constructed from a SNP supermatrix using RAxML
619
version 8.2.8 (http://www.exelixis-lab.org) with a GTR substitution matrix and 100 bootstrap
620
replicates.
621
622
Microsatellite genotyping and fragment analyses. 526 Pyricularia isolates (Table 1,
623
Supplementary table 1) were genotyped for 11 microsatellite loci (cnpt_mg-c013Tri, -c047, -
624
c060, -c065, -c108, -c129, -c147, -c168, -c233, -c248, and -p1e11) as described earlier [39,
625
87] (Supplementary Table 2). Briefly, amplifications were performed in a thermal cycler with
626
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

37
conditions as follows: initial denaturation at 95°C for 3 min; followed by 35 cycles of 95°C
627
for 25 s, 55°C or 60°C for 25 s, and 72°C for 25 s; with a final extension of 72°C for 15 min.
628
PCR reactions were diluted and combined in three sets for electrophoresis (Set 1: cnpt-mg-
629
c047, -c065, -c248, and -p1e11; Set 2: cnpt-mg-c013Tri, -c060, -c147, and -c168; and Set 3:
630
cnpt-mg-c108, -c129, and -c233). Isolates 12.1.111 and 10880 were included as controls in
631
every run of 93 samples. The fluorescent-labeled PCR products, along with a size standard
632
were separated on an ABI 3730xl capillary sequencer. The fragment analysis for detection
633
and discrimination among allele sizes was performed using Geneious R 9.1.5.
634
635
Analyses of population genetic structure. SSR datasets were used to calculate gene and
636
genotype diversity and genetic differentiation among populations, generate minimum
637
spanning networks among genotypes, and estimate contemporary patterns of migration and
638
gene flow. We inferred the predominant reproductive mode based on tests of gametic
639
equilibrium and frequencies of the mating type idiomorphs Mat1-1 and Mat1-2. Except for
640
the analyses of genotypic diversity, all analyses used clone-corrected datasets in which only
641
one individual from each multilocus microsatellite genotype was included per population.
642
643
Genotypic and genetic diversity and allelic richness. The multilocus microsatellite
644
genotype (MLMG) for each isolate was determined using Genodive v. 2.0b7 [57]. Isolates
645
exhibiting the same MLMG were considered clones. A minimum spanning network (MSN)
646
was constructed to show the distribution and genetic similarity among the MLMGs of Pygt
647
found in the nine populations. The MSN was constructed with the bruvo.msn distance
648
function [88] and the Prim algorithm of the igraph package, [89] using the poppr package
649
[90] in the R environment [91].
650
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

38
Measures of genotypic diversity included: a) number of MLMGs per population; b)
651
population-specific MLMGs; c) clonal fraction calculated as 1-(number of MLMGs)/(total
652
number of isolates); d) effective number of MLMGs (GO) [92]; and e) the evenness, an
653
indicator for how evenly the genotypes were distributed in the population, calculated as the
654
ratio of the effective number of distinct MLMGs scaled by the maximum number of expected
655
MLMGs. We tested the statistical significance of differences in genotypic diversity between
656
pairs of populations based on 1,000 bootstrap resamplings matching the size of the smallest
657
population (19 individuals) [61]. Allelic richness was estimated for each population as the
658
average number of alleles per locus using rarefaction [58]. To test whether populations
659
differed in allelic richness, p values for the significance of the pairwise comparisons were
660
obtained by 1,000 permutations. These calculations were computed using FSTAT v. 2.9.3.2
661
[93]. The probability of identical genotypes arising from sexual reproduction and random
662
mating and it is identical to the genotype probability was estimated with the Pgen index
663
previously described with GenAlEx v6.501 software [55, 56]
664
665
Population differentiation. AMOVA [94] was used to assess the distribution of gene
666
diversity and the degree of differentiation among geographical populations of the pathogen.
667
Populations were also grouped according to the host of origin. Degrees of differentiation
668
were compared using orthogonal contrasts. The sum of squared size differences (RST) was
669
used as the distance measure between two haplotypes [60]. The significance of the fixation
670
indexes was tested using 1,023 permutations by a nonparametric approach [94] at  = 0.05
671
after Bonferroni correction for multiple comparisons [62]. All calculations were carried out
672
with the program ARLEQUIN v. 3.11 [59].
673
674
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

39
Assessment of historical migration and demographic parameters. For migration analyses,
675
populations were grouped according to their host of origin. A maximum likelihood test based
676
on MCMC [68] was used to test four different models of migration between the populations
677
obtained from wheat and from other Poaceae. The migration models tested were: a) complete
678
panmixia; b) bidirectional; c) directional, with migration occurring from the wheat population
679
towards the other Poaceae population; and d) directional (inverse) with migration occurring
680
from the population obtained from other Poaceae towards the wheat population. Estimates of
681
gene flow were obtained using five runs, and the run with the highest likelihood was chosen
682
to represent each migration model. Then the likelihood values of the four migration models
683
were compared to select the one that best fit the data based on the Log of the Bayes Factor
684
(LBF). LBF was calculated as 2 [ln(Prob(Data | ModelX)) – ln (Prob(Data | best of the four
685
models))]; higher LBF values reflect better fits of the migration model to the data [64, 66].
686
For all migration analyses the data type chosen was microsatellite data with Brownian
687
motion and assuming a stepwise mutation model. Each of the five runs had ten short initial
688
chains, one long final chain, a static heating scheme with five temperatures (1, 100, 1000,
689
10,000 and 100,000), and swapping interval of 1. The initial chains were performed with
690
500-recorded steps, a sampling increment of 100, with 2,500 trees recorded per short sample.
691
The long chain was carried out with 8,334-recorded steps, a sampling increment of 500, six
692
concurrent chains (replicates) and 500 discarded trees per chain (burn-in). The final number
693
of sampled parameter values was 25,002,000 iterations. The values and confidence intervals
694
for the migration rate (M), and the effective population size (θ = 2Neμ for haploids, where Ne
695
= effective population size and μ = mutation rate inferred for each locus) were calculated
696
using a percentile approach. Migration analyses were implemented in MIGRATE-n v. 3.6.11
697
[64] at the CIPRES Science Gateway [63].
698
699
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

40
Tests for gametic equilibrium. Gametic equilibrium was assessed using a multilocus
700
association test (10). The hypothesis that genotypes at one locus are independent from
701
genotypes at another locus was tested using Fisher’s exact test at α = 0.05 and an MCMC
702
algorithm (with 1,000 batches and 1,000 iterations/batch) implemented using the program
703
GENEPOP v.3.4 [69]. The Bonferroni correction was applied to this analysis to avoid false
704
rejections of the null hypothesis due to the large number of comparisons performed [62]. Two
705
loci were in gametic equilibrium when their associated p value was not significant (p > 0.05).
706
We also measured the indexes of multilocus association (IA and r̅D) for each Pygt population
707
using Multilocus software ver 1.3b, according to Agapow and Burt [70].
708
709
Determination of mating type idiomorphs. The mating type idiomorph, Mat1-1 or Mat1-2
710
[95], was determined for each strain using a PCR assay [39]. To amplify Mat1-1, the primers
711
were A1:5’-AGCCTCATCAACGGCAA-3’ and A5: 5’-GGCACGAACATGCGATG-3’. For
712
Mat1-2 they were B15: 5’-CTCAATCTCCGTAGTAG-3’ and B16: 5’-
713
ACAGCAGTATAGCCTAC-3’. We included isolate Py46.2 as a positive control for Mat1-1
714
and a negative control for Mat1-2, and isolate Py5003 as a positive control for Mat1-2 and a
715
negative control for Mat1-1 [39].
716
717
Development of Pygt perithecia on senescing stems from several poaceous hosts.
718
Pygt strains Py33.1 (Mat1-1) and Py05046 (Mat1-2) were shown to be fertile in earlier
719
studies [39, 78]. The production of perithecia and asci on autoclaved wheat stems and
720
naturally senescing stems of wheat and other grasses was assessed after co-inoculation with
721
these strains. The other poaceous hosts assayed were: Avena strigosa (black oats) cv.
722
Embrapa 29 Garoa; Hordeum vulgare (barley) cvs. BR Elis and MN 743; Oryza sativa cvs.
723
BRS Primavera, BRSMG Relampago and Yin Lu 30 (red rice); Phalaris canariensis (canary
724
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

41
grass); Secale cereale (rye) cv. BR1; Setaria italica (foxtail millet); Triticum aestivum
725
(wheat) cvs. BRS 264 and MGS Brilhante; Triticale (xTriticosecale) cv. IAC Caninde;
726
Urochloa hybrid cv. Mulato (Urochloa ruziziensis x U. decumbens x U. brizantha). The
727
wheat cv. MGS Brilhante is classified as moderately resistant to wheat blast, while the other
728
wheat cultivar, barley, Urochloa spp., and oats are considered susceptible to wheat blast. In
729
contrast, rice cultivars are resistant to Pygt [15, 39]. The remaining hosts included in this
730
experiment have unknown susceptibility to Pygt.
731
Spores of isolates Py33.1 and Py05046 were harvested after 14 days of growth on
732
oatmeal agar [39] and combined in equal proportions at 1x104 conidia ml-1 for co-inoculation
733
as described earlier, with minor modifications [79]. Wheat stems consisted of 4-cm sections
734
collected from one-month old plants and autoclaved at 121°C for 20 min. Autoclaved wheat
735
stems or naturally senescing stems were placed in 90 mm Petri dishes containing water agar
736
(agar, 15g l-1) and were inoculated by injection of 0.3 mL of the spore mix. Inoculated
737
materials were kept in a growth chamber at 25⁰C under a 12 h dark /12 h fluorescent white
738
light photoperiod for 7 days. Subsequently, for perithecia development, the temperature was
739
lowered to 20⁰C and the samples were incubated for another 21 days (autoclaved stem
740
sections) or one month (senescing stem pieces) under the same photoperiod. The assays were
741
replicated once, with five repeats of each experimental unit each time. The development of
742
sexual structures was documented using light and scanning electron microscopes. The density
743
of proto-perithecia or perithecia on plant debris and on sections of senescing stems was
744
determined by analyzing at least three areas of approximately 0.5 mm2 on each plant species.
745
746
Virulence spectrum of Pygt on wheat seedlings and detached heads. The virulence spectra
747
of 173 isolates of Pygt representing 80 MLMG were assessed on seedlings and detached
748
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

42
heads of ten wheat cultivars and one barley cultivar. Within each MLMG, isolates were
749
selected at random from the eight populations sampled in 2012-2013, including 121 isolates
750
from wheat and 52 isolates from other poaceous hosts. The wheat cultivars included in the
751
tests were: Anahuac 75 (susceptible control), BR 18, BR 24, BRS 220, BRS 229, BRS 234,
752
BRS Buriti, CNT 8, MGS 3 Brilhante, Renan, and barley cv. PFC 2010123.
753
Detailed procedures for inoculum preparation, inoculation, incubation, disease
754
assessment and data analysis were described earlier [15, 39]. Briefly, inoculations were
755
conducted on 15-day-old seedlings at the 4-leaf stage and on detached heads harvested from
756
plants after anthesis. Seedling and head inoculation experiments were conducted using a two-
757
factor completely randomized balanced design. Two pots containing ten plants each were
758
used for the seedling test, while three foam blocks with ten detached heads apiece were
759
inoculated for each of the 173 isolates. Each inoculation test was conducted twice. In both
760
tests, disease was scored 5 days after inoculation. Cultivars were classified as resistant (R) or
761
susceptible (S) based on visual assessment of the percentage of leaf or detached head
762
showing typical blast symptoms. Pygt isolates were placed into seedling virulence groups
763
(SVGs) and head virulence groups (HVGs) according to their pathogenicity spectra on each
764
wheat cultivar.
765
766
Funding
767
This work was funded by FAPESP (São Paulo Research Foundation, Brazil) research grants
768
to P.C. Ceresini (2013/10655-4 and 2015/10453-8), EMBRAPA-Monsanto research grant
769
(Macroprogram II-02.11.04.006.00.00) to J.L.N. Maciel, and research grants from FINEP
770
(Funding Authority for Studies and Projects, Brazil) and FAPEMIG (Minas Gerais Research
771
Foundation, Brazil) to E. Alves (CAG-APQ-01975-15). P.C. Ceresini and E. Alves were
772
supported by research fellowships from Brazilian National Council for Scientific and
773
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

43
Technological Development - CNPq (Pq-2 307361/2012-8 and 307295/2015-0). V. L.
774
Castroagudin was supported by a Post-Doctorate research fellowship FAPESP (PDJ
775
2014/25904-2, from 2015–2016). S.I. Moreira was supported by a Postdoctoral researcher
776
fellowship PNPD from CAPES (Higher Education Personnel Improvement Coordination,
777
Brazil). A.L.D. Danelli was supported by a Doctorate research fellowship CAPES-PROSUP
778
(Programa de Suporte à Pós-Graduação de Instituições de Ensino Particulares, Brazil). We
779
thank CAPES for sponsoring the establishment of the ‘Centro de Diversidade Gentica no
780
Agroecossistema’ (Pro-equipamentos 775202/2012). Authorization for scientific activities #
781
39131-3 from the Brazilian Ministry of Environment (MMA) / ‘Chico Mendes’ Institute for
782
Conservation of Biodiversity (ICMBIO) / System for Authorization and Information in
783
Biodiversity (ICMBIO). DC is supported by the Swiss National Science Foundation (grant
784
31003A_173265). BAM is supported by the Swiss National Science Foundation (grant
785
31003A_155955) and the Bundesamt für Landwirtschaft (BLW Project PGREL-NN-0034).
786
787
Acknowledgements:
788
Primer sequences for MAT loci were provided by Didier Tharreau, INRA, Montpellier,
789
France.
790
791
Supporting information
792
S1 Table. Isolates of Pyricularia species included in the inference of genealogical
793
relationships. This table lists and describes all the isolates included in the inference of
794
genealogical relationships among wheat blast samples of Pyricularia graminis-tritici and
795
several other blast samples.
796
797
S2 Table. Isolates of Pyricularia species analyzed in this study. This table lists and
798
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

44
describes all the isolates examined in this study, including their original host, year and
799
location of sampling, mating type, multilocus microsatellite genotype, alleles found in 11
800
microsatellites loci, and seedling and head virulence group for each isolate.
801
802
S3 Table. Oligonucleotides. This table lists all primers for microsatellite loci and their
803
sequences in this study.
804
805
Author Contributions
806
Conceptualization: PCC, JLNM, EA, BAM
807
Data Curation: VLC, ALDD, JTAR, ALVB, CAF, JLNM, SIM, PCC, EA, DC
808
Formal Analysis: VLC, PCC, SIM, DC
809
Funding Acquisition: JLNM, EA, PCC, BAM, DC
810
Investigation: VLC, ALDD, SIM, JTAR, GC, JLNM, PCC, DC
811
Methodology: VLC, ALDD, SIM, GC, ALVB, JLNM, PCC, DC
812
Project Administration: VLC, JLNM, PCC, EA, DC
813
Resources: PCC, JLNM, EA, DC, BAM
814
Supervision: PCC, JLNM, CAF, BAM
815
Validation: VLC, ALDD, ALVB, SIM, JTAR, JLNM, EA, PCC, DC
816
Visualization: GC, PCC
817
Writing-Original Draft Preparation: VLC, SIM, PCC
818
Writing-Review & Editing: VLC, JNM, BAM, SIM, JLNM, PCC, DC
819
820
References
821
822
1. Urashima AS, Kato H. Pathogenic relationship between isolates of Pyricularia grisea of
823
wheat and other hosts at different host developmental stages. Fitopatol Bras. 1998; 23:30-5.
824
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

45
2. Takabayashi N, Tosa Y, Oh HS, Mayama S. A gene-for-gene relationship underlying the
825
species-specific parasitism of Avena/Triticum isolates of Magnaporthe grisea on wheat
826
cultivars. Phytopathology. 2002; 92:1182-8.
827
3. Murakami J, Tomita R, Kataoka T, Nakayashiki H, Tosa Y, Mayama S. Analysis of host
828
species specificity of Magnaporthe grisea toward foxtail millet using a genetic cross between
829
isolates from wheat and foxtail millet. Phytopathology. 2003; 93:42-5. doi:
830
10.1094/PHYTO.2003.93.1.42.; PMID: 18944155.
831
4. Couch BC, Fudal I, Lebrun M-H, Tharreau D, Valent B, van Kim P, et al. Origins of host-
832
specific populations of the blast pathogen Magnaporthe oryzae in crop domestication with
833
subsequent expansion of pandemic clones on rice and weeds of rice. Genetics. 2005;
834
170:613-30. doi: http://dx.doi.org/10.1534/genetics.105.041780. PubMed PMID:
835
PMC1450392.
836
5. Couch BC, Kohn LM. A multilocus gene genealogy concordant with host preference
837
indicates segregation of a new species, Magnaporthe oryzae, from M. grisea. Mycologia.
838
2002; 94:683-93. PMID: 21156541.
839
6. Hirata K, Kusba M, Chuma I, Osue J, Nakayashiki H, Mayama S, et al. Speciation in
840
Pyricularia inferred from multilocus phylogenetic analysis. Mycol Res. 2007; 111:799-808.
841
doi: 10.1016/j.mycres.2007.05.014. PMID: 17656080.
842
7. Klaubauf S, Tharreau D, Fournier E, Groenewald JZ, Crous PW, de Vries RP, et al.
843
Resolving the polyphyletic nature of Pyricularia (Pyriculariaceae). Stud Mycol. 2014;
844
79:85-120. doi: 10.1016/j.simyco.2014.09.004. PMID: 25492987
845
8. Huyse T, Poulin R, Theron A. Speciation in parasites: a population genetics approach.
846
Trends Parasitol. 2005; 21:469-75. doi: 10.1016/j.pt.2005.08.009. PMID: 16112615.
847
9. Stukenbrock EH, Banke S, Javan-Nikkhah M, McDonald BA. Origin and domestication of
848
the fungal wheat pathogen Mycosphaerella graminicola via sympatric speciation. Mol Ecol
849
Evol. 2007; 24:398-411. doi: 10.1093/molbev/msl169. PMID: 17095534.
850
10. Zaffarano PL, McDonald BA, Linde CC. Rapid speciation following recent host shifts in
851
the plant pathogenic fungus Rhynchosporium. Evolution. 2008; 62:1418-36.
852
11. Baker CJ, Harrington TC, Krauss U, Alfenas AC. Genetic variability and host
853
specialization in the Latin American clade of Ceratocystis fimbriata. Phytopathology. 2003;
854
93:1274-84. doi: doi:10.1094/PHYTO.2003.93.10.1274. PubMed PMID: 18944327.
855
12. Bucheli E, Gautschi B, Shykoff JA. Host-specific differentiation in the anther smut
856
fungus Microbotryum violaceum as revealed by microsatellites. J Evol Biol. 2000; 13:188-98.
857
doi: 10.1046/j.1420-9101.2000.00160.x.
858
13. Kato H, Yamamoto M, Yamaguchi-Ozaki T, Kadouchi H, Iwamoto Y, Nakayashiki H, et
859
al. Pathogenicity, mating ability and DNA restriction fragment length polymorphisms of
860
Pyricularia populations isolated from Gramineae, Bambusideae and Zingiberaceae plants. J
861
Gen Plant Pathol. 2000; 66:30-47. doi: 10.1007/PL00012919.
862
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

46
14. Stukenbrock EH, McDonald B. The origins of plant pathogens in agro-ecosystems. Ann
863
Rev Phytopatol. 2008; 46:75-100. doi: 10.1146/annurev.phyto.010708.154114. PMID:
864
18680424
865
15. Castroagudín VL, Moreira SI, Pereira DAS, Moreira SS, Brunner PC, Maciel JLN, et al.
866
Pyricularia graminis-tritici, a new Pyricularia species causing wheat blast. Persoonia. 2016;
867
37:199-216. doi: 10.3767/003158516X692149.
868
16. Igarashi S, Utiamada CM, Igarashi LC, Kazuma AH, Lopes RS. Pyricularia em trigo. 1.
869
Ocorrência de Pyricularia sp. no estado do Paraná. Fitopatol Bras. 1986; 11:351-2
870
17. Anjos JRND, Silva DBD, Charchar MJD, Rodrigues GC. Ocurrence of blast fungus
871
(Pyricularia grisea) on wheat and rye in the savanna region of Central Brazil. Pesq Agropec
872
Bras. 1996; 31:79-82.
873
18. Maciel JLN. Magnaporthe oryzae, the blast pathogen: current status and options for its
874
control. Plant Sci Rev. 2011; 2011:233-40. doi: 10.1079/PAVSNNR20116050.
875
19. Gutiérrez SA, Cúndom MA. Pyricularia oryzae affecting barley crops in Corrientes
876
(Argentina). Summa Phytopathol. 2015; 41:318-20. doi: 10.1590/0100-5405/2063.
877
20. Verzignassi RS, Poltronieri LS, Benchimol RL, França SKS, Carvalho EA, Fernandes
878
CD. Pyricularia grisea: new pathogen on Brachiaria brizantha cv. Marandu in Pará. Summa
879
Phytopathol. 2012; 38:254. doi: 10.1590/S0100-54052012000300016
880
21. Marangoni MS, Nunes MP, Fonseca N, Mehta YR. Pyricularia blast on white oats: a new
881
threat to wheat cultivation. Trop Plant Pathol. 2013; 38:198-202. doi: 10.1590/S1982-
882
56762013005000004.
883
22. Reges JTA, Negrisoli MM, Dorigan AF, Castroagudín VL, Maciel JLN, Ceresini PC.
884
Pyricularia pennisetigena and P. zingibericola from invasive grasses infect signal grass,
885
barley and wheat. Pesq Agropec Trop. 2016; 46:206-14. doi: 10.1590/1983-
886
40632016v4641335.
887
23. Martins TD, Lavorenti NA, Urashima AS. Methods to examine transmission of
888
Pyricularia grisea from seeds to seedlings of triticale. Fitopatol Bras. 2004; 29:425-8. doi:
889
10.1590/S0100-41582004000400011.
890
24. Crous PW, Wingfield MJ, Burgess TI, Hardy GESJ, Crane C, Barrett S, et al. Fungal
891
Planet description sheets: 469–557. Persoonia. 2016; 37:218-403. doi:
892
10.3767/003158516X694499.
893
25. Silva CP, Nomura E, Freitas EG, Brugnaro C, Urashima AS. Efficiency of alternative
894
treatments in the control of Pyricularia grisea in wheat seeds. Trop Plant Pathol. 2009; 34:
895
127-31. doi: 10.1590/S1982-56762009000200009
896
26. Castroagudín VL, Ceresini PC, Oliveira SC, Reges JTA, Maciel JLN, Bonato ALV, et al.
897
Resistance to QoI fungicides is widespread in Brazilian populations of the wheat blast
898
pathogen Magnaporthe oryzae. Phytopathology. 2015; 104:284-94. doi: 10.1094/PHYTO-06-
899
14-0184-R. PMID: 25226525.
900
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

47
27. Kohli MM, Mehta YR, Guzman E, Vierdma L, Cubilla LE. Pyricularia blast - a threat to
901
wheat cultivation. Czech J Genet Plant Breed. 2011; 47:S130-S4.
902
28. Duveiller E, Hodson D, Tiedmann A, editors. Wheat blast caused by Magnaporthe
903
grisea: a reality and new challenge for wheat research. 8th Int Wheat Conf; 2010 2010; St.
904
Petersburg. Vavilov Res. Inst. Plant Ind.: Vavilov Research Institute of Plant Industry.
905
29. Callaway E. Devastating wheat fungus appears in Asia for first time. Nature. 2016;
906
532:421-2. doi: 10.1038/532421a. PMID: 27121815
907
30. Croll D. The origin of wheat blast in Bangladesh 2016 [cited September 29, 2016.].
908
Available from: Available from: https://github.com/crolllab/wheat-blast.
909
31. Islam MT, Croll D, Gladieux P, Soanes DM, Persoons A, Bhattacharjee P, et al.
910
Emergence of wheat blast in Bangladesh was caused by a South American lineage of
911
Magnaporthe oryzae. BMC Biology. 2016; 14:84. doi: 10.1186/s12915-016-0309-7. PMID:
912
27716181.
913
32. Chaudhuri SR. Tackling wheat blast: Bengal government bans wheat cultivation within 5
914
km of Bangladesh border. Hindustan Times. 2017 19th July, 2017.
915
33. Press Trust of India. Deadly wheat blast disease spreads to Bengal districts. The New
916
Indian Express. 2017 9th March, 2017.
917
34. McTaggart AR, van der Nest MA, Steenkamp ET, Roux J, Slippers B, Shuey LS, et al.
918
Fungal genomics challenges the dogma of name-based biosecurity. PLoS Pathogens. 2016;
919
12:e1005475. doi: 10.1371/journal.ppat.1005475. PMID: 27149511.
920
35. Goulart ACP, Sousa PG, Urashima AS. Damages in wheat caused by infection of
921
Pyricularia grisea. Summa Phytopathol. 2007; 33:358-63. doi: 10.1590/S0100-
922
54052007000400007
923
36. Goulart ACP, Paiva FA, Andrade PJM. Qualidade sanitária de sementes de trigo
924
produzidos no Mato Grosso do Sul, safras 1987 a 1992. Summa Phytopathol. 1995; 21:235-8.
925
37. Urashima AS, Igarashi LC, Kato H. Host range, mating type, and fertility of Pyricularia
926
grisea from wheat in Brazil. Plant Dis. 1993; 77:1211-6.
927
38. Ceresini PC. Estrutura genética de populações do patógeno da brusone do trigo
928
Magnaporthe oryzae no Brasil. Ilha Solteira: Universidade Estadual Paulista (UNESP); 2011.
929
39. Maciel JLN, Ceresini PC, Castroagudin VL, Kema GHJ, McDonald BA. Population
930
structure and pathotype diversity of the wheat blast pathogen Magnaporthe oryzae 25 years
931
after its emergence in Brazil. Phytopathology. 2014; 104:95-107. doi: 10.1094/PHYTO-11-
932
12-0294-R. PMID: 23901831.
933
40. Inoue Y, Vy TTP, Yoshida K, Asano H, Mitsuoka C, Asuke S, et al. Evolution of the
934
wheat blast fungus through functional losses in a host specificity determinant. Science. 2017;
935
357:80-3. doi: 10.1126/science.aam9654. PMID: 28684523.
936
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

48
41. McDonald BA, Linde C. The population genetics of plant pathogens and breeding
937
strategies for durable resistance. Euphytica. 2002; 124:163-80. doi:
938
10.1023/A:1015678432355.
939
42. McDonald BA, Linde C. Pathogen population genetics, evolutionary potential, and
940
durable resistance. Ann Rev Phytopatol. 2002; 40:349-79. doi:
941
10.1146/annurev.phyto.40.120501.101443. PMID: 12147764
942
43. Urashima AS, Galbieri R, Stabili A. DNA fingerprinting and sexual characterization
943
revealed two distinct populations of Magnaporthe grisea in wheat blast from Brazil. Czech J
944
Genet Plant Breed. 2005; 41:238-45.
945
44. Urashima AS, Hashimoto Y, Don LD, Kusaba M, Tosa Y, Nakayashiki H, et al.
946
Molecular analysis of the wheat blast population in Brazil with s homolog of retrotransposon
947
MGR583. Ann Phytopathol Soc Jpn. 1999; 65:429-36.
948
45. Goulart ACP, Paiva FA. Incidência da brusone (Pyricularia oryzae) em diferentes
949
cultivares de trigo (Triticum aestivum) em condições de campo. Fitopatol Bras. 1992; 17:321-
950
5.
951
46. Goulart ACP, Paiva FA, Andrade PJM. Relação entre a incidência da brusone em espigas
952
de trigo e a presença de Pyricularia grisea nas sementes colhidas. Fitopatol Bras. 1995;
953
20::184-9.
954
47. Maciel JLN, Paludo EA, Só e Silva M, Scheeren PL, Caierão E. Reação à brusone de
955
genótipos de trigo do programa de melhoramento da Embrapa Trigo no estádio de planta
956
adulta. Passo Fundo: Embrapa Trigo; 2008. p. 14.
957
48. Cruz MFA, Prestes AM, Maciel JLN, Scheeren PL. Partial resistance to blast on common
958
and synthetic wheat genotypes in seedling and in adult plant growth stages. Trop Plant
959
Pathol. 2010; 35:24-31. doi: 10.1590/S1982-56762010000100004.
960
49. Urashima AS, Lavorenti NA, Goulart ACP, Mehta YR. Resistance spectra of wheat
961
cultivars and virulence diversity of Magnaporthe grisea isolates in Brazil. Fitop bras. 2004;
962
29:511-8. doi: http://dx.doi.org/10.1590/S0100-41582004000500007.
963
50. Urashima AS, Lavorent NA, Goulart ACP, Mehta YR. Resistance spectra of wheat
964
cultivars and virulence diversity of Magnaporthe grisea isolates in Brazil. Fitopatol Bras.
965
2004; 29:511-8. doi: 10.1590/S0100-41582004000500007.
966
51. Cruz MF, Maciel JLN, Prestes AM, Bombonatto EAS, Pereira JF, Consoli L. Molecular
967
pattern and virulence of Pyricularia grisea isolates from wheat. Trop Plant Pathol. 2009;
968
34:393-401. doi: 10.1590/S1982-56762009000600005.
969
52. Galbieri R, Urashima AS. Sexual characterization, compatibility and occurrence of sexual
970
reproduction among isolates of Pyricularia grisea from different hosts. Summa Phytopathol.
971
2008; 34:22-8. doi: 10.1590/S0100-54052008000100005
972
53. Urashima AS, Igarashi LC, Kato H. Host range, mating type, and fertility of Pyricularia
973
grisea from wheat in Brazil. Plant Dis. 1993; 77:1211-6. doi: 10.1094/PD-77-1211.
974
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

49
54. Urashima AS, Bruno AC. Sexual relationship between Magnaporthe grisea from wheat
975
and from other hosts. Fitopatol Bras. 2001; 26:21-6. doi: 10.1590/S0100-
976
41582001000100004.
977
55. Sydes MA, Peakall R. Extensive clonality in the endangered shrub Haloragodendron
978
lucasii (Haloragaceae) revealed by allozymes and RAPDs. Mol Ecol. 1998; 7: 87-93.
979
56. Peakall R, Smouse PE. GenAlEx 6: genetic analysis in Excel. Population genetic software
980
for teaching and research. Mol Ecol Notes 2006; 6:288-95. doi: 10.1111/j.1471-
981
8286.2005.01155.x.
982
57. Meirmans PG, Van Tienderen PH. Genotype and genodive: two programs for the analysis
983
of genetic diversity of asexual organisms. Mol Ecol Notes. 2004; 4:792-4. doi:
984
10.1111/j.1471-8286.2004.00770.x.
985
58. El Mousadik A, Petit RJ. High level of genetic differentiation for allelic richness among
986
populations of the argan tree [Argania spinosa (L.) Skeels] endemic to Morocco. Theor Appl
987
Genet. 1996; 92:832-9. doi: 10.1007/BF00221895. PMID: 24166548.
988
59. Excoffier L, Laval G, Schneider S. Arlequin (version 3.0): An integrated software
989
package for population genetics data analysis. Evol Bioinform Online. 2005; 1:47-50. PMID:
990
19325852.
991
60. Slatkin M. A measure of population subdivision based on microsatellite allele
992
frequencies. Genetics. 1995; 139:457-62.
993
61. Manly BFJ. Randomization, bootstrap and Monte Carlo methods in biology. 2 ed.
994
London: Chapman & Hall / CRC; 1991. 399 p.
995
62. Bonferroni CE. The calculation of confidence in test groups. (In Italian). Studi in onore
996
del professore Salvatore Ortu Carboni. Rome, Italy1935. p. 13-60.
997
63. Miller MA, Peiffer W, Schwartz T. Creating CIPRES Science Gateway for inference of
998
large phylogenetic trees. Proceedings of the Gateway Computing Environments Workshop
999
(GCE), 14 Nov 2010 New Orleans, LA. 2010.
1000
64. Beerli P, Palczewski M. Unified framework to evaluate panmixia and migration direction
1001
among multiple sampling locations. Genetics. 2010; 185:313-32. doi:
1002
10.1534/genetics.109.112532. PMID: 20176979.
1003
65. Beerli P. How to use migrate or why are markov chain monte carlo programs dificult to
1004
use? In: Bertorelle G, Bruford MW, Hauffe HC, Rizzoli A, Vernesi C, editors. Population
1005
Genetics for Animal Conservation. Conservation Biology. 17. Cambridge, UK: Cambridge
1006
University Press; 2009. p. 49-72.
1007
66. Beerli P. Comparison of Bayesian and maximum-likelihood inference of population
1008
genetic parameters. Bioinformatics. 2006; 22:341-5. doi: 10.1093/bioinformatics/bti803.
1009
PMID: 16317072.
1010
67. Beerli P, Felsenstein J. Maximum likelihood estimation of a migration matrix and
1011
effective population sizes in n subpopulations by using a coalescent approach. PNAS. 2001;
1012
98:4563-8. doi: 10.1073/pnas.081068098. PMID: 11287657.
1013
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

50
68. Beerli P, Felsenstein J. Maximun-likelihood estimation of migration rates and effective
1014
population numbers in two populations using a coalescent approach. Genetics. 1999;
1015
152:763-73. PMID: 10353916.
1016
69. Raymond M, Rousset F. GENEPOP (Version 1.2): Population genetics software for exact
1017
tests and ecumenicism. J Hered. 1995; 86:248-9.
1018
70. Agapow P-M, Burt A. Indices of multilocus linkage disequilibrium. Mol Ecol Notes.
1019
2001; 1:101-2. doi: 10.1046/j.1471-8278.2000.00014.x.
1020
71. Zadoks JC, Chang TT, Konzak CF. A decimal code for the growth stages of cereals.
1021
Weed Res. 1974; 14:415-21.
1022
72. Urashima AS, Leite SF, Galbieri R. Efficiency of aerial dissemination of Pyricularia
1023
grisea. Summa Phytopathol. 2007; 33:275-9. doi: 10.1590/S0100-54052007000300011.
1024
73. Saleh D, Xu P, Shen Y, Li C, Adreit H, Milazzo J, et al. Sex at the origin: an Asian
1025
population of the rice blast fungus Magnaporthe oryzae reproduces sexually. Mol Ecol. 2012;
1026
21:1330-44. doi: 10.1111/j.1365-294X.2012.05469.x. PMID: 22313491.
1027
74. Jank L, Barrios SC, do Valle CB, Simeão RM, Alves GF. The value of improved pastures
1028
to Brazilian beef production. Crop Pasture Sci. 2014:1132-7. doi: 10.1071/CP13319.
1029
75. Tharreau D, Fudal I, Andriantsimialona D, Utami D, Fournier E, Lebrun MH, et al.
1030
World population structure and migration of the rice blast fungus, Magnaporthe oryzae. . In:
1031
Wang GL, Valent B, editors. Advances in genetics, genomics and control of rice blast disease
1032
Dordrecht, The Netherlands.: Springer; 2009. p. 209-15.
1033
76. Saleh D, Milazzo J, Adreit H, Fournier E, Tharreau D. South-East Asia is the center of
1034
origin, diversity and dispersion of the rice blast fungus, Magnaporthe oryzae. New Phytol.
1035
2014; 201:1440-56. doi: 10.1111/nph.12627. PMID: 24320224.
1036
77. Takan JP, Chipili J, Muthumeenakshi S, Talbot NJ, Manyasa EO, Bandyopadhyay R, et
1037
al. Magnaporthe oryzae populations adapted to finger millet and rice exhibit distinctive
1038
patterns of genetic diversity, sexuality and host interaction. Mol Biotechnol. 2012; 50:145-58.
1039
doi: 10.1007/s12033-011-9429-z. PMID: 21701860.
1040
78. Moreira SI. Sexual reproduction studies with Pyricularia oryzae [Ph.D. Thesis]. Lavras,
1041
Minas Gerais, Brazil: Federal University of Lavras; 2015.
1042
79. Hayashi N, Li YC, Li JL, Naito H. In vitro production on rice plants of perithecia of
1043
Magnaporthe grisea from Yunnan, China. Mycol Res. 1997; 101:1308-10. doi:
1044
10.1017/S095375629700422X.
1045
80. Trail F. Sex and Fruiting in Fusarium. In: Brown DW, Proctor RH, editors. Fusarium:
1046
Genomics, Molecular and Cellular Biology. Norfolk, UK.: Caister Academic Press. ; 2013. p.
1047
10.
1048
81. Moreira SI, Ceresini PC, Alves E. Reprodução sexuada em Pyricularia oryzae. Summa
1049
Phytopathol. 2015; 41:175-82.
1050
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

51
82. Papaïx J, Burdon JJ, Zhan J, Thrall PH. Crop pathogen emergence and evolution in agro-
1051
ecological landscapes. Evol Appl. 2015; 8:385-402. doi: 10.1111/eva.12251. PMID:
1052
25926883.
1053
83. Cruz CD, Kiyuna J, Bockus WW, Todd TC, Stack JP, Valent B. Magnaporthe oryzae
1054
conidia on basal wheat leaves as a potential source of wheat blast inoculum. Plant Pathol.
1055
2015; 64:1491-8. doi: 10.1111/ppa.12414.
1056
84. Langmead B, Salzberg SL. Fast gapped-read alignment with Bowtie 2. Nat Methods.
1057
2012; 9:357-9. doi: 10.1038/nmeth.1923. PMID: 22388286.
1058
85. Trapnell C, Roberts A, Goff L, Pertea G, Kim D, Kelley DR, et al. ifferential gene and
1059
transcript expression analysis of RNA-seq experiments with TopHat and Cufflinks. . Nat
1060
Protoc. 2012; 7:562-78. doi: 10.1038/nprot.2012.016. PMID: 22383036.
1061
86. DePristo MA, Banks E, Poplin R, Garimella KV, Maguire JR, Hartl C, et al. Aframework
1062
for variation discovery and genotyping using next-generation DNA sequencing data. Nat
1063
Genet. 2011; 43. doi: 10.1038/ng.806. PMID: 21478889.
1064
87. Pereira JF, Consoli L, Bombonatto EAS, Bonato ALV, Maciel JLN. Development of
1065
genomic SSR markers and molecular characterizaton of Magnaporthe oryzae isolates from
1066
wheat in Brazil. Biochem Genet. 2014; 52:52-70. doi: 10.1007/s10528-013-9627-4. PMID:
1067
24271825.
1068
88. Bruvo R, Michiels NK, D’Souza TG, Schulenburg H. A simple method for the
1069
calculation of microsatellite genotype distances irrespective of ploidy level. Mol Ecol. 2004;
1070
13:2101-6. doi: 10.1111/j.1365-294X.2004.02209.x. PMID: 15189230.
1071
89. Csárdi G, Nepusz T. The igraph software package for complex network research. Inter J
1072
Complex Syst. 2006:1965.
1073
90. Kamvar ZN, Tabima JF, Grunwald N. Poppr: an R package for genetic analysis of
1074
populations with clonal, partially clonal, and/or sexual reproduction. PeerJ. 2014:2:e281. doi:
1075
10.7717/peerj.281.; PMID: 24688859.
1076
91. R-Core-Team. R: a language and environment for statistical computing. R foundation for
1077
statistical computing. Vienna, Austria.2013.
1078
92. Stoddart JA, Taylor JF. Genotypic diversity: estimation and prediction in samples.
1079
Genetics. 1988; 118:705-11.
1080
93. Goudet J. FSTAT (Version 1.2): A computer program to calculate F-statistics. J Hered.
1081
1995; 86:485-6.
1082
94. Excoffier L, Smouse PE, Quattro JM. Analysis of molecular variance inferred from
1083
metric distances among DNA haplotypes: application to human mitochondrial DNA
1084
restriction data. Genetics. 1992; 131:479-91.
1085
95. Xu JR, Hamer JE. Assessment of Magnaporthe grisea mating type by spore PCR. Fungal
1086
Genet Newsl. 1995; 42:80. doi: 10.1079/CJB200545.
1087
1088
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

52
Figure Legends.
1089
1090
Fig 1. Population genomic analyses of transcriptomic single nucleotide polymorphisms
1091
among isolates of Pyricularia graminis-tritici from wheat and several other poaceous hosts in
1092
Brazil, P. oryzae, P. grisea from Digitaria sanguinalis and other Pyricularia spp. from
1093
Setaria italica, and Eleusine indica. The scale bar shows the number of informative sites. The
1094
samples included 47 rice blast strains with publically available genome sequences, 32
1095
Brazilian wheat and other poaceous blast strains, seven strains from various additional hosts
1096
and four wheat blast samples collected in Bangladesh in spring 2016. The dataset contained
1097
only SNPs reliably called in the transcriptomic sequences of the Bangladesh sample 12 and
1098
genotyped in at least 90% of all other strains. We retained 55,041 informative SNPs. A
1099
maximum likelihood phylogeny was constructed using RAxML version 8.2.8 with a GTR
1100
substitution matrix and 100 bootstrap replicates. Pygt and PoT stands for the formerly
1101
described P. graminis-tritici and P. oryzae pathotype Triticum.
1102
1103
Fig 2. Geographical location of populations of Pyricularia graminis-tritici and P. oryzae
1104
examined in this study. The distinct colors in each population indicate the proportion of
1105
clones, while light gray indicates the proportion of distinct genotypes. Population 2005W was
1106
included because it represents a collection of MLMG genotypes sampled earlier in 2005 from
1107
central-southern Brazil.
1108
1109
Fig 3. Minimum spanning network based on Bruvo distance for comparing 219 multilocus
1110
microsatellite genotypes (MLMG) of Pyricularia graminis-tritici isolates obtained from
1111
wheat and other poaceous hosts, and P. oryzae obtained from rice. Each node in the network
1112
represents a single haploid MLMG determined using 11 microsatellite loci. The size of the
1113
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

53
node (circle) represents the frequency of the sampled MLMGs. The shading (colors) of the
1114
nodes represents the membership of each population, while the thickness of the connecting
1115
lines and shading represent the degree of relationship between MLMGs. The line length is
1116
arbitrary. MLMGs shared among populations of P. graminis-tritici from wheat and other
1117
grasses are shown in red, while MLMGs associated only with one host are showed in black.
1118
1119
Fig 4. Development of proto-perithecia and perithecia of Pyricularia graminis-tritici induced
1120
by injection of living conidia of isolates PY33.1 (Mat1-1) and PY05046 (Mat1-2) within
1121
autoclaved stems sections of wheat (Triticum aestivum) cv. MGS Brilhante. Panel A, site of
1122
injection (arrow 1) and fungal colonization within plant tissues (arrow 2); development of
1123
proto-perithecia (B) and perithecia (C) inside stems; D, perithecia developing from the
1124
internal plant tissues to beak emersion; proto-perithecia at interface (E) and on surface of
1125
plant tissues (F and G). H, Control composed of autoclaved stems without inoculation. The
1126
images of panels A, B, E and G were acquired by scanning electron microscope. Images of
1127
panels C, D and F were acquired by light microscope.
1128
1129
Fig 5. Development of proto-perithecia and perithecia of Pyricularia graminis-tritici induced
1130
by injection of living conidia of isolates Py33.1 (Mat1-1) and Py5046 (Mat1-2) within
1131
senescing stems sections of different Poaceae species. Panel A, Procedure for inoculation by
1132
injection of living spores into the host stems. B, Pieces of stem placed at 120 mm Petri dishes
1133
to incubation in humid chamber 1 month after inoculation. C, Stems with proto-perithecia
1134
and/or perithecia development after incubation in humid chamber (arrows). Fruiting body in
1135
different plant species: D, canary seeds (Phalaris canariensis); E, rice (O. sativa) cv.
1136
Primavera; F, rice cv. Relampago; G, red rice (O. sativa) cv. Yin Lu 30; H, Brachiaria cv.
1137
Hybrid Mulato; I, barley (Hordeum vulgare) cv. BR Elis; J, barley cv. MN 743; K,Rye
1138
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

54
(Secale cereale) cv. BR1; L, black oat (Avena strigosa) cv. Embrapa 29 Garoa; M, foxtail
1139
millet (Setaria italica); N, wheat (Triticum aestivum) cv. BRS 264; O, wheat cv. MGS
1140
Brilhante; P, triticale (x Triticosecale) cv. IAC Caninde. The images of panels D to P were
1141
acquired by bright field microscopy.
1142
1143
Fig 6. Virulence spectrum and typical blast lesions on wheat seedlings caused by isolates of
1144
Pyricularia graminis-tritici belonging to the predominant seedling virulence group (SVG L)
1145
on the differential set of ten wheat (Triticum aestivum) cultivars and one barley (Hordeum
1146
vulgare) cultivar. The differential set was consist of ten wheat cultivars: a) Anahuac 75; b)
1147
BR 18; c) BR 24; d) BRS 220; e) BRS 229; f) MGS 3 Brilhante; g), BRS Buruti; h) CNT 8;
1148
j) Renan; k) BRS 234; and one barley cultivar: i) PFC 2010123. Varieties indicated in bold
1149
showed resistant reaction. Isolate inoculated: 12.1.109.
1150
1151
Fig 7. Distribution of seeding virulence groups (SVGs) of the wheat blast pathogen
1152
Pyricularia graminis-tritici in ten populations from central-southern Brazil.
1153
1154
Fig 8. Virulence spectrum and typical blast lesions on wheat heads caused by isolates of
1155
Pyricularia graminis-tritici belonging to the predominant head virulence group (HVG A) on
1156
the differential set of ten wheat (Triticum aestivum) cultivars and one barley (Hordeum
1157
vulgare) cultivar. The differential set was consist of ten wheat cultivars: a) BRS 229; b) CNT
1158
8; d) BR 234; e) Anahuac 75; f) BR 24; g) BRS 220; h), BR 18; i) Renan; j) BRS Buriti; k)
1159
MGS 3 Brilhante; and one barley cultivar: c) PFC 2010123. All cultivars showed susceptible
1160
reactions. Isolate inoculated: 12.1.170.
1161
1162
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

55
Fig 9. Distribution of head virulence groups (HVGs) of the wheat blast pathogen Pyricularia
1163
graminis-tritici in ten populations from central-southern Brazil.
1164
1165
Fig 10. Pyricularia graminis-tritici life cycle and wheat blast disease cycle. At the end of a
1166
cropping season (Ae), wheat blast infection on ears will result in seed infection (B, C),
1167
providing inoculum for either local or long distance dispersal of the pathogen [72]. Crop
1168
residues remaining in the field after harvesting, especially under no tillage conditions, serves
1169
as a niche for sexual reproduction of the fungus (D, 1-4); the resulting mature perithecia
1170
release ascospores by deliquescence of asci (D1), giving rise to new fungal individuals by the
1171
germination of terminal cells (D2), which is followed by fungal vegetative growth and
1172
subsequent conidiogenesis (D3) [81]; primary conidia originating from this process are
1173
released (D4) and constitute airborne inoculum for leaf infection on other poaceous hosts,
1174
either invasive or contiguous to wheat fields (E, F) [1, 53, 82]. Perithecia can be formed also
1175
in other infected poaceous hosts and major pasture grasses and ascospores released out onto a
1176
nearby wheat crop (E). Seedborne inoculum (B, C) results in primary infections in a newly
1177
established wheat crops. (F4) conidia released from leaf blast lesions on other poaceous hosts
1178
nearby wheat crops also contributes inoculum for wheat blast on ears [1, 53]. Conidia
1179
production on leaves (Af) in the lower canopy of certain wheat cultivars coinciding with
1180
spike emergence under field conditions and could be an important trigger for wheat blast
1181
epidemics on ears (Ae) [83].
1182
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

12.0.368
12.0.345
12.0.326
12.0.555i
Digitaria, Urochloa, Echinochloa,
Avena and wheat, Brazil
88
100
100
100
100
90
100
100
100
100
100
100
100
100
100
100
100
100
100
PY86.1
PY05002
PY05010
12.1.127
12.1.169
12.1.204
Urochloa, Brazil
12.0.009i
12.0.012i
12.0.007i
PY06025
PY06045
PY06047
PY25.1
PY05035
PY06017
BR32
PY05033
PY35.3
PY36.1
PY05003
Wheat, Brazil
12.1.205
12.1.032i
7.LIB1751
12.LIB21752
5.LIB21750
2.LIB1749
PY0925
12.1.053i
12.0.534i
12.0.642i
12.1.117
12.1.037 Wheat, Brazil
Wheat, Bangladesh
Eleusine, Cenchrus, Brazil
Eleusine indica
CD156
4091-5-8
Digitaria sanguinalis
BR29
Setaria
italica
US71
363
K96-07
K98-10
Mo1106.2
Mo2303.1
98-06
MoHN06
MoJS25
P131
PR72
CA205
Mo4403.2
IT-10
FR13
K93-16
K96-11
MoAH06
MoGD22
MoCQ11
MoJL10
MoZJ15
MoNX37
MoHB12
MoSC05
K91-10
K91-13
MoKJ201
K88-24
K98-02
GOV41
Mo903.4
MoK84-01
MoGX01
Y34
Mo401.4
Mo1801.4
Mo1836.3
Mo4603.4
FJ81278
B157
MG01
K91-30
K88-07
PH14
TH12
TH16
HN19311
70-15
GY11
BM1-24
10,000 informative SNPs
Rice
(Oryza sativa)
Weeping lovegrass isolate (cross)
Pygt
PoT
PoT
PoT
Pygt
PoO
Pyricularia oryzae
Pyricularia graminis-tritici
Wheat, Brazil
Hordeum
vulgare
Pyricularia grisea
Figure 1 Click here to download Figure Fig 1.pdf
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

Latitude (S - N)
Longitude (E - W)
0 200 400 km
Uruguay
Argentina
Paraguay
Bolivia
South America
Altantic Ocean
Brazil
-30° -25° -20° -15° -10° -5° 0°
-70° -65° -60° -55° -50° -45° -40° -35°
MS
RS
MG
SP
!"
Pyricularia oryzae
patho type
Triticum (Po-T)
Pyricularia
graminis-tritici
(Pgt)
!"#$%
Pyricularia oryza e
path otype
Triticum (Po-T)
Pyricularia
graminis-tritici
(Pgt)
!"
Pyricularia oryza e
pat hotype
Triticum (Po-T)
Pyricularia
graminis-tritici
(Pgt)
!"
Pyricularia oryzae
pathotype
Triticum (Po-T)
Pyricularia
graminis
-tritici
(Pgt)
!"
Pyricularia oryzae
pathotype
Triticum (Po-T)
Pyricularia
graminis-tritici
(Pgt)
Pyricularia oryzae
pathoty pe
Triticum (Po-T)
Pyricularia
graminis-tritici
(Pgt)
MS
74
DF-GO
86
62
PR
82
28
52
31
PR
31
Triticum
2005
7 9
Distinct colors: Proportion of clones
Light gray: Proportion of distinct
genotypes
Total N = 526 isolates
W
W
W
W
W
W
W
p
p
Figure 2 Click here to download Figure Fig 2.pdf
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

1
2
3
4
5
67
8
9
10
11
12
13
14
15
16
17
18 19
20
21
22
23
24 25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91 92
93
94
95
96
97
98
99
100
101 102
103
104 105
106
107
108
109
110
111 112
113
114
115
116
117
118
119
120
121
122
123
124
125
126 127
128
129
130
131
132
133
134
135 136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
Populations
2005_W
DFGO_W
MG_W
MS_W
PR_W
RS_W
SP_W
MS_P
PR_P
Derived from other Poaceae
Derived from wheat (Triticum aestivum)
Figure 3 Click here to download Figure Fig 3.pdf
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

100µm 10µm
100µm
100µm
100µm 10µm
10µm
1
2
AB
CD
EF
GH
Figure 4 Click here to download Figure Fig 4.pdf
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

50µm
50µm 50µm 50µm 50µm
50µm 50µm 50µm 50µm
50µm 50µm 50µm 50µm
ABC D
EFGH
IJKL
M P
NO
Figure 5 Click here to download Figure Fig 5.pdf
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

a b c d e
g h i j k
f
Figure 6 Click here to download Figure Fig 6.pdf
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

Fig 5.
A
B
C
D
E
F
G
H
I
J
K
L
M
N
O
P
Q
R
S
T
U
V
W
X
Y
0
20
40
60
Race
A
B
C
D
E
F
G
H
I
J
K
L
M
N
O
P
Q
R
S
T
U
V
W
X
Y
0
20
40
60
Race
Percentage
A
B
C
D
E
F
G
H
I
J
K
L
M
N
O
P
Q
R
S
T
U
V
W
X
Y
0
20
40
60
Race
Percentage
A
B
C
D
E
F
G
H
I
J
K
L
M
N
O
P
Q
R
S
T
U
V
W
X
Y
0
20
40
60
Race
A
B
C
D
E
F
G
H
I
J
K
L
M
N
O
P
Q
R
S
T
U
V
W
X
Y
0
20
40
60
Race
Percentage
A
B
C
D
E
F
G
H
I
J
K
L
M
N
O
P
Q
R
S
T
U
V
W
X
Y
0
20
40
60
Race
Percentage
A
B
C
D
E
F
G
H
I
J
K
L
M
N
O
P
Q
R
S
T
U
V
W
X
Y
0
20
40
60
Race
A
B
C
D
E
F
G
H
I
J
K
L
M
N
O
P
Q
R
S
T
U
V
W
X
Y
0
20
40
60
Race
Percentage
A
B
C
D
E
F
G
H
I
J
K
L
M
N
O
P
Q
R
S
T
U
V
W
X
Y
0
20
40
60
Race
Percentage
DF-GO
W
MG
W
MS
W
PR
W
RS
W
SP
W
MS
P
PR
P
All populations
Percentage
Percentage
Percentage
Seedling virulence groups (SVGs)
Figure 7 Click here to download Figure Fig 7.pdf
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

a b c d e f h i j k
g
Figure 8 Click here to download Figure Fig 8.pdf
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

Fig 7
A
AA
B
BB
C
CC
D
DD
T
0
20
40
60
Race
A
AA
B
BB
C
CC
D
DD
T
0
20
40
60
Race
Percentage
A
AA
B
BB
C
CC
D
DD
T
0
20
40
60
Race
Percentage
A
AA
B
BB
C
CC
D
DD
T
0
20
40
60
Race
A
AA
B
BB
C
CC
D
DD
T
0
20
40
60
Race
Percentage
A
AA
B
BB
C
CC
D
DD
T
0
20
40
60
Race
Percentage
A
AA
B
BB
C
CC
D
DD
T
0
20
40
60
Race
A
AA
B
BB
C
CC
D
DD
T
0
20
40
60
Race
Percentage
A
AA
B
BB
C
CC
D
DD
T
0
20
40
60
Race
Percentage
DF-GO
W
MG
W
MS
W
PR
W
RS
W
SP
W
MS
P
PR
P
All populations
Head virulence groups (HVGs)
PercentagePercentage Percentage
Figure 9 Click here to download Figure Fig 9.pdf
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;

Figure 10 Click here to download Figure Fig 10 .jpg
.CC-BY-NC-ND 4.0 International licenseIt is made available under a
(which was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
The copyright holder for this preprint. http://dx.doi.org/10.1101/203455doi: bioRxiv preprint first posted online Oct. 16, 2017;