http://orcid.org/0000-0002-9809-2163
Figures
Abstract
The wheat curl mite, Aceria tosichella Keifer, one of the most destructive arthropod pests of bread wheat worldwide, inflicts significant annual reductions in grain yields. Moreover, A. tosichella is the only vector for several economically important wheat viruses in the Americas, Australia and Europe. To date, mite-resistant wheat genotypes have proven to be one of the most effective methods of controlling the A. tosichella—virus complex. Thus, it is important to elucidate A. tosichella population genetic structure, in order to better predict improved mite and virus management. Two genetically distinct A. tosichella lineages occur as pests of wheat in Australia, Europe, North America, South America and the Middle East. These lineages are known as type 1 and type 2 in Australia and North America and in Europe and South America as MT-8 and MT-1, respectively. Type 1 and type 2 mites in Australia and North America are delineated by internal transcribed spacer 1 region (ITS1) and cytochrome oxidase I region (COI) sequence differences. In North America, two A. tosichella genotypes known as biotypes are recognized by their response to the Cmc3 mite resistance gene in wheat. Aceria tosichella biotype 1 is susceptible to Cmc3 and biotype 2 is virulent to Cmc3. In this study, ITS1 and COI sequence differences in 25 different populations of A. tosichella of known biotype 1 or biotype 2 composition were characterized for ITS1 and COI sequence differences and used to model spatio-temporal dynamics based on biotype prevalence. Results showed that the proportion of biotype 1 and 2 varies both spatially and temporally. Greater ranges of cropland and grassland within 5000m of the sample site, as well as higher mean monthly precipitation during the month prior to sampling appear to reduce the probability of occurrence of biotype 1 and increase the probability of occurrence of biotype 2. The results suggest that spatio-temporal modeling can effectively improve A. tosichella management. Continual integration of additional current and future precipitation and ground cover data into the existing model will further improve the accuracy of predicting the occurrence of A. tosichella in annual wheat crops, allowing producers to make informed decisions about the selection of varieties with different A. tosichella resistance genes.
Citation: Khalaf L, Timm A, Chuang W-P, Enders L, Hefley TJ, Smith CM (2020) Modeling Aceria tosichella biotype distribution over geographic space and time. PLoS ONE 15(5): e0233507. https://doi.org/10.1371/journal.pone.0233507
Editor: Satyanarayana Tatineni, USDA Agricultural Research Service, UNITED STATES
Received: December 31, 2019; Accepted: May 6, 2020; Published: May 29, 2020
Copyright: © 2020 Khalaf et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The nucleotide sequences of ITS1 region and COI used in phylogenetic analyses have been deposited in GenBank (accession numbers MT336812-MT337241 for ITS1 region, and MT370025-MT370073 for COI).
Funding: This work was supported by USDA/NIFA North Central IPM Program award 2013-34103-21208 to CMS, the Kansas Wheat Alliance, funding by the University of Baghdad to LK, and Kansas State University Research and Extension. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
The wheat curl mite, Aceria tosichella Keifer, is a global pest of bread wheat Triticum aestivum L. The mite reduces grain yield by causing both direct feeding damage and as a vector of several viral wheat pathogens [1–7]. Yield losses caused by A. tosichella feeding may be up to 30% [4, 8] due to leaf rolling and trapping [9]. Aceria tosichella transmits three damaging viruses to the wheat plant—Wheat Streak Mosaic Virus (WSMV, family Potyviridae, genus Tritimovirus), High Plains wheat mosaic virus (HPWMoV, genus Emaravirus, formerly High plains virus; www.ictvonline.org/proposals-15/2015.018aP.A.v3.Emaravirus_sp.pdf), and Triticum Mosaic Virus (TriMV, family Potyviridae, genus Poacevirus). Aceria tosichella nymphs obtain WSMV after feeding for as little as 30 min on infected plants and can spread the virus for at least 7d postfeeding. [9–13]. Although WSMV infections occur at a greater incidence than HPWMoV or TriMV [14], co-infections are common [15–17]. WSMV causes wheat yield losses ranging from 2.5 to 7% on at least five different continents [18] depending on climate, virus acquisition time and wheat cultivar [13, 19–24]. Aceria tosichella detection is complicated by the mite’s small size (150–225 μm length), ability to attain maximum concealment through cryptic behavior and wide host range [10, 18, 25–31]. To date, no effective acaricides exist to manage A. tosichella and the viruses it transmits [32, 33]. The cultural practice of controlling over-summering hosts such as volunteer wheat and weed grass hosts can provide effective management of the A. tosichella-virus complex if producers use this management approach [34].
Aceria tosichella is a complex global mixture of at least 29 different genetic lineages [28, 30, 31]. However, two genetically distinct lineages occur as pests of wheat in Australia, Europe, North America, South America and the Middle East. These lineages are known as type 1 and type 2 in Australia and North America and in Europe and South America type 1 and type 2 are known as MT-8 and MT-1, respectively [28]. In North America, internal transcribed spacer 1 region (ITS1) sequence differences were used to delineate two lineages in mites collected in Kansas, Montana, Alberta Canada, and Nebraska [35]. More recently, type 1 and type 2 lineages have been delineated using ITS1- and cytochrome oxidase I region (COI) sequence differences [29, 36]. Evidence also exists to show that the A. tosichella type 1 and type 2 lineages differ in their ability to transmit WSMV, HPWMoV and TriMV in Australia and North America [37, 38, 39].
The development and use of A. tosichella-resistant wheat cultivars to reduce A. tosichella populations and WSMV infection has progressed in North America since 1995 [40–45]. During this process, North America A. tosichella populations also began to be referred to as biotypes because of the ability of one biotype to overcome the effect of (exhibit virulence to) A. tosichella resistance gene(s) in wheat. Currently, biotype 1 is referred to as being avirulent to the effects of the rye:wheat translocation resistance gene (Cmc3 [curl mite colonization]) in the wheat variety TAM 107. Biotype 2 is recognized as being virulent to Cmc3 [46].
Virulence in A. tosichella to Cmc3 has remained stable for the past 20 years [34] but recent field assessments [47] determined that 24% of A. tosichella collected from multiple locations in North America are virulent to Cmc3. Therefore, there is a real need for new information about the current geographic distribution of A. tosichella biotypes or genetic lineages throughout the U. S. Great Plains and the potential changes occurring in each. In order to obtain new knowledge for more effective A. tosichella management programs, a regional study was conducted to assess the current genetic variation of A. tosichella. Our hypothesis was that changes in A. tosichella genetic composition are dynamic and have spatial and temporal structure that may be correlated with climate and landscape features. To test this hypothesis, experiments were conducted to assess the geographic distribution of A. tosichella in six U. S. Great Plains wheat-producing states in 2014 and 2015 based on internal transcribed spacer 1 (ITS1) region and cytochrome oxidase I (COI) polymorphisms and plant phenotypic reactions. An additional experiment was conducted to compare in-depth sequence analyses of A. tosichella populations at four locations in Kansas, Missouri and Nebraska in 2016 to determine variation over local scales. Finally, temporal variation in A. tosichella lineages over a 2-year period was used to develop a generalized additive spatio-temporal model to predict the prevalence of biotypes 1 and 2 in the Great Plains.
Results
Phylogenetic analyses
A total of 430 A. tosichella were collected in 2014, 2015, and 2016 from 12 locations in 2014, 13 locations in 2015, and 12 locations in 2016 (S1 and S2 Tables). These samples yielded regions of 618 bases for ITS1 analysis in all samples and 506 bases for COI analysis in 49 samples. All mites analyzed were A. tosichella [29, 36]. Bayesian phylogenetic analyses revealed clearly distinct differences between biotypes, based on 8 ITS1 haplotypes and 3 COI haplotypes. Genbank sequences that were used for comparison confirmed biotype 1 and 2 designations, as well as MT-8 and MT-1 designations [48]. Genetic distance values between haplotypes based on ITS1 polymorphism ranged from 0.003 to 0.028 (S3 Table), whereas these values ranged from 0.002 to 0.177 based on COI polymorphism (S4 Table).
Aceria tosichella biotype distribution
Biotypes 1 and 2 were present in 2014 and 2015 in all sampling sites with the exception of six fields in Missouri and two fields in Texas, where only biotype 2 was present (Fig 1A). Biotype 1 occurred as five haplotypes and biotype 2 was present as three haplotypes. One biotype 1 haplotype with a 1 base pair (bp) difference to the primary haplotype was observed in about one third of the populations and was present primarily in Kansas and Missouri. Although biotype 2 was present in six states (Fig 1A), most of the variation in this biotype was observed in Kansas and Texas. Interestingly, a new biotype 1 haplotype differing by 1bp to the dominant haplotype appeared in all 3 years of sampling (Fig 1A and 1B).
(A) and 2016 (B) within biotypes 1 and 2. Smaller circles indicate fewer individuals in a haplotype. Hash marks on lines connecting haplotypes symbolize base-pair differences.
Each biotype was present at all sample locations during the study, with the exception of two locations in Nebraska and two locations in Texas in 2014 and 2015 (Table 1). Individual grain head samples obtained in 2016 from Missouri, Nebraska and Kansas also contained both biotypes Table 2). However, biotype 1 occurred in greater frequency in Missouri locations, where the average ratio of mites sampled at six locations was 70% biotype 1 and 30% biotype 2 (Table 1). A higher probability of occurrence of biotype 2 occurred at four Nebraska locations (20% biotype 1, 80% biotype 2; and at two Texas locations that each yielded 100% biotype 2 (Table 1). Biotype ratios were evenly distributed in Kansas and South Dakota (50% for each biotype); and in North Dakota (40% biotype 1, 60% biotype 2) (Table 1).
A total of three fields at each location were sampled, five individuals were collected at each of three sites in each field for a total of 45 individuals per field.
Mites collected from individual wheat heads in fields in Kansas, Missouri and Nebraska during 2016 were mixtures of both biotypes, indicating that both occur on a single grain head simultaneously (Fig 2). Of 36 heads examined, 28 contained both biotypes. However, eight heads contained a single biotype. Seven of these eight heads contained only biotype 1 and occurred in Barton County Missouri (field 3, head 2); Cape Girardeau County Missouri (field 1 heads 2 and 3, field 2 head 3; field 3 heads 1 and 3) and in Hayes County Nebraska (field 2 head 2) (Table 2) (Fig 3A and 3B). In a single instance, only biotype 2 was found on head 2 collected in Cape Girardeau County Missouri field 3 (Fig 3C).
(A) Barton county Missouri field 1 and (B) field 2; and (C) Ellis County Kansas field 1. Each field contained biotype 1 and biotype 2 in all heads sampled. Smaller circles indicate fewer individuals in a haplotype. Hash marks on lines connecting haplotypes symbolize base-pair differences.
(A) Cape Girardeau county Missouri field 1 containing biotype 1 in all heads and biotype 2 in one head only; (B) Cape Girardeau county Missouri field 2 containing biotype 1 in two heads and biotype 2 in all heads; (C) Cape Girardeau county Missouri field 3 containing biotype 1 in two heads and biotype 2 in one head only. Smaller circles indicate fewer individuals in a haplotype. Hash marks on lines connecting haplotypes symbolize base-pair differences.
Samples of individual grain heads collected in 2016 in Nebraska, Kansas and Missouri revealed some changes in biotype ratios from those in 2015. In Hayes County Nebraska the ratio changed slightly from 4:6 in 2015 to 5:5 in 2016; and in Cape Girardeau County Missouri, the ratio remained primarily biotype 1, shifting from 9:1 to 8:2 (Tables 1 and 2). Similarly, the ratio shifted slightly from 8:2 in 2015 to 6:4 in 2016 in Barton County Missouri. However, the ratio in Ellis County Kansas shifted significantly from 7:3 in 2015 to 4:6 in 2016 (Tables 1 and 2).
Predicted A. tosichella biotype occurrence
The probability of occurrence of each biotype showed distinct spatio-temporal patterns, which were influenced by predictor variables (Fig 4). The percentage of grass/pasture and crop land within 5000 m of a sample location had negative coefficient estimates with 90% confidence intervals (CI) that did not include zero. Negative coefficient estimates indicate that as the predictor variable increases (e.g., percentage of crop land with 5000 m) the probability of an A. tosichella individual being biotype 1 decreases. Similar to the land cover predictor variables, precipitation during the month of sampling and the month prior to sampling had negative coefficient estimates, but in this case the upper limit of the 90% CI was slightly greater than zero, indicating less certainty regarding the sign and magnitude of the effects. All other climate predictors (i.e., mean monthly temperature) and landscape effects (i.e., the interaction of crop land and grass/pasture) had 90% CI that indicated a substantial level of uncertainty about the sign and magnitude of the effect. We refrain from interpreting the impact of these variables on the occurrence of the biotypes. In summary, our results demonstrate that increases in crop and grass/pasture within 5000 m of the sample location along with monthly precipitation during the month prior to and month of sampling reduced the probability of occurrence of biotype 1, and consequently increased the probability of occurrence of biotype 2 in some locations in Nebraska, North Dakota, South Dakota, and Kansas. Fig 4 shows the expected probability of each biotype at each sample location assuming a given point was wheat and that mites were present at that location.
Maps do not consider specific agricultural landscape and can be interpreted as the potential number of biotype 1 or 2 expected if a given map location was wheat. Red areas—higher probability of biotype 1 presence; blue areas—higher probability of biotype 2 presence. Note: These data were collected during the periods May 21 to July 10, 2014; June 25 to July 12, 2015; and June 245 11 to June 16, 2016. As a result, the predictive maps are an extrapolation if used to infer dynamics outside of the time periods in which these data were collected.
The spatio-temporal statistical model predicted that the probability of occurrence of biotype 1 was particularly high at the majority of dates and locations in Missouri and South Dakota in 2014, 2015 and 2016 (Fig 4, S1A–S1E and S2A–S2C Figs). In contrast, the model predicted a decreased probability of occurrence of biotype 1 in 2014 and 2015 in North Dakota and a gradual increase of biotype 1 probability by the end of 2016 (Fig 4, S3A and S3B Fig). An even greater decrease in the probability of occurrence of biotype 1 was predicted at the majority of locations in Kansas and Nebraska (Fig 4B–4I). For example, predicted probability of biotype 1 occurrence was low at most locations in Kansas and Nebraska in 2015 (Fig 4D–4F), but increased to a nearly equal probability of occurrence of each biotype in 2016 (Fig 4G–4I, S4A–S4C and S5 Figs) with the exception of the eastern-most sample location in Nebraska, where biotype 2 was predicted to be predominant in 2015 and 2016 (Fig 4D–4I, S4D Fig). Finally, the model predicted biotype 2 occurrence at the two Texas locations in all three sample years (Fig 4, S5 Fig).
Discussion
Sequencing ITS1 and COI polymorphisms in A. tosichella samples obtained in the current study indicate that these genes remain useful genomic regions for A. tosichella biotype discrimination. Results of phylogenetic analyses based on polymorphisms of each gene support previous conclusions that A. tosichella has two dominant haplotypes in North America [29, 36] with divergence in the ITS1 region similar to the Australian haplotypes EU734729 (WCM1) and EU734726 (WCM2) [29]. Both ITS1 and COI polymorphisms indicated that A. tosichella samples obtained in the current study corresponded to lineages MT-1 and MT-8 based on global sampling [48] although ITS sequences generated for the current study were significantly longer (~350bp vs ~617bp). However, several other haplotypes appear to be divergent and may potentially represent additional biotypes [29, 30, 49]. Thus, it was not surprising to determine that the ITS1 and COI polymorphisms determined in our results identified eight new A. tosichella haplotypes from wheat with no matches in the GenBank database. These results also confirm previously identified large-scale co-occurrences of biotype 1 and 2 in individual field populations in North America [29, 36, 37], as well as within individual wheat grain heads [50].
Australian types 1 and 2 also co-occur across Australian wheat production areas, although biotype 1 occurs more often in the southeast and biotype 2 more frequently in the west [29, 37]. Results of our experiments further demonstrate possible genetic drift or host shifts in North American A. tosichella populations [51], although variation is less than that in observed previously in Turkey [52]. Possible explanations include genetic drift resulting from an invasive A. tosichella population [53, 54] or a host shift resulting from A. tosichella adaptation to mite resistance genes in wheat [55]. The greater numbers of unique ITS1 region haplotypes in biotype 1 than in biotype 2 (Figs 1, 2 and 3), support conclusions by Harvey et al. [46] that biotype 2 developed after deployment of the Cmc3 A. tosichella resistance gene in wheat cultivar TAM 107 in 1983 [56].
Our results also revealed eight haplotypes based on ITS1 region sequence variants, while those of Hein et al. [36] determined only two. The differences in results between the two studies are likely due to differences in geographic scope of sampling (25 sample sites in our experiments versus 5 sites in Hein et al.) and the year of sample collection (2014–2016 versus 1999). Finally, differences in the results of the two studies may involve an increase in biotype diversity resulting from the recent release of cultivars containing the Cmc4 gene for A. tosichella resistance in Montana and Oklahoma [57, 58], as well as cultivation of cultivars in Colorado, Kansas, Oklahoma and Texas with the Dn7 gene for resistance the Russian wheat aphid, Diuraphis noxia (Kudjumov), and the H21 gene for resistance to Hessian fly, Mayetiola destructor Say, both of which have recently been shown to be resistant to A. tosichella [59].
Enders et al. [60] employed a flexible phenological model similar to that used in our experiments to account for spatial or temporal autocorrelation that may have been generated by population dynamics of different cereal aphid virus vectors. Adoption of such models on an area-wide basis in North America could provide an enhanced understanding of A. tosichella biotype geographic distribution and improve predictions of the risk of infestations by A. tosichella and the viruses they transmit.
Previous reports of the presence of each A. tosichella biotype in North America were based on data from mites collected at five locations in one year [36, 61]. In contrast, our experiments are the first to determine A. tosichella presence at 38 unique locations from a 1.2 mill km2 area in the North American Great Plains over a 3-year period. As a result, these data provide the first ratios of the two biotypes over multiple locations and years. The use of these data to develop spatial-temporal predictions of A. tosichella biotype variation provide the first demonstration of the effects of precipitation and land cover on biotype distribution. The predicted biotype ratios based on 2014, 2015, and 2016 infestations necessitate continued and coordinated monitoring of North American A. tosichella biotype variation in order to anticipate future mite infestation intensity and biotype composition.
Materials and methods
Sample collection
Aceria tosichella was collected from wheat T. aestivum heads at 25 locations in the U.S. Great Plains wheat production area from May 21 to July 10, 2014; June 25 to July 12, 2015; and June 11 to June 16, 2016. No special permits were required to collect samples, as verbal permission was given by producers at each sample collection site. The geographic coordinates of each sample location in 2014 and 2015 (S1 Table) or 2016 (S2 Table) were recorded using a hand-held GPS device. In 2014 and 2015, three field sites were sampled at each GPS location and 30 wheat heads were sampled within each of the three field sites, resulting in a total of 90 heads per location. To avoid bias, the heads were grouped, and 10 heads were arbitrarily selected from each 90 head group. One individual live female was transferred using a 30X microscope, from each of the 10 selected wheat heads to a cold microcentrifuge PCR tube and centrifuged at 14,000 rpm at 4°C for 1 min to position the mite near or in the bottom of the tube before storage at -80°C. An 8 h recess was observed between transfers to prevent cross-contamination between populations [9]. In 2016, additional collections were made at one location in Kansas, two locations in Missouri, and one location in Nebraska at the same locations as in 2014 and 2015, or in the nearest wheat field in locations where wheat had been rotated (S2 Table). Three wheat fields were sampled at each location and in each field, five heads were sampled from plants at each of three different sites in each field, resulting in a total of 45 heads per location. Each head was kept separate in a plastic bag in order to distinguish genetic differences between mites within a field and a grain head. Mites from each location were previously classified as biotype 1 avirulent to Cmc3 or biotype 2 virulent to Cmc3 [47] using methods of Harvey et al. [46].
Aceria tosichella DNA processing and amplification
Aceria tosichella DNA was extracted using the MyTaq™ Extract-PCR kit (Bioline USA Inc. Taunton, MA). A master mix was prepared for each reaction using 35 μl nuclease-free water (Ambion Co., Lewisville, TX), 10 μl Buffer A and 5 μl Buffer B (total 50 μl). This solution was added to each tube containing a specimen of A. tosichella. Tubes were incubated at 75°C and 95°C for 10 min each and thereafter held at 12°C for ∞. Mite DNA extracts were stored at 4°C. Polymerase chain reactions were performed to amplify 618 base pairs (bp) of the nuclear ribosomal internal transcribed spacer 1 (ITS1) region. Primer3Plus [62] was used to design primers to amplify 600 bp of this gene (Table 3). Using primers indicated in Table 3, a subsample of specimens was subjected to cytochrome oxidase I (COI) analysis to confirm whether biotype groupings/designations were correct, and to determine whether analysis of this gene was more adept at detecting variation at local geographic scales. For this reason, the majority of specimens (42 out of a total of 49) for which the COI gene was sequenced were collected in 2016.
All PCRs were conducted in a 40 μl volume including 1 μl DNA extract, 20 μl Taq DNA polymerase (Bioline Inc. Taunton, MA), 0.5 pmol each of the forward and reverse primers (Table 3), 1 μl MgCl2 (Thermo Scientific, New Hampshire, MA), and 17 μl nuclease-free water, using a T100 thermal cycler (Bio-Rad, Hercules, CA). The ITS1 region amplification protocol was 95°C for 3 min (initial denaturation), four cycles of 95°C for 20 sec, 56°C for 15 sec, 72°C for 20 sec, followed by 34 cycles of 95°C for 20 sec, 45°C for 15 sec, 72°C for 20 sec, and 72°C for 15 min. The COI amplification protocol was 95°C for 3 min (initial denaturation), 40 cycles of 95°C for 20 sec, 45°C for 15 sec, 72°C for 20 sec, and 72°C for 15 min. 5 μl of each PCR product was mixed with 1 μl loading dye (Promega, Madison, WI) and run on a 1% agarose gel (Fisher Scientific, Suwanee, GA), stained with GelGreen® Nucleic Acid Gel Stain (Bioline Inc. Taunton, MA) for 60 min and visualized under UV light (Bio-Rad Gel Doc EZ System Gel Imaging System, San Jose, CA) to determine amplification products. PCR product sizes were assessed using the Hi-Lo™ DNA marker (Minnesota Molecular, Inc. Minneapolis, MN) and product concentration was measured by comparison with Lambda DNA of standard concentrations (Promega) and Nanodrop spectrophotometry (Thermo Scientific).
Sanger sequence data were then generated using GeneWiz Inc. (South Plainfield, NJ). Because of large sample sizes, PCR products were sequenced for a few specimens in both directions (F and R) using the same primers used for PCR. However, the majority of our specimens were sequenced in one direction (F) only. Sequences for A. tosichella and related species obtained from GenBank were aligned and edited using BioEdit V. 7 software [63].
Bayesian phylogenetic analyses
Bayesian phylogenetic analyses of the data were performed using MrBayes 3.2 [64]. DNAsP v. 5.10.01 [65] was used to test sequence polymorphism among individuals with ITS1 region and COI, and ITS1 region sequences were imported as nexus files to POPART [66] to create haplotype network diagrams. The nucleotide sequences of ITS1 region and COI used in phylogenetic analyses have been deposited in GenBank (accession numbers MT336812-MT337241 for ITS1 region, and MT370025-MT370073 for COI).
To delineate between biotype 1 and 2 ITS1 region nucleotide sequences, comparisons were made using GenBank sequence EU734729 serving as a reference for biotype 1 and EU734726 serving as a reference for biotype 2 [29]. For this gene, comparisons were also made using sequences that correspond to lineages MT-1, MT-7 and MT-8 [48]. To delineate between biotype 1 and 2 COI nucleotide sequences, comparisons were made using GenBank sequence JQ248920 serving as a reference for biotype 1. No explicit reference sequence for biotype 2 was available for this gene at the time of the analyses. However, comparisons were also made for this gene using sequences from lineages MT-1, MT-7 and MT-8 [48]. Genbank sequence JF920113 obtained from Aceria eximia was used as an outgroup control in the ITS1 sequence analyses, and Genbank sequence FJ387563 obtained from Aceria tulipae was used as an outgroup control in the COI sequence analyses. The threshold of sequence similarity required to determine a biotype was 99% identical to a known biotype ITS sequence.
Spatio-temporal prediction of A. tosichella biotype
A generalized additive model [67] was used to capture the spatio-temporal dynamics in the prevalence of A. tosichella biotypes 1 and 2, incorporating weather and land cover as dependent variables with temporal changes in A. tosichella population dynamics. A binomial distribution was assumed, with the number of “trials” of the binomial distribution being the number of mites sampled at each unique site and time period, which was 10 in 2014–2015 and 15 in 2016.
For each sample obtained, the PRISM database [68] was used to obtain the average monthly temperature and precipitation occurring during the month and the month prior to sample collection, and the 2011 National Land Cover Database [69] was used to determine either grass/pasture or cropland land cover covariates at the 30 m by 30 m resolution. NLCD classes 71 and 82 defined grass/pasture and class 42 defined cropland. Land cover was assumed to influence mite prevalence at a scale larger than 30 m x 30 m resolution. The effective scale influencing the response was determined by calculating the percentage of grass/pasture and cropland within circular regions centered at the sample location with a diameter of 100-, 500-, 1000-, 2500-, 5000-, and 10000m.
Spatio-temporal effects unrelated to weather or land cover covariates i.e., autocorrelation [70] were included using a categorical factor composed of the year of data collection and thin plate regression splines, a type of basis function that models “smooth” effects of spatial location or time [70]. The interaction between grass/pasture and cropland land cover at the 500m scale was included in a given model, but candidate models were constructed for spatial scales at 100-, 500-, 1000-, 2500-, 5000-, and 10000m. The appropriate scale was chosen from the candidate model with the lowest Akaike's information criterion (AIC) score [71] and calculating the AIC. The drivers of the prevalence of each A. tosichella biotype were assumed to covariates with coefficients within 90% confidence intervals that did not contain zero.
Supporting information
S1 Fig. Probability of A. tosichella biotype 1 occurrence (mean ± SE) at locations in Barton (A), Pettis (B), Cape Girardeau (C), Pike (D), Copper (E) and Stoddard (F) county Missouri in 2014, 2015 and 2016.
https://doi.org/10.1371/journal.pone.0233507.s001
(PPTX)
S2 Fig. Probability of A. tosichella biotype 1 occurrence (mean ± SE) at locations in Hughes (A), Lake (B) and Tripp (C) county South Dakota in 2014, 2015 and 2016.
https://doi.org/10.1371/journal.pone.0233507.s002
(PPTX)
S3 Fig. Probability of A. tosichella biotype 1 occurrence (mean ± SE) at locations in Bottineau (A) and Ward (B) county North Dakota in 2014, 2015 and 2016.
https://doi.org/10.1371/journal.pone.0233507.s003
(PPTX)
S4 Fig. Probability of A. tosichella biotype 1 occurrence (mean ± SE) at locations in Cheyenne (A), Hayes (B), Furnas and Saunders (D) county Nebraska in 2014, 2015 and 2016.
https://doi.org/10.1371/journal.pone.0233507.s004
(PPTX)
S5 Fig. Probability of A. tosichella biotype 1 occurrence (mean ± SE) at locations in Barton (A), Ellis (B), Geary (C), Dickinson (D), Greeley (E), Ellsworth (F), Saline (G) and Finney (H) county Kansas in 2014, 2015 and 2016.
https://doi.org/10.1371/journal.pone.0233507.s005
(PPTX)
S6 Fig. Probability of A. tosichella biotype 1 occurrence (mean ± SE) at locations in Dallam (A) and Randall (B) county Texas in 2014, 2015 and 2016.
https://doi.org/10.1371/journal.pone.0233507.s006
(PPTX)
S1 Table. State, county, and geographic coordinates for locations of A. tosichella samples collected in 2014 and 2015.
https://doi.org/10.1371/journal.pone.0233507.s007
(DOCX)
S2 Table. State, county, and geographic coordinates for locations of A. tosichella samples collected in 2016.
https://doi.org/10.1371/journal.pone.0233507.s008
(DOCX)
S3 Table. Estimates of genetic distance and genetic identity in eight U.S. A. tosichella populations collected in 2014, 2015 and 2016, using variation among unique ITS1 haplotypes.
https://doi.org/10.1371/journal.pone.0233507.s009
(DOCX)
S4 Table. Estimates of genetic distance three U.S. A. tosichella populations collected in 2014, 2015 and 2016, using variation among unique COI haplotypes.
https://doi.org/10.1371/journal.pone.0233507.s010
(DOCX)
Acknowledgments
Authors gratefully acknowledge the USDA/NIFA North Central IPM Program, the Kansas Wheat Commission, the Kansas Association of Wheat Growers, University of Baghdad and Kansas State University Research and Extension for supporting this research. Contribution no. 20-114-J from the Kansas Agricultural Experiment Station.
References
- 1. Nault LR, Styer WE. Transmission of an Eriophyid-borne wheat pathogen by Aceria tulipae. Phytopathol. 1970; 60:1616–1618.
- 2. Harvey TL, Martin TJ. Resistance to the wheat curl mite (Acari: Eriophyidae) in common wheat. Cer Res Comm. 1992; 20:63–66.
- 3.
Amrine JW, Stasny TA. Catalog of the Eriophyoidea (Acarina: Prostigmata) of the World. Indira Publishing House, West Bloomfield. 1994; 798pp.
- 4. Harvey TL, Martin TJ, Seifers DL. Effect of nonviruliferous wheat curl mites on yield of winter wheat. J Agric Urban Entomol. 2000; 17:9–13.
- 5. Seifers DL, Martin TJ, Harvey TL, Fellers JP, Michaud JP. Identification of the wheat curl mite as the vector of Triticum mosaic virus. Plant Dis. 2009; 93:25–29.
- 6. Murugan M, Cardona PS, Duraimurugan P, Whitfield AE, Schneweis D, Smith CM. Wheat curl mite resistance: interactions of mite feeding with Wheat streak mosaic virus infection. J Econ Entomol. 2011; 104:1406–1414.
- 7. Navia D, De Mendonça RS, Skoracka A, Szydło W, Knihinicki D, Hein GL, et al. Wheat curl mite, Aceria tosichella, and transmitted viruses: an expanding pest complex affecting cereal crops. Exp Appl Acarol. 2013; 59:95–143.
- 8. Harvey TL, Martin TJ, Seifers DL. Wheat yield reduction due to wheat curl mite (Acari: Eriophyidae) infestations. J Agric Urban Entomol. 2002; 19:9–13.
- 9. Orlob GB. Feeding and transmission characteristics of Aceria tulipae Keifer as vector of Wheat streak mosaic virus. J. Phytopathol. 1966; 55:218–238.
- 10. Seifers DL, Harvey TL, Martin TJ, Jensen SG. Identification of the wheat curl mite as the vector of the High Plains virus of corn and wheat. Plant Dis. 1997; 81:1161–1166.
- 11. Seifers DL, Harvey TL, Martin TJ, Jensen SG. A partial host range of the High Plains virus of corn and wheat. Plant Dis. 1998; 82:875–879.
- 12. Hadi BAR, Langham MAC, Osborne L, Tilmon KJ. Wheat streak mosaic virus on wheat: biology and management. J Integrated Pest Manag. 2011; 2:J1–J5.
- 13.
Appel J, DeWolf E, Bockus W, Todd T. Preliminary 2015 Kansas wheat disease loss estimates. Kansas Cooperative Plant Disease Survey. 2015. https://agriculture.ks.gov/docs/default-source/pp-disease-reports
- 14. Seifers DL, Martin TJ, Fellers JP. Occurrence and yield effects of wheat infected with Triticum mosaic virus in Kansas. Plant Dis. 2011; 95:183–188.
- 15. Mahmood T, Hein GL, Jensen SG. Mixed infection of hard red winter wheat with high plains virus and Wheat streak mosaic virus from wheat curl mites in Nebraska. Plant Dis. 1998; 82:311–315.
- 16. Seifers DL, Martin TJ, Harvey TL, Fellers JP, Stack JP, Ryba-White M, et al. Triticum mosaic virus: A new virus isolated from wheat in Kansas. Plant Dis. 2008; 92:808–817.
- 17. Byamukama E, Seifers DL, Hein GL, DeWolf E, Tisserat NA, Langham MAC, et al. Occurrence and distribution of Triticum mosaic virus in the central Great Plains. Plant Dis. 2013; 97:21–29.
- 18.
Navia D, Ochoa R, Welbourn C, Ferragut F. Adventive eriophyoid mites: a global review of their impact, pathways, prevention and challenges. In Eriophyoid Mites: Progress and Prognoses. 2009; (pp. 225–255). (Eds) Springer Netherlands.
- 19. Tosic M. Virus diseases of wheat in Serbia. Phytopathol. 1971; 70:145–162.
- 20. Sim T IV, Willis WG, Eversmeyer MG. Kansas plant disease survey. Plant Dis. 1988; 72:832–836.
- 21. Christian ML, Willis WG. Survival of Wheat streak mosaic virus in grass hosts in Kansas for wheat harvest to fall wheat emergence. Plant Dis. 1993; 77:239–242.
- 22. Seifers DL, Harvey TL, Kofoid KD, Stegmeier WD. Natural infection of pearl millet and sorghum by Wheat streak mosaic virus in Kansas. Plant Dis. 1996; 80:179–185.
- 23. Sánchez-Sánchez H, Henry M, Cárdenas-Soriano E, Alvizo-Villasana HF. Identification of Wheat streak mosaic virus and its vector Aceria tosichella in Mexico. Plant Dis. 2001; 85:13–17.
- 24. Rotenberg D, Bockus WW, Whitfield AE, Hervey K, Baker KD, Ou Z, et al. Occurrence of viruses and associated grain yields of paired symptomatic and nonsymptomatic tillers in Kansas winter wheat fields. Phytopathol. 2016; 106:202–210.
- 25.
Somsen HW, Sill WH. The wheat curl mite, Aceria tulipae Keifer, in relation to epidemiology and control of wheat streak mosaic. Agricultural Experiment Station, Kansas State University of Agriculture and Applied Science. 1970; 162:1–24.
- 26. Harvey TL, Seifers DL, Martin TJ. Host range differences between two strains of wheat curl mites (Acari: Eriophyidae). J Agric Urban Entomol. 2001; 18:35–41.
- 27. Skoracka A, Rector BG, Hein GL. The interface between wheat and the wheat curl mite, Aceria tosichella, the primary vector of globally important viral diseases. Frontiers in Plant Science 2018; 9 (1098), 1–8, https://doi.org/10.3389/fpls.2018.01098).
- 28. Skoracka A, Lopes LF, Alves MJ, Miller A, Lewandowski M, Szydło W, et al. Genetics of lineage diversification and the evolution of host usage in the economically important wheat curl mite, Aceria tosichella Keifer, 1969. BMC Evol Biol. 2018; 18:122, https://doi.org/10.1186/s12862-018-1234-x.
- 29. Carew M, Schiffer M, Umina P, Weeks A, Hoffmann A. Molecular markers indicate that the wheat curl mite, Aceria tosichella Keifer, may represent a species complex in Australia. Bull Entomol Res. 2009; 99:479–486.
- 30. Skoracka A, Kuczyńsk L, de Mendonca RS, Dabert M, Szydło W, Knihinicki D, et al. Cryptic species within the wheat curl mite Aceria tosichella (Keifer) (Acari, Eriophyoidea) revealed by mitochondrial, nuclear and morphometric data. Invert Syst. 2012; 26:417–433.
- 31. Skoracka A, Kuczyński L, Szydło W, Rector B. The wheat curl mite Aceria tosichella (Acari: Eriophyoidea) is a complex of cryptic lineages with divergent host ranges: evidence from molecular and plant bioassay data. Biol J Linn Soc. 2013; 109:165–180.
- 32. Martin TJ, Harvey TL, Bender CG, Seifers DL, Hatchet JH. Control of Wheat streak mosaic virus with vector resistance in wheat. Phytopathol. 1984; 74:963–964.
- 33. Velandia M, Rejesus RM, Jones DC, Price JA, Workneh F, Rush CM. Economic impact of Wheat streak mosaic virus in the Texas High Plains. Crop Prot. 2010; 29:699–703.
- 34. Chuang W-P, Aguirre-Rojas LM, Khalaf LK, Zhang G, Fritz AK, Whitfield AE, et al. Wheat genotypes with combined resistance to wheat curl mite, wheat streak mosaic virus, wheat mosaic virus, and triticum mosaic virus. J Econ Entomol. 2017; 110:711–718.
- 35.
Malik R. Molecular genetic characterization of wheat curl mite, Aceria tosichella Keifer (Acari: Eriophyidae), and wheat genes conferring wheat curl mite resistance. PhD Dissertation, Kansas State University. 2001.
- 36. Hein GL, French R, Siriwetwiwat B, Amrine JW. Genetic characterization of North American populations of the wheat curl mite and dry bulb mite. J Econ Entomol. 2012; 105:1801–1808.
- 37. Schiffer M, Umina P, Carew M, Hoffmann A, Rodoni B, Miller A. The distribution of wheat curl mite (Aceria tosichella) lineages in Australia and their potential to transmit Wheat streak mosaic virus. Ann Appl Biol. 2009; 155:371–379.
- 38. Seifers DL, Harvey TL, Louie R, Gordon DT, Martin TJ. Differential transmission of isolates of the High Plains virus by different sources of wheat curl mites. Plant Dis. 2002; 86:138–142.
- 39. McMechan AJ, Tatineni S, French R, Hein GL. 2014. Differential transmission of Triticum mosaic virus by wheat curl mite populations collected in the Great Plains. Plant Dis. 2014; 98:806–810.
- 40. Andrews JE, Slykhuis JT. Reaction of winter wheat varieties and Triticum× Agropyron hybrids when inoculated with Wheat streak mosaic virus by the mite vector Aceria tulipae Keifer. Plant Dis Rep. 1956; 40: 513–516.
- 41. Carrera SG, Davis H, Aguirre–Rojas LM, Murugan M, Smith CM. Multiple categories of resistance to wheat curl mite (Acari: Eriophyidae) expressed in accessions of Aegilops tauschii. J Econ Entomol. 2012; 105:2180–2186.
- 42. Conner RL, Thomas JB, Whelan EDP. Comparison of mite resistance for control of wheat streak mosaic. Crop Sci. 1991; 31:315–318.
- 43. Thomas JB, Conner RL, Graf RJ. Comparison of different sources of vector resistance for controlling wheat streak mosaic in winter wheat. Crop Sci. 2004; 44:125–130.
- 44. Harvey TL, Martin TJ, Seifers DL. Importance of plant resistance to insect and mite vectors in controlling virus diseases of plants: resistance to the wheat curl mite (Acari: Eriophyidae). J Agric Entomol. 1994; 11:271–277.
- 45. Wood EA, Sebesta EE, Webster JA, Porter DR. Resistance to wheat curl mite (Acari: Eriophyidae) in greenbug-resistant ‘Gaucho’triticale and ‘Gaucho’x wheat crosses. J Econ Entomol. 1995; 88:1032–1036.
- 46. Harvey TL, Martin TJ, Seifers DL, Sloderbeck PE. Change in virulence of wheat curl mite detected on TAM 107 wheat. Crop Sci. 1997; 37:624–625.
- 47. Khalaf LK, Chuang W-P, Aguirre-Rojas LM, Klein P, Smith CM. Differences in Aceria tosichella population responses to wheat resistance genes and wheat virus transmission. Arthropod-Plant Inte. 2019; 13:807–818.
- 48. Skoracka A, Rector B, Kuczyńsk L, Szydło W, Hein G, French R. Global spread of wheat curl mite by its most polyphagous and pestiferous lineages. Ann Appl Biol. 2014; 116:222–235.
- 49. Skoracka A, Lewandowski M, Rector BG, Szydło W, Kuczyński L. Spatial and host-related variation in prevalence and population density of wheat curl mite (Aceria tosichella) cryptic genotypes in agricultural landscapes. PLoS One. 2017; 12:e0169874.
- 50.
Siriwetwiwat, B. 2006. Interactions between the wheat curl mite, Aceria tosichella Keifer (Eriophyidae), and wheat streak mosaic virus, and distribution of wheat curl mite biotypes in the field. Ph.D. Diss., University of Nebraska, Lincoln.
- 51. Harrison S. Local extinction in a metapopulation context: an empirical evaluation. Biological J Linn Soc. 1991; 42: 73–88.
- 52. Szydło W, Hein G, Denizhan E, Skoracka A. Exceptionally high levels of genetic diversity in wheat curl mite (Acari: Eriophyidae) populations from Turkey. J Econ Entomol. 2015; 108:2030–2039.
- 53. Tsutsui ND, Suarez AV, Holway DA, Case TJ. Reduced genetic variation and the success of an invasive species. Proc Natl Acad Sci. USA. 2000; 97:5948–5953.
- 54. De Barro P, Ahmed M.Z. Genetic networking of the Bemisia tabaci cryptic species complex reveals pattern of biological invasions. PLoS One. 2011; 6: p.e25579.
- 55.
Smith CM. Plant resistance to arthropods: molecular and conventional approaches. Springer, Berlin, Germany. 2005.
- 56. Martin TJ, Harvey TL, Bender CG, Seifers DL, Hatchett JH. Wheat curl mite resistant wheat germplasm. Crop Sci. 1983; 23:809.
- 57. Hofer P, Berg JE, Huang L, Graf RJ, Bruckner PL. Registration of ‘Yellowstone’ winter wheat backcross–derived lines incorporating leaf rust and wheat curl mite resistance. J Plant Regis. 2011; 5:422–425.
- 58. Carver BF, Smith CM, Chuang W-P, Hunger RM, Edwards JT, Yan L, et al. Registration of OK05312, a high-yielding hard winter wheat donor of for wheat curl mite resistance. J Plant Regis. 2016; 10:75–79.
- 59. Aguirre-Rojas LM, Khalaf LK, Garcés-Carrera S, Sinha DK, Chuang W-P, Smith CM. Resistance to wheat curl mite in arthropod-resistant rye-wheat translocation lines. Agronomy. 2017; 7:1–9.
- 60. Enders L, Hefley T, Girvin J, Whitworth R, Smith CM. Spatio-temporal distribution and environmental drivers of Barley yellow dwarf virus and vector abundance in Kansas. Phytopathol. 2018. (ja). https://doi.org/10.1094/PHYTO-10-17-0340-R
- 61. Wosula E, McMechan A, Oliveira-Hofman C, Wegulo S, Hein G. Differential transmission of two isolates of wheat streak mosaic virus by five wheat curl mite populations. Plant Dis. 2016; 100:154–158.
- 62. Untergasser A, Nijveen H, Rao X, Bisseling T, Geurts R, Leunissen JA. Primer3Plus, an enhanced web interface to Primer3. J Nucleic Acids Res. 2007; 35:W71–W74.
- 63. Hall TA. BioEdit: A user-friendly biological sequence alignment editor and analysis program for Windows 95/98/ NT. Nucleic Acids Symp Ser. 1999; 41:95–98.
- 64. Ronquist F, Teslenko M, Van Der Mark P, Ayres DL, Darling A, Höhna S, et al. MrBayes 3.2: efficient Bayesian phylogenetic inference and model choice across a large model space. Syst Biol. 2012; 61:539–542.
- 65. Librado P, Rozas J. DnaSP v5: a software for comprehensive analysis of DNA polymorphism data. Bioinformatics. 2009; 25:1451–1452.
- 66. Leigh JW, Bryant D. POPART: full-feature software for haplotype network construction. Methods Ecol Evol. 2015; 6:1110–1116.
- 67.
Wood SN. Generalized Additive Models: An introduction with R, Second Edition. CRC Press. 2017.
- 68.
PRISM Climate Group, Oregon State University, http://prism.oregonstate.edu, accessed 27 Feb 2017.
- 69. Homer CG, Dewitz JA, Yang L, Jin S, Danielson P, Xian G. Completion of the 2011 National Land Cover Database for the conterminous United States-Representing a decade of land cover change information. Photogram Engin Remote Sens. 2015; 81:345–354.
- 70. Hefley TJ, Broms KM, Brost BM, Buderman FE, Kay SL, Scharf HR, et al. The basis function approach for modeling autocorrelation in ecological data. Ecol. 2017; 98:632–646.
- 71.
Burnham KP, Anderson DR. Model selection and multimodel inference: A practical information-theoretic approach. Springer Verlag. 2003.
- 72. Simon C, Frati F, Beckenbach A, Crespi B, Liu H, Flook P. Evolution, weighing, and phylogenetic utility of mitochondrial gene sequences and a compilation of conserved polymerase chain reaction primers. Ann Entomol Soc Am. 1994; 87:651–701.
- 73. Skoracka A, Dabert M. The cereal rust mite Abacarus hystrix (Acari: Eriophyoidea) is a complex of species: evidence from mitochondrial and nuclear DNA sequences. Bull Entomol Res. 2010; 100:263–272.