INTRODUCTION
Silphium integrifolium Michx (commonly known as rosinweed or
silflower) is a native perennial prairie species being domesticated for
production as a forage and oilseed species (Vilela et al., 2018; Van
Tassel et al., 2017). The plant’s perennial nature allows it to provide
novel ecosystem services in agronomic production systems (Glover et al.,
2010; Van Tassel et al., 2017) while producing oilseeds with similar
nutritional composition to annual sunflower (Helianthus annuus L.
(Asterales: Asteraceae) seeds (Reinert et al., 2018). However, S.
integrifolium suffers infections from diverse pathogen enemies across
its native range. The stems, leaves, and flowers can be infected with
multiple strains (Turner, unpublished observations) of Puccinia
silphii Schwein (1832), or Silphium rust (Turner et al., 2018),Colletotrichum silphii Davis (1919) (Horst 2008) syn. andColletotrichum dematium (Pers.) Grove (Cybernome; Farr 1989)
called leaf blotch hereafter. Recently, Silphium clear vein (SCV)
leaf and stem disease has been described and hypothesized to be caused
by a virus based on the identification of viral sequences inSilphium integrifolium that are similar to the Dahlia common
Mosaic Virus (DCMV) and Dahlia endogenous plant pararetroviral sequence
(DvEPRS, formerly DMV-D10) (Almeyda, 2014; Cassetta et al., 2023).
Understanding how S. integrifolium defends against these
pathogens is crucial for progress in breeding programs and will
elucidate evolutionary questions posed in the species (Cassetta et al.,
2023, Van Tassel et al, 2017). In this study, plants from 12 wild
populations of S. integrifolium were planted in a reciprocal
transplant design to better understand adaptation to disease pressures
along a rainfall gradient. Phenotypic data revealed that these wild
populations have distinctive morphologies and showed variable levels of
tolerance to biotic stresses, such as pests and pathogens (Peterson et
al., in prep). Individuals with resistance to diseases of S.
integrifolium were visually identified in the study, and have been
previously identified in breeding populations (Turner et al., 2018), but
the genetic basis of disease resistance has not been investigated. Here,
we look at the diversity of disease resistance genes in wild populations
of S. integrifolium to answer evolutionary questions regarding
adaptation of wild populations to their pathogen pressures and suggest
applications for germplasm enhancement.
Environments with high precipitation are conducive to increased pathogen
pressures (Clarkson et al., 2014; Granke and Hausbeck 2010; Islam and
Toyota, 2004; Magarey et al., 2005; Rowlandson et al., 2015). Several
studies have found patterns of resistance alleles consistent with
pathogen selection intensity increasing with precipitation (Wahl 1970,
Abbott, Brown, and Burdon 1991, Dong et al. 2009). We hypothesize thatS. integrifolium populations evolved in areas with high
precipitation will have a more diverse complement of disease resistance
genes to deal with this challenge. To test this hypothesis, we collected
seeds from S. integrifolium stands from four prairie remnants
each in three geographic regions in the Central Plains of the United
States. The regions are situated along a gradient of effective
precipitation, or the amount of rainfall that is not lost to potential
evapotranspiration, referred to in this study as the climate moisture
index (CMI). We refer to these regions as “West”, from West-Central
Kansas, “Central”, from Eastern Kansas, and “East”, from Central
Illinois (Figure 1). Prairie remnants in the West receive around -55
kg/m2/month of CMI per year, whereas sites in the East
receive closer to -15 kg/m2/month (Table 1). We note
that plant defensive gene diversity increased along this gradient inAndropogon gerardii , a native perennial grass that co-occurs withS. integrifolium (Rouse et al. 2011) and soil pathogen diversity
also increased along this same gradient (Delavaux et al. 2021).
A major class of disease resistance genes in plants are the cytoplasmic
nucleotide binding and leucine-rich repeat (NLR) genes. (Caplan et al.,
2008; Eitas and Dangl, 2010). These NLRs, (hereafter, R genes), often
contain an N-terminal domain that is either of a coiled-coil type,
called CNLs, or a toll-interleukin-like type (TNLs), whose divergence
predates the split between monocots and eudicots (Meyers et al., 1999),
estimated to be about 160 Mya by TimeTree (Kumar et al., 2017). R genes
are highly diverse, numbering in the hundreds in many plant species
(Jupe et al., 2012; Toda et al., 2020). These genes evolve rapidly and
are enriched in presence/absence variation even at the intraspecies
level, limiting the utility of a single reference genome to elucidate
the diversity of R genes within a species. This study employed R-gene
enrichment sequencing (RenSeq; Jupe et al, 2013), which uses DNA baits
to reduce complexity of genomic DNA libraries. The baits we designed
come from the common sunflower, Helianthus annuus L., which is
separated from S. integrifolium by between 22 (Meireles et al.,
2020) and 33 million years ago (Zhang et al., 2021) . We sequenced the
RenSeq libraries from S. integrifolium plants across the rainfall
gradient using both Illumina short-read and PacBio long-read
technologies.
RenSeq is not a new technology – having been developed in 2013. It has
been used to address basic problems in plant immunity (Jupe et al.,
2013) and applied questions in plant breeding in numerous crop species
and their wild relatives for crop improvement (e.g. Arora et al., 2019)
and to understand mechanisms of plant immunity (e.g. Witek et al.,
2021). Despite the utility of RenSeq in plant breeding projects, and in
helping us better understand plant disease resistance, this technology
has not been used to address basic evolutionary and biogeographical
questions.
Using both simple linear models as well as mixed effect models, we found
a significant positive correlation between the effective precipitation
of a plant’s host prairie and the number of R genes detected in both the
Illumina and PacBio datasets. While this observation does not itself
prove the broader theory that disease resistance correlates with
rainfall in plants, it is certainly consistent with theoretical
expectations, and should motivate additional evaluation of these ideas
in other taxa. Because we demonstrate the economic utility of the RenSeq
approach in non-model systems, our work not only points to data
consistent with our hypothesis but paves the way towards evaluating the
generality of this finding.