ABSTRACT
This study provides a phylogeographic insight into the population diversity of Xanthomonas translucens strains causing bacterial leaf streak disease of small-grain cereals in Iran. Among the 65 bacterial strains isolated from wheat, barley, and gramineous weeds in eight Iranian provinces, multilocus sequence analysis and typing (MLSA and MLST) of four housekeeping genes (dnaK, fyuA, gyrB, and rpoD), identified 57 strains as X. translucens pv. undulosa, while eight strains were identified as X. translucens pv. translucens. Although the pathogenicity patterns on oat and ryegrass weed species varied among the strains, all X. translucens pv. undulosa strains were pathogenic on barley, Harding’s grass, rye (except for XtKm35) and wheat, and all X. translucens pv. translucens strains were pathogenic on barley and Harding’s grass, while none of the latter group was pathogenic on rye or wheat (except for XtKm18). MLST using the 65 strains isolated in Iran, as well as the sequences of the four genes from 112 strains of worldwide origin retrieved from the GenBank database, revealed higher genetic diversity (i.e., haplotype frequency, haplotype diversity, and percentage of polymorphic sites) among the Iranian population of X. translucens than among the North American strains of the pathogen. High genetic diversity of the BLS pathogen in Iran was in congruence with the fact that the Iranian Plateau is considered the center of origin of cultivated wheat. However, further studies using larger collections of strains are warranted to precisely elucidate the global population diversity and center of origin of the pathogen.
IMPORTANCE Bacterial leaf streak (BLS) of small-grain cereals (i.e., wheat and barley) is one of the economically important diseases of gramineous crops worldwide. The disease occurs in many countries across the globe, with particular importance in regions characterized by high levels of precipitation. Two genetically distinct xanthomonads—namely, Xanthomonas translucens pv. undulosa and X. translucens pv. translucens—have been reported to cause BLS disease on small-grain cereals. As seed-borne pathogens, the causal agents are included in the A2 list of quarantine pathogens by the European and Mediterranean Plant Protection Organization (EPPO). Despite its global distribution and high economic importance, the population structure, genetic diversity, and phylogeography of X. translucens remain undetermined. This study, using MLSA and MLST, provides a global-scale phylogeography of X. translucens strains infecting small-grain cereals. Based on the diversity parameters, neutrality indices, and population structure, we observe higher genetic diversity of the BLS pathogen in Iran, which is geographically close to the center of origin of common wheat, than has so far been observed in other areas of the world, including North America. The results obtained in this study provide a novel insight into the genetic diversity and population structure of the BLS pathogen of small-grain cereals on a global scale.
INTRODUCTION
Among the biotic constraints of small-grain cereals in the Poaceae family, bacterial leaf streak (BLS), caused by different pathovars of Xanthomonas translucens (i.e., X. translucens pv. translucens and X. translucens pv. undulosa), is considered one of the most important seed-borne pathogens around the globe (1, 2). The disease was first described on barley (Hordeum vulgare) in the United States in 1917, and currently, the causal agent is widespread through all six continents. Xanthomonas translucens is included in the A2 list of quarantine pathogens by the European and Mediterranean Plant Protection Organization (EPPO) (3). The pathogen is seed-borne, and infected seeds are the main source of inoculum for long-distance distribution. Furthermore, a humid and warm environment coupled with frequent overhead irrigation maximizes bacterial growth and plant-to-plant spread in the field, leading to 10 to 40% yield losses (2). When the symptoms are found on wheat glumes (a leaf-like structure attached to the flowers of Gramineae), the disease is called “black chaff” and is characterized by black, longitudinal, and parallel stripes (1). Yield loss due to BLS is proposed to be attributable to disease severity on the wheat spikes and flag leaf (the uppermost leaf of the plant). In severe cases of disease incidence, 5 to 10% of the wheat spikes may be sterile due to the BLS infection. It has also been noted that 50 and 100% disease severity on the flag leaf will result in 8 to 13% and 13 to 34% kernel weight losses, respectively (2).
Until recently, discrimination of X. translucens pathovars has mostly relied on the host ranges and symptom profiles of the pathogens on a set of gramineous species (4–6). For instance, it has been noted that X. translucens pv. translucens, including the strains previously designated X. translucens pv. hordei, is limited to barley, while X. translucens pv. undulosa displays a broader host range, infecting wheat, barley, oat, rye, bromes, and triticale (2, 5). Pathovar designation of X. translucens strains is complicated not only because of the overlapping host ranges (5, 7) but also because of similar genotypic features of the pathotype strains (3, 6). Conventional bacteriological typing techniques (e.g., repetitive element palindromic PCR [rep-PCR], amplified fragment length polymorphism [AFLP], and protein/fatty acid profiling) can only partly solve the taxonomic issues among the X. translucens pathovars (4, 5, 8, 9). However, it has recently been shown that multilocus sequence analysis (MLSA) and typing (MLST) using partial sequences of four housekeeping genes (dnaK, fyuA, gyrB, and rpoD) could discriminate between X. translucens strains isolated from gramineous crops, and the resulting scheme was in congruence with their experimental host range and host of isolation, as well as draft-genome-sequence-based phylogeny of the strains (3, 6).
Although Curland et al. (6) pinpointed the best method so far for evaluating the genetic diversity of X. translucens strains, the exploration was limited to North American strains, thus leaving the Old World population of the pathogen uninvestigated. Considering that the center of origin of cultivated wheat was determined to be in the Karacadağ Mountains in southeastern Turkey (10, 11), a large-scale phylogenetic analysis is warranted to shed light on the population structure and diversity parameters of X. translucens in the Old World and to provide a global-scale phylogeography of the species. While field symptoms resembling BLS infection have occasionally been observed for a long time on barley and wheat in Iran, the presence of X. translucens was scientifically first described in Kerman Province (southern Iran) in 1983 (12). However, information on the geographic distribution and phylogenetic position of the causal agent(s) is mostly lacking in the country. Given the fact that wheat is a staple food in Iran, with the highest economic importance among the agricultural commodities of the country, precise characterization and identification of the BLS pathogen will aim to develop sustainable methods to manage the disease. Indeed, by 2016, wheat cultivation occupied over 50% of the total area dedicated to annual crops in Iran, and the areas cultivated with wheat accounted for more than 5,928,000 ha, with an annual production of 14.6 million tonnes (13). Hence, a multiphasic characterization and comparative phylogenetic analysis is warranted to decipher the population structure of the BLS pathogen in Iran, which at the same time will help to elucidate the phylogeography of cereal-infecting X. translucens strains at a global scale.
The objectives of the present study were to (i) assess the host range of X. translucens strains isolated from wheat, barley, rye, and ryegrass (Lolium spp.) in Iran and (ii) determine the phylogenetic positions of the strains using the 4-gene (dnaK, fyuA, gyrB, and rpoD) MLSA/MLST scheme. Furthermore, a global-scale phylogeographic analysis was conducted using the strains isolated in Iran and the publicly available sequences of 112 X. translucens strains of worldwide origin (6, 14). Haplotype diversity and sequence variation statistics showed that the strains isolated in Iran possess higher genetic diversity than the New World strains of the pathogen.
RESULTS
Bacterial strains and host range assays.Sixty-five Gram-negative xanthomonad-like bacterial strains were isolated from wheat, barley, and gramineous weeds in cereal-growing areas in Iran from 2008 to 2017 (Table 1). The number of surveyed fields in each area per year depended on the BLS incidence and severity reports received from the local crop sanitary inspectors. The 65 strains were isolated in eight provinces throughout southern (Bushehr, Fars, and Kerman), western (Kurdistan, Hamadan, Lorestan, and Kermanshah), and northwestern (Zanjan) Iran (Fig. S1 in the supplemental material). All of the strains formed pale-yellow colonies on yeast extract-dextrose-calcium carbonate (YDC) medium and elicited a hypersensitivity reaction (HR) on tobacco leaves. Furthermore, the specific primer pair T1/T2 (7) amplified a 139-bp DNA fragment of the tRNAAla locus in all 65 strains that is useful in identifying X. translucens species. Consequently, all of the strains were subjected to the pathogenicity tests and host range assays.
Xanthomonas translucens strains isolated in Iran and their host plants, geographic areas, years of isolation, and sequence types based on the partial sequences of four housekeeping genes
In the host range assays, all strains were shown to be pathogenic on their host of isolation except strain XtKm15, which was isolated from ryegrass but did not induce symptoms on this species. On the other hand, all but seven strains (XtKm4, XtKm7, XtKm8, XtKm9, XtKm11, XtKm33, and XtKm34) were pathogenic on wheat, although the symptoms produced by XtKm18 on wheat plants were less severe than those produced by the other strains. All of the strains that were nonpathogenic on wheat plants had been isolated from barley. Although all of the strains were pathogenic on barley plants, the barley strain XtKm9, as well as the type strain of X. translucens pv. translucens (DSM 18974, used as a positive control), were pathogenic only on barley plants. All of the strains isolated from wheat were pathogenic on wheat and barley, while only half of the barley strains (7 out of 14) were pathogenic on wheat (Table 1; Fig. 1). In general, the host range of the wheat strains was broader than that of the barley strains. Twenty wheat strains, four barley strains (XtHn9, XtHn11, XtLr7, and XtLr9), and the rye strain XtKm3 were pathogenic on all six small-grain cereals tested (Fig. 1). The few remaining strains were nonpathogenic on at least one of the plant species tested under the greenhouse conditions used.
Phylogeny of Xanthomonas translucens strains isolated in Iran based on the concatenated sequences of four housekeeping genes (dnaK, fyuA, gyrB, and rpoD). Xanthomonas translucens pv. poae (LMG 728T) was used as an out-group to root the tree. The bar presents numbers of substitutions per site. Among the 65 Iranian strains, 57 strains were identified as X. translucens pv. undulosa, while 8 strains were identified as X. translucens pv. translucens. The blue background indicates X. translucens pv. undulosa strains, while red, yellow, and green backgrounds indicate different subclusters of X. translucens pv. translucens. The black-and-white table at the right shows hosts of isolation (Host) and host ranges of the strains on six gramineous species, including wheat (W), barley (B), oat (O), rye (R), ryegrass (Rg), and Harding’s grass (P [Phalaris]), where black squares indicate pathogenicity and white squares indicate nonpathogenicity. A dash mark in the host range table means that pathogenicity has not been determined. The atypical strain XtKm18 (indicated with an asterisk inside a black square), which was isolated from barley and identified as X. translucens pv. translucens, induced water-soaking symptoms on wheat plants.
Phylogenetic analyses.Multilocus sequence analysis of four housekeeping genes (dnaK, fyuA, gyrB, and rpoD) was conducted on all 65 strains used in this study. BLAST searches using the sequences of each of the individual genes revealed that the strains isolated in Iran had 99 to 100% sequence identity with the type/reference strains of X. translucens, i.e., X. translucens pv. undulosa (LMG 892T), X. translucens pv. secalis (LMG 883T), and X. translucens pv. translucens (ICMP 5752T). A phylogenetic tree constructed using the data set of concatenated sequences of dnaK, fyuA, gyrB, and rpoD genes confirmed the results obtained from the BLAST analyses. All of the strains isolated in Iran clustered in a monophyletic clade within X. translucens, although they were scattered through different pathovars and phylogroups of the species (Fig. 1; Fig. S2). Similar results were obtained when the sequences of individual housekeeping genes were subjected to the phylogenetic analysis (data not shown).
Eighty-seven percent of X. translucens strains isolated in Iran (57 of 65 strains) were clustered with the pathotype strains of X. translucens pv. undulosa and X. translucens pv. secalis. These strains were divided into two main subclusters (subcluster I, including 23 strains, and subcluster II, including 18 strains), as well as 16 unclustered strains (Fig. 1). Among the aforementioned 57 strains, 49 strains were isolated from wheat, 6 strains were isolated from barley, and each of the 2 remaining strains was isolated from rye or from ryegrass. According to the host of isolation, we refer to the wheat, barley and ryegrass strains as X. translucens pv. undulosa, while the rye strain (XtKm3) is considered X. translucens pv. secalis. Regardless of their geographic origin (province/county), all 49 wheat strains, as well as the 2 strains isolated from weed species (XtKm3 and XtKm15, isolated from rye and ryegrass, respectively), were identified as X. translucens pv. undulosa, while only 43% (6 of 14) of the barley strains were clustered in this pathovar (Fig. 1). The remaining eight barley strains were clustered among the strains of X. translucens pv. translucens; however, they were scattered through three subclusters. Interestingly, all of the Iranian strains identified as X. translucens pv. translucens were isolated from barley in Kerman Province (Fig. S3). Although strains XtKm18, XtKm33, XtKm34, XtKm8, and XtKm7 were clustered in a monophyletic group, XtKm7 was differentiated from the remaining four strains and placed in a separate subcluster (Fig. 1). Strain XtKm7 showed one and five nucleotide differences in the gyrB and rpoD gene sequences, respectively, compared to the sequences of the subcluster consisting of strains XtKm8, XtKm18, XtKm33, and XtKm34. Furthermore, strains XtKm11, XtKm4, and XtKm9, which clustered together with the strains isolated from barley in the United States, had one, three, and six nucleotide differences in the dnaK, fyuA, and rpoD gene sequences, respectively, from the sequences of the subgroup containing strains XtKm8, XtKm18, XtKm33, and XtKm34. Moreover, there were three nucleotides differences in the fyuA sequences and one nucleotide difference in the dnaK, gyrB, and rpoD sequences between strain XtKm7 and the three strains XtKm4, XtKm9, and XtKm11.
The phylogenetic positions of the strains were in congruence not only with their host of isolation but also with the experimental host range of the strains. For instance, all the strains identified as X. translucens pv. undulosa were pathogenic on wheat, barley, Harding’s grass (Phalaris aquatica), and rye except the wheat strain XtKm35, which was not pathogenic on rye, whereas all X. translucens pv. translucens strains were pathogenic on barley and Harding’s grass and none of them was pathogenic on wheat and rye except strain XtKm18, which induced water soaking symptoms on wheat plants. The pathogenicity patterns on ryegrass and oat plants were variable among the strains regardless of their host of isolation and phylogenetic position. Interestingly, all 23 strains that clustered in subcluster I of X. translucens pv. undulosa were pathogenic on all six plant species evaluated. However, neither oat nor ryegrass plants were infected with the 18 strains in subcluster II. More specifically, none of the strains XtFa2, XtFa3, XtHn2, XtKm10, XtKm22, XtKm24, XtKm30, XtKr1, XtLr1, and XtZa1, which were in the same multilocus haplotype (MH), was pathogenic on oat.
Genetic diversity.The strains isolated in Iran carried different allelic forms and sequence types (STs) and corresponded to 26 MHs based on the concatenated sequences of four housekeeping genes (Table S1). As for the individual genes in the Iranian strains, 15, 3, 5, and 11 STs were detected for the dnaK, fyuA, gyrB, and rpoD genes, respectively (Fig. S3), while only 8, 3, 2, and 8 STs were detected for these genes in 112 strains from other countries (Table S1). Sequence variation statistics and diversity parameters among the strains isolated in Iran were calculated using DnaSP version 5.10 software and compared with those of 112 X. translucens strains isolated in six different (non-Iran) countries (Table 2). As for the 58 X. translucens pv. undulosa strains isolated in Iran (strains isolated in this study and the reference strain ICMP 11055 isolated in southern Iran in 1983), there were 8, 1, 4, and 7 STs in the sequences of dnaK, fyuA, gyrB, and rpoD genes, respectively, while 7, 2, 3, and 4 STs, respectively, were found in the X. translucens pv. translucens strains (Fig. 2). The haplotype frequency (HF; ratio of the number of haplotypes versus the number of strains) indices were 0.137, 0.017, 0.068, and 0.120 for X. translucens pv. undulosa and 0.875, 0.250, 0.375, and 0.500 for X. translucens pv. translucens strains in the dnaK, fyuA, gyrB, and rpoD genes, respectively, indicating higher genetic diversity among the X. translucens pv. translucens strains in Iran. Furthermore, comparison of HF and haplotype diversity (HD) indices between the strains isolated in Iran and those of 112 non-Iran strains revealed higher genetic diversity among the Iranian strains of X. translucens (Table 2).
Sequence variation statistics of the partial sequences of four housekeeping genes from Xanthomonas translucens strains isolated in Iran in comparison to those of the strains isolated in six other countries
TCS multilocus haplotype (MH) network generated using the POPART program from the concatenated partial sequences of four housekeeping genes (dnaK, fyuA, gyrB, and rpoD) in 178 Xanthomonas translucens strains from Iran and seven countries around the globe. The size of a circle indicates the relative frequency of sequences belonging to a particular MH. Hatch marks along the branches indicate the numbers of mutations. Each color indicates a different geographic area. To provide precise insight into the diversity of the pathogen, strains isolated in southern Iran and western Iran were considered distinct groups. The number to the left of a slash represents the MH number (Table S1), while the number to the right of a slash indicates the number of strains in a given MH. Xanthomonas translucens pv. undulosa strains are surrounded by blue lines, while X. translucens pv. translucens strains are surrounded by red lines. MH2, consisting of 56 X. translucens pv. undulosa strains, was considered the founder genotype of the pathovar as confirmed by the phylogeographic analyses using the eBURST program (Fig. S5).
Several strains isolated in different geographical areas in Iran shared the same MH with strains isolated in other countries (Fig. 2). For instance, strain XtKm3 isolated in Kerman (Iran) and strains CIX36, CIX53, and CIX54 isolated in Minnesota belonged to the same MH, while strain XtKm9 shared the same MH with barley strain UPB458 isolated in India. Interestingly, MH2 included 15 and 39 strains isolated in Iran and the United States, respectively, as well as the pathotype strains of X. translucens pv. undulosa (LMG 892T) and X. translucens pv. secalis (LMG 883T). On the other hand, several MHs were represented by only one Iranian strain (e.g., XtKm4, XtKm7, XtKm8, XtKm11, XtKm18, XtKm33, and XtKm34 for X. translucens pv. translucens and XtFa4, XtHn7, XtKr4, XtLr5, XtLr6, XtLr8, and XtZa4 for X. translucens pv. undulosa), which indicates a higher allelic richness of the pathogen in Iran (Fig. 2; Table S1).
Phylogenetic networks were generated using the splits-decomposition method for all the individual genes, as well as for the concatenated-sequence data set. When 178 worldwide strains were analyzed, in spite of minor reticulations in the splits-decomposition network, the pairwise homoplasy index (PHI) test did not find statistically significant evidence for recombination either in the individual genes or in concatenated data sets (data not shown). The results of the PHI test were in congruence with those of the Recombination Detection Program (RDP), which did not find statistically significant evidence for recombination. On the other hand, none of the population neutrality indices (i.e., Tajima’s D, Fu and Li’s D*, and Fu and Li’s F*) was significant among the 66 strains isolated in Iran, thus indicating the lack of evidence for selection among the population and that the population is evolving per mutation-drift equilibrium (Table 2). On the other hand, the non-Iran X. translucens pv. undulosa strains showed significant departure from the mutation-drift equilibrium in their rpoD gene sequences. The negative value of neutrality indices in rpoD gene sequences of non-Iran X. translucens pv. undulosa strains suggests a recent selective sweep or/and population expansion after a recent bottleneck, although further investigations using the sequences of additional housekeeping genes are needed to prove these observations.
Phylogeography of strains of worldwide origin.Several lines of evidence of geographic structure in the ST distribution were detected in the networks built by the TCS algorithm implemented in POPART version 1.7 (Fig. 3). Although a number of STs were shared among the strains of worldwide origin, country-specific STs were more prevalent in the dnaK and rpoD networks. For instance, in the dnaK, gyrB, and rpoD sequences, there were 11, 4, and 7 STs consisting of Iranian strains, while only 2 and 4 STs in the dnaK and rpoD sequences, respectively, corresponded to those in United States strains (Fig. 3). On the other hand, strain DAR61454 isolated in Australia shared the same MH with the strains isolated in Iran, Canada, and the United States, while strain UPB458 isolated in India shared the same MH with strains isolated in southern Iran (Fig. 2; Table S1), raising the speculation that strain UPB458 originated in Iran. Furthermore, six strains isolated in Canada (XT5523, XT5770, XT5791, XT8, LMG 883, and LMG 892), as well as strains UPB513 and UPB787 isolated in Mexico and Paraguay, respectively, shared the same MH with strains isolated in the Midwestern United States (Fig. S4). The patterns of MH positioning in the TCS network (Fig. 2 and 3) were in congruence with the MrBayes bifurcating trees; however, reticulations in the TCS networks of the dnaK and rpoD sequences, as well as of the concatenated sequences, suggest a more complex evolutionary history of X. translucens than can be observed in the phylogenetic trees (Fig. 1).
TCS haplotype network generated using the POPART program from the partial sequences of four housekeeping genes (dnaK, fyuA, gyrB, and rpoD) in 178 Xanthomonas translucens strains. The size of a circle indicates the relative frequency of sequences belonging to a particular sequence type. Hatch marks along the branches indicate the numbers of mutations. Each color indicates a different geographic area. To provide precise insight into the diversity of the pathogen, strains isolated in southern Iran and western Iran were considered distinct groups. The number to the left of a slash represents the ST number (Table S1), while the number to the right of a slash indicates the number of strains in a given ST. Strains isolated in Iran generated several unique haplotypes, indicating a higher genetic diversity of the pathogen in Iran.
Using the eBURST analysis, 178 X. translucens strains were assigned to 46 MHs within three groups and four singleton sequences. Group one consisted of 143 strains and 37 MHs (Fig. S5), while groups two and three, both consisting of only X. translucens pv. translucens strains, included 30 and 2 strains, as well as 4 and 2 MHs, respectively (data not shown). Although the founder genotype of X. translucens pv. undulosa (MH2) and the respective subgroups in group one were supported with 99 to 100% bootstrap values (Fig. S5), assignment of the founder genotype in groups two and three was not supported by high bootstrap value (>20), and hence, they were not considered reliable founder genotypes. The founder genotype MH2 included 56 strains (as shown in Fig. 2), with 20, 9, and 4 single-, double-, and triple-locus variants, respectively. Two single-locus variants of MH2 (MH10 and MH28) were identified as the subgroup founders, each with its own cluster of linked single-locus variants (Fig. S5). Among the 56 strains in MH2, 15 strains were isolated in Iran, while 40 and 1 strain were isolated in the United States and Australia, respectively. The results obtained with eBURST analyses were in congruence with those of POPART analyses, suggesting MH2 as the primary founder genotype for X. translucens pv. undulosa (Fig. 2; Fig. S5).
DISCUSSION
This study describes the host range and phylogenetic positions of 65 X. translucens strains isolated from gramineous plants in Iran among those of the worldwide population of the pathogen. All the wheat strains isolated in Iran were identified as X. translucens pv. undulosa, while the strains isolated from barley were shown to belong to two pathovars (X. translucens pv. undulosa and X. translucens pv. translucens). Only one (XtKm18) of the X. translucens pv. translucens strains isolated in Iran was pathogenic on wheat, while all of the strains were pathogenic on barley and Harding’s grass. The relatively higher number of wheat strains (49/65) isolated in our surveys might be attributed to the differences in the cultivated areas covered by the two crops and the proportionally higher number of wheat fields surveyed in comparison to the number of barley fields. Similar results in terms of multispecies etiology and variable host range were observed in the bacterial agents infecting solanaceous vegetables (15), edible dry beans (16), and sugar beets (17) in Iran.
As revealed by several diversity parameters (haplotype frequency, haplotype diversity, and percentage of polymorphic sites), the strains isolated in Iran possess significantly higher genetic diversity than the strains isolated in the North American countries. Furthermore, a global-scale phylogeographic analysis supports the hypothesis that Iranian strains of X. translucens have acted as the founder population in the other countries (Fig. S4 and S5 in the supplemental material). Several unique MHs were determined among the Iranian strains, highlighting the allelic richness of the X. translucens population in Iran. All these lines of evidence support the hypothesis that a given plant pathogen possesses higher genetic diversity in the center of origin of its host plant, which is the case for the BLS pathogen. The results of neutrality tests (Fu and Li’s D*and Fu and Li’s F*) on the rpoD gene sequences of non-Iran X. translucens pv. undulosa strains also support the hypothesis, suggesting a selective sweep and population expansion after a recent bottleneck in North America, while additional housekeeping gene sequences need to be examined to further validate the hypothesis (Table 2). Since the center of origin of cultivated wheat was determined to be in southeastern Turkey (10, 11), we speculate that X. translucens should be an endemic pathogen in the Middle Eastern countries, including Iran. Unlike modern crops, such as tomato (Solanum lycopersicum), which have only a short history of cultivation and trade on the global scale, dating to the early 16th century (18), the history of the worldwide distribution of common wheat by humankind is not clear, making the phylogeographic study of cereal pathogens complicated. BLS was reported in the neighbor countries of Iran, and hence, a comprehensive epidemiological investigation will decipher the distribution, population structure, and genetic diversity of X. translucens in the center of origin of cultivated wheat.
According to the eBURST and POPART analyses, MH2—including 56 of 178 strains—was determined to be the founder genotype of X. translucens pv. undulosa (Fig. S5). Interestingly, the pathotype strains of X. translucens pv. undulosa (LMG 892T) and X. translucens pv. secalis (LMG 883T) (both were isolated in southern Canada in 1966) belong to MH2 (Table S1). The pathotype strain of X. translucens pv. undulosa was isolated from Triticum turgidum, while the pathotype strain of X. translucens pv. secalis was isolated from rye (6). It is likely that the two pathotypes originated from the same founding genotype in the North America; however, the prevalence of the “one host-one species” concept in the mid-20th century led to the designation of the two strains as two different taxa at that time (4). Triticum turgidum, from which the pathotype strain of X. translucens pv. undulosa was isolated, is also known as Khorasan wheat (19). The common name refers to Khorasan Province in northeastern Iran and indicates the origin of the crop in the ancient civilizations in the Iranian Plateau or Fertile Crescent (19). Khorasan wheat was accidentally introduced into the United States in 1949 and planted for the first time in Fort Benton, MT, while extensive cultivation of the crop in the United States only started in 1964 (19). Since the founding genotype of the pathogen is still prevalent in the Iranian wheat cultivation areas, one can hypothesize that X. translucens pv. undulosa was transmitted from the Iranian Plateau into the United States via infected seeds of Khorasan wheat in the mid-20th century. All these lines of evidence bring us to the speculation that the existing population of X. translucens pv. undulosa in North America was introduced into the New World from the Iranian Plateau and dispersed throughout the continent during the past 50 years (Fig. S4) (6). However, further genotypic and phylogeographic examinations using a broader set of strains are needed to validate this hypothesis.
In conclusion, the results of the present study revealed a greater diversity among the worldwide population of X. translucens than had been described before. More specifically, based on the phylogeographic analyses, we have noted that the strains isolated in the Iranian Plateau—which is considered the center of origin of cultivated wheat—were genetically more diverse than the strains isolated in the New World. These results suggest that the phylogeny of wheat- and barley-pathogenic members of X. translucens should be reexamined using a larger collection of strains from all the known hosts of the pathogen, including gramineous weeds, from all over the world. Moreover, despite a few preliminary studies (20–23), these findings raise questions about whether there is any source of resistance among the wild population of wheat species (Triticum spp.) in the center of origin of this crop and emphasize at the same time the need for a more detailed investigation in this regard. Only future studies, based on population genetics, comparative genomics, and pathogenicity assays of a wide collection of strains isolated from different geographical regions, can shed more light on these questions.
MATERIALS AND METHODS
Bacterial strains and their identification.The bacterial strains used in this study were obtained during a decade-long (2008 to 2017) series of field surveys across cereal-growing areas of northwestern, western, central, and southern Iran. Furthermore, seed lots of small-grain cereals deposited in the seed banks of the Seed and Plant Certification and Registration Institute (Karaj, Iran) were investigated for BLS infestation. The surveying strategy and sampling procedure were the same as described previously (24, 25). Bacterial strains were isolated from the plant tissues and seed samples using the methods described previously (26) with minor modifications (27). All of the strains (Table 1) were subjected to preliminary phenotypic tests (26). Gram reaction and colony characteristics on yeast extract-dextrose-calcium carbonate (YDC) agar medium were determined. The hypersensitivity reaction (HR) was evaluated on tobacco (Nicotiana tabacum cv. Turkish) leaves, using the bacterial suspension from a 48-h culture on NA medium (26) at a concentration of 108 CFU/ml. Reference strains of X. translucens pv. undulosa (ICMP 11055) and X. translucens pv. translucens (DSM 18974T) were included as controls in all of the tests. Phenotypic tests were repeated twice.
Pathogenicity tests and host range assays.All the bacterial strains (Table 1) were evaluated for their pathogenicity on six gramineous species, including their host of isolation, as well as common gramineous weeds in the region. Barley (cv. Valfajr), Harding’s grass, oat, rye, ryegrass, and wheat (cv. Pishtaz) plants were inoculated with the bacterial suspensions using a procedure described previously (27). The inoculated plants were maintained under greenhouse conditions with 90 to 95% relative humidity and periodically (daily) monitored for symptom development until 25 days postinoculation (dpi). The same numbers of plants were treated with the reference strains of X. translucens pv. translucens (DSM 18974T) and X. translucens pv. undulosa (ICMP 11055) as positive controls. Negative-control plants were treated in the same manner except that sterile distilled water was used instead of the bacterial suspensions. Koch’s postulates were accomplished by reisolating the inoculated strains from plants showing symptoms on YDC medium. Confirmation of the reisolated bacteria was made using X. translucens-specific PCR primers T1/T2 (Table 3) (7). The pathogenicity tests were repeated twice.
Primer pairs used in this study
Molecular phylogenetic analysis.All strains showing xanthomonad-like characteristics (i.e., pale-yellow mucoidal colonies on YDC medium and hypersensitivity reaction on tobacco leaves) were subjected to the diagnostic PCR test using the X. translucens-specific primer pair T1/T2 (Table 3) (7). The bacterial DNA extraction and PCR procedures were described previously (27); the primer sequences and annealing temperatures are provided in Table 3. Sixty-five X. translucens strains were identified using the species-specific PCR primer and were subjected to MLSA/MLST analyses using the sequences of four housekeeping genes (dnaK, fyuA, gyrB, and rpoD), as recommended previously (6).
To determine the phylogenetic positions of the strains isolated in Iran and access a global-scale phylogeography for the BLS pathogen, sequences of the four housekeeping genes in a collection of 112 worldwide X. translucens strains were retrieved from the GenBank database and included in the phylogenetic analyses (Table S1 in the supplemental material) (6, 14). The sequences were concatenated following the alphabetic order of the genes, resulting in a sequence of 2,464 bp, as follows: nucleotides 1 to 763 for dnaK, 764 to 1285 for fyuA (522 bp), 1286 to 1790 for gyrB (505 bp), and 1791 to 2464 for rpoD (674 bp) genes. Phylogenetic trees were constructed using MrBayes 3.1.0, running the chain for 1 × 106 generations (bootstrapping at 1,000 replications) and setting the burn-in at 1,000 (28). The substitution model for Bayesian analysis was determined, using MrModelTest version 2.3 software, to be AIC+[GTR+I] (29). The resulting phylogenetic trees were annotated and presented using MEGA 6.06 software (30).
Genetic diversity and recombination analyses.Nucleotide diversity, numbers of STs and MHs, haplotype diversity, haplotype frequency, percentages of polymorphic sites, minimum numbers of recombination events, and the ratio of the rate of nonsynonymous to synonymous mutations were estimated using DnaSP 5.10 software (31). The class I neutrality tests (Tajima’s D, Fu and Li’s D*, and Fu and Li’s F* statistics) were also calculated for detecting departure from the mutation-drift equilibrium (31). Detection of potential recombinant sequences and identification of likely parental sequences were carried out using a set of seven nonparametric detection methods (RDP, Geneconv, MaxChi, Chimera, BootScan, SiScan, and 3Seq) implemented in Recombination Detection Program (RDP) version 4.80 (32). The analysis was performed with the default settings for the different detection methods, and the Bonferroni-corrected P value cutoff was set at 0.05. Recombination events were accepted when they were identified by at least four of the seven detection methods (32). A splits-decomposition network was constructed, and the pairwise homoplasy index (PHI) was calculated using SplitsTree version 4.14.4 (33). These calculations were performed once for 66 Iranian strains and again for the entire data set (178 strains) for all the individual genes, as well as for the concatenated sequences (33).
Phylogeographic analyses.To visualize the phylogeographic relationships between the strains, haplotype networks were inferred for individual housekeeping genes using the TCS algorithm (34) implemented in POPART version 1.7 (35). The geographic distribution of 178 strains collected in seven countries (Australia, Canada, India, Iran, Paraguay, Mexico, and the United States) was employed in building the haplotype network of each of the four housekeeping genes, as well as concatenated sequences as described by Leigh and Bryant (35). The country-specific STs, as well as the shared MHs between the countries, were estimated based on the resulting haplotype networks (Table S1). Furthermore, a global-scale geographical structure of X. translucens was visualized using the eBURST software (36). The eBURST analysis was used to explore the patterns of evolutionary descent among the strains used in this study as described by Feil et al. (36). The analyses were performed using the stringent (default) group definition, in which sequence types are included within the same clonal complex only if they share identical alleles at three or four of the four MLST loci with at least one other ST in the population. Furthermore, as a default setting, 1,000 resamplings were performed for bootstrapping the resultant evolutionary network.
Accession number(s).The nucleotide sequences of the strains isolated in Iran were deposited in the GenBank database with the following accession numbers: MK622011 to MK622075 for dnaK, MK622076 to MK622140 for fyuA, MK622141 to MK622205 for gyrB, and MK622206 to MK622270 for rpoD sequences. The GenBank accession numbers for the 112 strains of worldwide origin are listed in Table S1 in the supplemental material.
ACKNOWLEDGMENTS
Financial support for this study was coprovided by Shiraz University and Vali-e-Asr University of Rafsanjan, Iran. C.B., R.K., and E.O. benefited from interactions promoted by COST Action CA16107 EuroXanth (https://euroxanth.eu/), supported by COST (European Cooperation in Science and Technology).
E.O. and M.K. conceived and designed the study, with assistance from S.M.T., H.H., P.K., and R.K. M.K. carried out the experiments. E.O. analyzed and interpreted the data with assistance from M.K., H.H., R.K., and C.B. M.K. and E.O. prepared the paper, with assistance from C.B., R.K., and G.C. E.O. revised the final manuscript and acted as the corresponding author.
The authors declare no competing financial interests.
FOOTNOTES
- Received 8 July 2019.
- Accepted 6 August 2019.
- Accepted manuscript posted online 16 August 2019.
Supplemental material for this article may be found at https://doi.org/10.1128/AEM.01518-19.
- Copyright © 2019 American Society for Microbiology.
