Management Implications
Invasive knotweeds (Reynoutria spp.) pose a threat to ecosystems across Europe and North America. There are numerous genetic entities in the invaded range, including two species (Reynoutria japonica [Japanese knotweed] and Reynoutria sachalinensis [giant knotweed]) and several distinct kinds of hybrids (collectively known as Bohemian knotweed [Reynoutria ×bohemica]). The different genetic entities vary with respect to attributes that affect their invasiveness, including reproductive strategies, growth rate, and competitive ability. It benefits land managers to know the genetic identities of their invasive Reynoutria plants so that they can develop effective strategies for prevention and control. In this study, we expand on previous work and provide additional data regarding the genetic makeup of invasive Reynoutria. We present novel methods for identifying plant material, including a next-generation sequencing method and a cost-effective method that uses polymerase chain reaction (PCR). The newly developed methods should enable rapid and confident identification of invasive Reynoutria plants.
Introduction
Plant invasions frequently manifest as two or more congeneric species with similar ecological profiles (Gaskin et al. Reference Gaskin, West and Rector2024; Green Reference Green1966; Kartesz Reference Kartesz2015; Lubell et al. Reference Lubell, Brand, Lehrer and Holsinger2008; Maddox et al. Reference Maddox, Byrd and Serviss2010). When two invasive congeners grow together, these can hybridize to produce a new invasive taxon (Ellstrand and Schierenbeck Reference Ellstrand and Schierenbeck2000; Schierenbeck and Ellstrand Reference Schierenbeck and Ellstrand2009; Welles and Ellstrand Reference Welles and Ellstrand2020). It is essential to identify species and hybrids correctly, because the different genetic entities could have different ecological tolerances or responses to control methods (David et al. Reference David, Zarnetske, Hacker, Ruggiero, Biel and Seabloom2015; Tataridas et al. Reference Tataridas, Jabran, Kanatas, Oliveira, Freitas and Travlos2022; Yang et al. Reference Yang, Ferrari and Shea2011). Moreover, the legal status of one taxonomic entity may differ from another, and this discrepancy can delay an effective management response in cases of uncertain identifications (Fox and Gordon Reference Fox and Gordon2009; Randall et al. Reference Randall, Morse, Benton, Hiebert, Lu and Killeffer2008).
Invasive knotweeds in the genus Reynoutria Houtt. (syn.: Polygonum L.; Polygonaceae) are a growing concern in several parts of the world. Originally restricted to a relatively narrow native distribution in eastern Asia, these plants have become noxious weeds in Europe (Lavoie Reference Lavoie2017), North America (Del Tredici Reference Del Tredici2017), Oceania (Desjardins et al. Reference Desjardins, Pashley and Bailey2023b), and Africa (Germishuizen Reference Germishuizen1986). Japanese knotweed [Reynoutria japonica Houtt.; syn.: Polygonum cuspidatum Siebold & Zucc.]), giant knotweed [Reynoutria sachalinensis (F. Schmidt ex Maxim.) Nakai; syn.: Polygonum sachalinense F. Schmidt ex Maxim.]), and their hybrid known as Bohemian knotweed [Reynoutria ×bohemica Chrtek & Chrtková; syn.: Polygonum ×bohemicum (J. Chrtek & Chrtková) Zika & Jacobson [cuspidatum × sachalinense]]) have similar invasive potential in temperate habitats, owing to their aggressive clonal growth (Bailey et al. Reference Bailey, Bímová and Mandák2009), their allelopathic effects on neighboring plants (Murrell et al. Reference Murrell, Gerber, Krebs, Parepa, Schaffner and Bossdorf2011; Parepa et al. Reference Parepa, Schaffner and Bossdorf2012; Vrchotová and Šerá Reference Vrchotová and Šerá2008), and their tenacious underground rhizomes that resist eradication (Drazan Reference Drazan2022; Lawson et al. Reference Lawson, Fennell, Smith and Bacon2021). The negative impacts of invasive Reynoutria species include ecosystem disruption (Drazan et al. Reference Drazan, Smith, Anderson, Becker and Clark2021; Gerber et al. Reference Gerber, Krebs, Murrell, Moretti, Rocklin and Schaffner2008; Lecerf et al. Reference Lecerf, Patfield, Boiché, Riipinen, Chauvet and Dobson2007; Maerz et al. Reference Maerz, Blossey and Nuzzo2005; Serniak et al. Reference Serniak, Corbin, Pitt and Rier2017; Siemens and Blossey Reference Siemens and Blossey2007) and a variety of negative consequences for humans (Mclean Reference Mclean2010).
Although these Reynoutria species and hybrids are all considered invasive outside their native range, differences in their ecological attributes may impact their mechanism of spread, habitat tolerances, competitive abilities, and susceptibility to control methods. For example, the hybrid R. ×bohemica has been shown to grow more rapidly and inhibit the growth of native species more strongly than either R. japonica or R. sachalinensis (Bímová et al. Reference Bímová, Mandák and Pyšek2003; Mandák et al. Reference Mandák, Pyšek and Bímová2004; Moravcová et al. Reference Moravcová, Pyšek, Jarošik and Zákravský2011; Parepa et al. Reference Parepa, Fischer, Krebs and Bossdorf2014). Control methods often are applied indiscriminately to invasive Reynoutria regardless of taxon, but there can be some value to using taxon-specific management approaches (Camargo et al. Reference Camargo, Kurose, Post and Lommen2022; Clements et al. Reference Clements, Larsen and Grenz2016; Kadlecová et al. Reference Kadlecová, Vojík, Kutlvašr and Berchová-Bímová2022; Yoshimoto and Szűcs Reference Yoshimoto and Szűcs2024). Therefore, it would be valuable to identify which taxon (or taxa) is present before attempting to manage it (Bailey et al. Reference Bailey, Bímová and Mandák2007). Although morphological characters exist for differentiating the various invasive Reynoutria taxa (Zika and Jacobson Reference Zika and Jacobson2003), molecular methods offer an independent method for improving confidence and confirming the identification of morphologically intermediate plants (Tippery et al. Reference Tippery, Olson and Wendtlandt2021). Fortunately, molecular methods can identify invasive Reynoutria, including hybrids, using DNA sequence data from individual genes (Park et al. Reference Park, Bhandari, Won, Park and Park2018; Tippery et al. Reference Tippery, Olson and Wendtlandt2021) or from a wide range of variable sites across the genome (Drazan Reference Drazan2022).
LEAFY (LFY or LFY3, also known as FLORICAULA/FLO; Arabidopsis AT5G61850; Berardini et al. Reference Berardini, Reiser, Li, Mezheritsky, Muller, Strait and Huala2015) is a low-copy nuclear gene that encodes a transcriptional regulator involved in floral meristem development (Frohlich and Meyerowitz Reference Frohlich and Meyerowitz1997; Gao et al. Reference Gao, Chen, Li and Zhang2019; Moyroud et al. Reference Moyroud, Tichtinsky and Parcy2009). Nuclear gene sequences can provide biparentally inherited information for reconstructing phylogenetic relationships, and more importantly, such genes can be useful for distinguishing parental species and their allopolyploid descendants. LEAFY has been used to infer the parental species involved in generating allopolyploid species in Fagopyrum Mill. and Persicaria Mill. (Kim et al. Reference Kim, Sultan and Donoghue2008; Nishimoto et al. Reference Nishimoto, Ohnishi and Hasegawa2003) as well as Polygonaceae more broadly (Sanchez and Kron Reference Sanchez and Kron2008). In recent years, the utility of LEAFY has expanded substantially to include plants from across the angiosperm phylogeny (e.g., Archambault and Bruneau Reference Archambault and Bruneau2004; Grob et al. Reference Grob, Gravendeel and Eurlings2004; Howarth and Baum Reference Howarth and Baum2005).
The LEAFY gene structure is conserved across angiosperms and consists of three exons and two introns (Weigel et al. Reference Weigel, Alvarez, Smyth, Yanofsky and Meyerowitz1992). Phylogenetic studies have relied upon the second intron, which can be reliably sequenced and is sufficiently large and variable for many phylogenetic applications. Prior studies of Reynoutria have produced more than 20 unique sequences for the LEAFY second intron and provided support for the phylogenetic distinctness of several species: R. japonica and R. sachalinensis, as well as Reynoutria compacta [dwarf knotweed] (Hook.f.) Nakai and Reynoutria multiflora [tuber fleeceflower] (Thunb.) Moldenke (Desjardins et al. Reference Desjardins, Bailey, Zhang, Zhao and Schwarzacher2023a; Park et al. Reference Park, Bhandari, Won, Park and Park2018; Tippery et al. Reference Tippery, Olson and Wendtlandt2021). Analysis of LEAFY second intron sequences for invasive Reynoutria in the United States recovered four kinds of sequence that were attributed to R. japonica and two sequences that were attributed to R. sachalinensis (Tippery et al. Reference Tippery, Olson and Wendtlandt2021). The same publication describes how five different LEAFY sequence combinations were recovered from plants identified as R. ×bohemica, suggesting a high amount of interbreeding among species and hybrids.
The existing LEAFY second intron sequences for Reynoutria species have been effective for identifying invasive species and hybrids, and we anticipate that gathering additional sequence data would improve the utility of this region for species identification and phylogenetic analysis. For example, a study in the related genus Fagopyrum recovered nearly complete LEAFY sequences (including substantial portions of all introns and exons), and these provided some of the first data for LEAFY gene structure outside the model organism Arabidopsis thaliana (Nishimoto et al. Reference Nishimoto, Ohnishi and Hasegawa2003). We endeavored to obtain sequences from the first intron and second exon to learn about the protein-coding portion of the LEAFY gene and to compare the amount of variation between the two intron regions. We anticipate that the enhanced dataset will enable more precise and more efficient identifications of Reynoutria species and hybrids by providing a larger set of nucleotide polymorphisms that can be compared.
Materials and Methods
Morphological Data Collection
Samples of Reynoutria species were collected from Wisconsin, USA (Figure 1; Supplementary Table S1) and were targeted to include all genetic variants that had been documented previously (Tippery et al. Reference Tippery, Olson and Wendtlandt2021). Morphological data for lamina base, abaxial vein hairs, inflorescence size, and reproductive condition were recorded from dried specimens, as described previously (Tippery et al. Reference Tippery, Olson and Wendtlandt2021). Additionally, we measured the length of the longest available lamina for each specimen (i.e., located in the distal 20 cm of a branch) as the length from the petiole attachment point to the tip of the lamina apex. Characters were scored numerically as follows: lamina base (0 = truncate without apparent lobes, 0.5 = intermediate with lobes <2 cm, 1 = with lobes >2 cm), abaxial vein hairs (0 = hairs absent or evident only as scabrous protrusions, 0.5 = hairs unicellular and <2-mm long, 1 = hairs multicellular and >2-mm long), inflorescence size (0 = shorter than leaf at the same node, 0.5 = approximately equal, 1 = longer than leaf at the same node), and reproductive condition (0 = male sterile, 1 = male fertile). Lamina length was recorded as a continuous numerical variable.
Figure 1. Map of Reynoutria collection sites in Wisconsin. Shapes indicate taxonomic identity (square, R. japonica; circle, R. ×bohemica; triangle, R. sachalinensis), and colors depict unique composite genotypes. Numbers inside the shapes refer to specimen ID (Supplementary Table S1). Site positions are “jittered” to facilitate viewing adjacent sites, using a random uniform distribution of 0.1° in both latitude and longitude.
Molecular Data Collection
Primers for polymerase chain reaction (PCR) were developed initially by aligning and comparing sequences of the LEAFY gene from a phylogenetic study of Fagopyrum (Nishimoto et al. Reference Nishimoto, Ohnishi and Hasegawa2003). Subsequent primers were designed using novel sequence data from Reynoutria taxa. Ultimately, the following primers were used most effectively: Flint1-F1, Japo8F, and Reyn3F, located in the first exon; Japo5R and Reyn4F, both located in the second exon; and MLFYI2-2385R (Schuster et al. Reference Schuster, Wilson and Kron2011), located in the third exon (Table 1). The plastid matK region was amplified using the primers AF and 8R (Yan et al. Reference Yan, Pang, Jiao, Zhao, Shen and Zhao2008).
Table 1. DNA sequences for oligonucleotide primers that were used to amplify portions of the LEAFY gene in Reynoutria species.
a Intron 1 is located at alignment positions 299–834, and intron 2 is located at positions 1223–2375.
Genomic DNA was extracted using the CTAB method (Doyle and Doyle Reference Doyle and Doyle1987), modified as described by Tippery et al. (Reference Tippery, Pesch, Murphy and Bautzmann2020). PCR was conducted using the Phire Hot Start II DNA Polymerase (Thermo Fisher Scientific, Waltham, MA, USA), with a 58 C annealing temperature and 30-s extension time. PCR products were cleaned enzymatically using the ExoI and FastAP enzymes (Thermo Fisher Scientific). Sanger sequencing (Sanger et al. Reference Sanger, Nicklen and Coulson1977) used the same primers that were used for PCR reactions and was conducted through Eurofins Genomics (Louisville, KY, USA). In situations where Sanger sequencing produced polymorphic results, separate sequences were obtained either by subcloning PCR products into bacterial vectors or by conducting next-generation sequencing. For bacterial subcloning, the PCR products were cleaned using the Wizard® SV Gel and PCR Clean-Up System (Promega, Madison, WI, USA) and cloned into bacteria using the CloneJET PCR Cloning Kit (Thermo Fisher Scientific), then colonies of transformed cells were subjected to PCR and sequencing as above. For next-generation sequencing, PCR products were cleaned using 0.1 volumes of 3 M ammonium acetate and 4 volumes of 100% ethanol, followed by one wash with 80% ethanol. Sequencing was conducted by Eurofins Genomics using Oxford Nanopore Technologies (Oxford, UK).
Molecular Data Analysis
Newly obtained sequences were combined with previously published sequences (Desjardins et al. Reference Desjardins, Bailey, Zhang, Zhao and Schwarzacher2023a; Park et al. Reference Park, Bhandari, Won, Park and Park2018; Tippery et al. Reference Tippery, Olson and Wendtlandt2021) and aligned manually using Mesquite v. 3.81 (Maddison and Maddison Reference Maddison and Maddison2023). LEAFY sequence variants were attributed to species according to their prior identification (Desjardins et al. Reference Desjardins, Bailey, Zhang, Zhao and Schwarzacher2023a; Park et al. Reference Park, Bhandari, Won, Park and Park2018) or their phylogenetic relatedness to previously identified sequences (see “Results and Discussion”). Sequence alignments were analyzed using the ape package v. 5.7.1 (Paradis and Schliep Reference Paradis and Schliep2009) for R v. 4.4.1 (R Core Team 2024). After model selection using IQ-TREE v. 2.0.5 (Minh et al. Reference Minh, Schmidt, Chernomor, Schrempf, Woodhams, von Haeseler and Lanfear2020; Nguyen et al. Reference Nguyen, Schmidt, von Haeseler and Minh2015), phylogenetic analyses were conducted in BEAST v. 1.10.4 (Suchard et al. Reference Suchard, Lemey, Baele, Ayres, Drummond and Rambaut2018) with the GTR+G model of evolution (Tavaré Reference Tavaré1986), using 10 million generations of Markov chain Monte Carlo (MCMC) (Hastings Reference Hastings1970; Metropolis et al. Reference Metropolis, Rosenbluth, Rosenbluth, Teller and Teller1953), sampled every 5,000 generations, and trees were summarized after the first 25% of trees were discarded as burn-in. Sequence alignment data for the two LEAFY introns were evaluated separately and together.
To evaluate distances among novel and previously published sequences effectively, the sequence alignment was trimmed to include only regions having complete data for all nucleotide sequences (i.e., only the LEAFY second intron). Insertions/deletions (indels) were coded using simple indel coding (Simmons and Ochoterena Reference Simmons and Ochoterena2000) in the program SeqState v. 1.4.1 (Müller Reference Müller2005). After removal of identical sequences from the data matrix, genetic distances among LEAFY gene sequences were visualized using the program SplitsTree v. 6.3.25 (Huson and Bryant Reference Huson and Bryant2006) with p distances (Hamming Reference Hamming1950) and NeighborNet method (Bryant and Huson Reference Bryant and Huson2023; Bryant and Moulton Reference Bryant and Moulton2004).
Putative amino acid translations of the LEAFY exons were generated in Mesquite after aligning Reynoutria and Fagopyrum sequences and coding equivalent nucleotide positions as codons (Nishimoto et al. Reference Nishimoto, Ohnishi and Hasegawa2003). Similarity comparisons were made by conducting a protein BLAST search of the putative translations (Altschul et al. Reference Altschul, Gish, Miller, Myers and Lipman1990; Johnson et al. Reference Johnson, Zaretskaya, Raytselis, Merezhuk, McGinnis and Madden2008).
Diagnostic PCR
Primers for diagnostic PCR reactions were developed by identifying regions that consistently differed among types of Reynoutria sequences. Four novel primers were developed and named according to whether they amplified R. japonica (Japo) or R. sachalinensis (Sach) sequence variants. Japo1F or Sach1F was used in combination with MLFYI2-2385R. Japo1R or Sach1R was used with Reyn3F. All diagnostic primers were designed to anneal to locations in the LEAFY second intron. Diagnostic PCR reactions were run as described earlier for other primers, and amplicons were run using electrophoresis on a 2% agarose gel containing GelStar™ nucleic acid stain (Lonza Bioscience, Walkersville, MD, USA). PCR products (4 µl of each) were run alongside 2 µl of GeneRuler 1 kb Plus DNA ladder (Thermo Fisher Scientific).
Morphological Analyses
Morphological data were evaluated for individuals with known LEAFY gene sequences. Data were visualized using the ggplot2 package v. 3.5.1 in R (R Core Team 2024; Wickham Reference Wickham2016). Plants were grouped by their composite genetic makeup (i.e., all types of LEAFY sequence variants that were recovered for each individual), and the differences among genetic variants were evaluated using ANOVA (Fisher Reference Fisher1921). Specifically, we used the aov function in the stats package, followed by the Tukey’s test (Tukey Reference Tukey1949) via the HSD.test function in the agricolae package v. 1.3.7 in R (Mendiburu Reference Mendiburu2021; R Core Team 2024). Principal component analysis (Pearson Reference Pearson1901) was done via the prcomp function in the R stats package (R Core Team 2024) using the variables of lamina length, lamina base, and abaxial vein hairs, after removing individuals that lacked data for one or more of the variables.
Results and Discussion
Molecular Data Summary
Sequence data from the LEAFY gene were obtained for 156 Reynoutria specimens (Figure 1; Supplementary Table S1). Five unique sequence variants were identified, and these were assigned codes that corresponded to previously identified variants (Tippery et al. Reference Tippery, Olson and Wendtlandt2021). (It remains unclear whether any of the LEAFY sequence variants are inherited as alleles for homologous chromosomes, and therefore we avoid referring to sequence variants as “alleles.” We refer to the collection of all sequence variants obtained for a single individual as its “composite genotype.”) Three sequence variants were attributable to R. japonica (J1, J2, J3), and two variants were attributed to R. sachalinensis (S1, S2) (Park et al. Reference Park, Bhandari, Won, Park and Park2018; Tippery et al. Reference Tippery, Olson and Wendtlandt2021).
Up to three variants were recovered from any one individual. Thirty-two accessions contained only R. japonica sequence variants (composite genotypes J1/J3 and J1/J2/J3), and 20 accessions contained only R. sachalinensis sequence variants, specifically the S2 variant. The remaining plants all contained a combination of sequence variants from both species, which is consistent with their identification as hybrids. There was minimal variation among sequences that were identified herein as corresponding to the same sequence variant (Supplementary Figures S1 and S2).
Sequence alignment for the LEAFY second intron was 1,149 nucleotides in length and included 26 polymorphic sites and 42 indels (Figure 2). The first intron was 483 to 528 nucleotides in length and included 19 polymorphic sites and 10 indels. The second exon was 402 to 408 nucleotides in length and included 5 polymorphic sites and 1 indel. The second intron was 599 to 1,056 nucleotides in length and included 27 polymorphic sites and 14 indels. The second intron of the J3 sequence variant included a 369-nucleotide insertion relative to the other variants (Figure 2). Sequence data for the first exon were incomplete but comprised 240 nucleotides, of which 3 sites were polymorphic, with no indels. Because of the location of the MLFYI2-2385R primer (Schuster et al. Reference Schuster, Wilson and Kron2011), we obtained only a negligible amount of sequence data for the third exon.
Figure 2. Alignment of LEAFY sequences for invasive Reynoutria taxa, showing the most commonly encountered sequences for each sequence variant. Sequences are identified as originating from R. japonica (J1/J2/J3) or R. sachalinensis (S1/S2). Intron and exon borders are indicated; the third exon is not shown but would appear shortly to the right of the last nucleotide shown.
Molecular Data Analyses
Phylogenetic analysis of LEAFY sequences recovered two clades that corresponded to R. japonica and R. sachalinensis, respectively (Figure 3; Supplementary Figures S1 and S2). Newly reported sequence variants frequently were most similar to previously reported sequences from invasive plants in New Zealand or the United Kingdom. For example, our J2 variant closely matched accession ON586877 from New Zealand, the J3 variant matched ON586876 from New Zealand, and both S1 and S2 were very similar to JF831231 from the United Kingdom. Notably, the J1 variant did not share close similarity with any previously published sequences from outside the United States. The phylogenetic position of R. compacta relative to R. japonica and R. sachalinensis was poorly supported and inconclusive.
Figure 3. Phylogenetic relationships among sequences obtained from Reynoutria species. Sequences are identified to species according to their previous identifications (Desjardins et al. Reference Desjardins, Bailey, Zhang, Zhao and Schwarzacher2023a; Park et al. Reference Park, Bhandari, Won, Park and Park2018; Schuster et al. Reference Schuster, Wilson and Kron2011) or by their similarity to previously published sequences. Previously published sequences are labeled with GenBank accession numbers. Sequences from plants in Wisconsin (Tippery et al. Reference Tippery, Olson and Wendtlandt2021; this study) have a unique identifier (e.g., R019) that is referenced in Supplementary Table S1. Sequence variant names (e.g., J1) are given for plants collected from the invasive range. Nodal support values indicate posterior probability; values less than 0.5 are not shown.
Translation of the protein-coding regions of the LEAFY gene produced amino acid sequences that were comparable with equivalent regions in Fagopyrum. Protein BLAST of the translated coding region from the longest sequence among the novel accessions (R046: 213 amino acids) returned a maximum sequence identity of 80.1% with Fagopyrum macrocarpum [largeseed buckwheat] Ohsako & Ohnishi (GenBank accession no. BAC76911). Comparison of translated LEAFY gene sequences among Reynoutria taxa showed that two amino acid positions were consistently different between sequences identified as R. japonica or R. sachalinensis. Additionally, there was an indel, two amino acids in length, in a region that consisted of either four or six glutamine residues, in a region that also was variable among Fagopyrum sequences (Nishimoto et al. Reference Nishimoto, Ohnishi and Hasegawa2003).
Data from the plastid matK region placed most of the accessions into the clade corresponding to R. japonica, except for accessions that had only R. sachalinensis LEAFY sequences (with no evidence of hybridization), in which case the latter group had matK sequences that matched R. sachalinensis (data not shown). The fact that plastid sequences for hybrids match R. japonica is consistent with that species being the original maternal parent of all hybrids, and this also is consistent with the evidence that R. japonica is male-sterile throughout its invasive range (Forman and Kesseli Reference Forman and Kesseli2003; Grimsby et al. Reference Grimsby, Tsirelson, Gammon and Kesseli2007; Hollingsworth and Bailey Reference Hollingsworth and Bailey2000; Tippery et al. Reference Tippery, Olson and Wendtlandt2021).
The SplitsTree network for LEAFY sequences separated R. japonica sequence variants from those of R. sachalinensis (Figure 4). The shortest distance between sequence variants was between S1 and S2 of R. sachalinensis. The J3 variant of R. japonica showed the greatest dissimilarity among sequences assigned to the same variant. The LEAFY sequence for R. compacta was most similar to the J1 sequence variant of R. japonica, from which it differed by 13 changes. Sequence variants for R. japonica sequences were separated from R. sachalinensis sequences by a minimum of 12 sequence differences. The maximum distance between any 2 R. japonica sequence variants was 30 differences, and the maximum distance between R. sachalinensis sequence variants was 5 differences.
Figure 4. SplitsTree network of Reynoutria LEAFY sequences. Each label represents a unique sequence variant that is connected to similar sequences by lines. Previously published sequences are identified with GenBank accession numbers. Sequences from plants in Wisconsin (Tippery et al. Reference Tippery, Olson and Wendtlandt2021; this study) have a unique identifier (e.g., R019) that is referenced in Supplementary Table S1. Sequence variant names (e.g., J1) are given for plants collected from the invasive range.
Diagnostic PCR reactions corroborated the genotype evidence that was obtained using DNA sequencing (Figure 5). The expected lengths of PCR products were as follows: Japo1F/MLFYI2-2385R: 638 to 678 bp; Sach1F/MLFYI2-2385R: 548 to 549 bp; Reyn3F/Japo1R: 1,102 to 1,106 bp; Reyn3F/Sach1R: 1,055 to 1,091 bp. Individuals that had only R. japonica sequence variants failed to amplify products using the Sach1F or Sach1R primers, individuals with only R. sachalinensis sequence variants failed to produce products using Japo1F or Japo1R, and all diagnostic primer pairs produced PCR products for individuals whose genetic makeup included sequence variants from both species (Figure 5). A variety of hybrid genotypes were tested, including individuals for each of the single-species genotypes as well as multiple kinds of hybrids.
Figure 5. Results of diagnostic polymerase chain reaction (PCR) using primers that selectively amplify either Reynoutria japonica or Reynoutria sachalinensis sequence variants. The top row shows the result of two primer pairs, separated by a dotted line, both using the MLFYI2-2385R reverse primer located in the LEAFY third exon, with a discriminant forward primer located in the second intron. The bottom row shows the result of two primer pairs, separated by a dotted line, and both using the Reyn3F forward primer located in the LEAFY first exon, with a discriminant reverse primer located in the second intron. The same source DNAs were used for all four PCR reactions: (A) R012, (B) R039, (C) R026, and (D) R071 (Supplementary Table S1). R012 has only R. japonica sequence variants, R071 has only R. sachalinensis sequence variants, and the remaining two DNAs contain sequence variants from both species. Two sizes (in bp) are labeled for the size standards, corresponding to the brightest bands at those locations.
This study investigated the feasibility and diagnostic value of gathering additional sequence data from the LEAFY first intron and second exon, regions that have seldom been sequenced in angiosperms. We were able to obtain sequence data from these regions for all previously identified genetic variants in the United States, with the result that about twice as many nucleotides (relative to prior studies) could be evaluated to assess genetic variability and phylogenetic relationships. We developed novel PCR primers that could be used to sequence additional Reynoutria taxa as well as species in related genera. Currently, there are no comparable first intron or second intron sequences for plants in the native range of Reynoutria or other portions of the invaded range, and these would be valuable to compare.
Morphological Data Summary
We obtained morphological data for the same Reynoutria accessions that were used to extract molecular data (Figure 1; Supplementary Table S1). There were 60 specimens with complete morphological data for all characters evaluated, whereas the remaining specimens either lacked reproductive material or had lost all flowers to abscission. Inflorescence length data were recorded for 93 specimens. Lamina base morphology and abaxial vein hair morphology were recorded for all 156 specimens. Lamina length across all specimens (all species) ranged from 8 to 29 cm.
Morphological Data Analyses
Our extensive sampling of Reynoutria plants in Wisconsin has enabled us to evaluate the relationship between composite genotype and morphology. It has been established previously that hybrid R. ×bohemica plants have intermediate morphological characteristics (Tippery et al. Reference Tippery, Olson and Wendtlandt2021; Zika and Jacobson Reference Zika and Jacobson2003), but the morphologies of various hybrids have not been explored in light of their genetic differences. Genetically variable hybrids may have morphological traits that resemble those of one or the other parental species (Jordon-Thaden et al. Reference Jordon-Thaden, Spoelhof, Viccini, Combs, Gomez, Walker, Soltis and Soltis2023; Mitchell et al. Reference Mitchell, Luu, Owens, Rieseberg and Whitney2022; Rieseberg et al. Reference Rieseberg, Raymond, Rosenthal, Lai, Livingstone, Nakazato, Durphy, Schwarzbach, Donovan and Lexer2003). Thus, it is important to evaluate morphological differences among Reynoutria plants that have different genomic contributions from R. japonica and R. sachalinensis.
Plants in our study with only R. japonica sequence variants (J1/J2/J3 and J1/J3 composite genotypes) mostly exhibited the expected morphology for that species: truncate lamina base, absent abaxial vein hairs, and male-sterile flowers (Figure 6). Long abaxial vein hairs were observed in most (but not all) specimens having the S2 composite genotype and rarely in other composite genotypes, whereas intermediate vein hairs were found occasionally in several composite genotypes (Figure 6C). Inflorescence length was expected to be longer than leaf length for R. japonica, but in fact very few specimens of any genotype had long inflorescences (Figure 6D). Plants with only R. sachalinensis sequence variants (S2 composite genotype) mostly had the expected traits of long basal lamina lobes, multicellular abaxial vein hairs, and male-fertile flowers.
Figure 6. Morphological data for Reynoutria plants with known composite genotypes. In each panel, shapes correspond to taxonomic identity (square, R. japonica; circle, R. × bohemica; triangle, R. sachalinensis), and colors depict unique composite genotypes. Positions are “jittered” by a value of 0.2 to facilitate viewing adjacent points. The first five panels each show morphological data for one trait: (A) lamina length, (B) lamina base, (C) abaxial vein hairs, (D) relative inflorescence size, and (E) reproductive condition. The final panel (F) shows the principal component (PC) analysis, with vectors showing the relative contributions of the three morphological traits that were evaluated.
Among the hybrids, basal lamina lobes >2 cm were rarely encountered, except in the J2/S2 composite genotype (Figure 6B). Notably, this was the only hybrid composite genotype with an S2 sequence variant. Male-sterile plants were rarely encountered among hybrids (Figure 6E). Lamina length was widely variable overall (Figure 6A), yet R. sachalinensis plants and plants with the J2/S1 or J2/S2 composite genotypes had significantly longer laminae than R. japonica plants (one-way ANOVA: F(7, 143) = 22.37, P < 2−16). The J1/S1 and J2/S2 hybrids had a significantly different mean for their lamina base morphology (Figure 6B), but it should be noted that both truncate bases and lobes of intermediate size were present in both kinds of hybrids. The lamina base trait was evaluated using dried herbarium specimens, and it may be possible to gain a clearer understanding of the differences by using more precise measurements on fresh leaves.
Importantly, the morphological analysis revealed that some hybrid individuals with the J1/S1 or J2/S2 composite genotypes are indistinguishable from non-hybrid R. sachalinensis plants, and other individuals with these same composite genotypes were not different from R. japonica plants. Principal component (PC) analysis (Figure 6F) separated plants on the PC1 axis mostly by lamina base, and the other two morphological traits contributed strongly to the PC2 dimension. Plants largely clustered by composite genotype, and overall the two most commonly encountered hybrid genotypes (J1/S1 and J2/S2) were poorly differentiated from each other. Nonetheless, the hybrids largely could be distinguished from R. japonica by having longer laminae and from R. sachalinensis by lacking basal lamina lobes and multicellular abaxial vein hairs.
Widespread Hybridization
The high similarity between LEAFY sequences for R. japonica and R. sachalinensis (Figures 2–4) supports prior evidence that these species are closely related and occasionally hybridize in their native range (Park et al. Reference Park, Bhandari, Won, Park and Park2018; Tippery et al. Reference Tippery, Olson and Wendtlandt2021). Hybrid genotypes appear prominently in the invaded range, where they are associated with measurably different invasiveness patterns (Bímová et al. Reference Bímová, Mandák and Pyšek2003; Mandák et al. Reference Mandák, Pyšek and Bímová2004; Moravcová et al. Reference Moravcová, Pyšek, Jarošik and Zákravský2011). Our study corroborates prior evidence that the invasive U.S. Reynoutria hybrids comprise at least five distinct genotypic combinations, presumably the result of separate original hybridization events (Tippery et al. Reference Tippery, Olson and Wendtlandt2021).
The existence of multiple versions of the LEAFY gene in some Reynoutria individuals likely reflects polyploidy in the genus, which has been documented in the native and invasive ranges (Iwatsubo et al. Reference Iwatsubo, Kodate and Naruhashi2004; Kim and Park Reference Kim and Park2000; Mandák et al. Reference Mandák, Pyšek, Lysák, Suda, Krahulcova and Bímová2003). In Europe, different ploidy levels characterize each taxon, with octaploid R. japonica and tetraploid R. sachalinensis having hybridized to produce the predominantly hexaploid R. ×bohemica (Bailey et al. Reference Bailey, Bímová and Mandák2007; Te Beest et al. Reference Te Beest, Le Roux, Richardson, Brysting, Suda, Kubešová and Pyšek2012), although a variety of ploidy levels were observed for the latter taxon (Saad et al. Reference Saad, Tiébré, Hardy, Mahy and Vanderhoeven2011). We recovered no more than three sequence variants from any one individual, and the only taxon we encountered with one variant per individual was R. sachalinensis (containing the S2 sequence variant only). The genetically diverse array of hybrids has yet to be correlated with ploidy levels, and a chromosomal investigation may produce insights into the origins of hybrid genotypes. We maintain that using a single hybrid name, R. ×bohemica, may hinder a more nuanced understanding of hybridization in Reynoutria, and we support efforts to distinguish hybrids by their genetic composition rather than simply by their parent species (Tippery et al. Reference Tippery, Olson and Wendtlandt2021).
Four of the five LEAFY gene variants we identified were very similar to sequences that were recovered from invasive plants in New Zealand and the United Kingdom (Figures 3 and 4). Invasive species frequently are genetically similar across their invasive ranges (Benoit et al. Reference Benoit, Les, King, Na, Chen and Tippery2019; Tippery et al. Reference Tippery, Harms, Purcell, Hong, Häfliger, Killoy, Wolfe and Thum2023), resulting in part from introduction bottlenecks and anthropogenic movement (Dlugosch and Parker Reference Dlugosch and Parker2008; Smith et al. Reference Smith, Hodkinson, Villellas, Catford, Csergő, Blomberg, Crone, Ehrlén, Garcia, Laine and Roach2020). Only the J1 variant from our study lacked a comparable sequence from outside the United States, and this could indicate that the variant is only present in U.S. invasive plants or that the collection and molecular analysis protocols that have been used thus far in other countries have failed to locate a plant with the J1 variant. We recommend that molecular tools should be used across the invasive range to determine the genetic similarity among invasive Reynoutria plants worldwide and potentially to reconstruct the origins of various composite hybrid genotypes. Moreover, it may be possible to ascertain whether hybridization continues to generate novel genetic combinations.
Comparisons among amino acid sequences from the second exon, both within Reynoutria and in the related genus Fagopyrum, are consistent with the possibility that all versions of LEAFY could be functional. At this point we have not obtained evidence to suggest that any sequence variants have become nonfunctional, as can happen in polyploids (Adams and Wendel Reference Adams and Wendel2005; Edger and Pires Reference Edger and Pires2009; Roulin et al. Reference Roulin, Auer, Libault, Schlueter, Farmer, May, Stacey, Doerge and Jackson2013). Currently, Fagopyrum is the most phylogenetically similar genus whose LEAFY protein sequence can be compared with that of Reynoutria, and it would be beneficial to obtain equivalent sequences from the more closely related genera Fallopia Adans. or Muehlenbeckia Meisn. (Schuster et al. Reference Schuster, Wilson and Kron2011), as well as other Reynoutria species.
The term “next-generation sequencing” describes a variety of methods that can produce sequences from individual fragments of sample DNA (Goodwin et al. Reference Goodwin, McPherson and McCombie2016). For heterogeneous pools of DNA, such methods can efficiently separate out allelic variants (e.g., Macas et al. Reference Macas, Kejnovský, Neumann, Novák, Koblížková and Vyskot2011). Thus, next-generation sequencing offers an alternative to bacterial subcloning, which also can be effective for isolating DNA sequence variants (e.g., Moody and Les Reference Moody and Les2002). In this study, we employed both next-generation sequencing and bacterial subcloning and obtained equivalent results. We found the next-generation sequencing method to be cost-effective and more rapid than bacterial subcloning, and we recommend this method as a viable strategy for future plant identifications.
Several diagnostic primers were tested, and we confirmed their effectiveness for identifying plants that have R. japonica and/or R. sachalinensis sequence variants. Conducting PCR, followed by gel electrophoresis (and sometimes also including restriction enzyme digest), is a faster and less expensive alternative to DNA sequencing (e.g., Saltonstall Reference Saltonstall2003; Wendell et al. Reference Wendell, Huang, Gryspeerd and Freeland2021). The diagnostic primers reported herein are able to identify plants as species or hybrids; however, they were not designed to distinguish specific sequence variants from each species. The newly reported LEAFY gene sequences also enable additional diagnostic options to be explored, such as variant-specific sequencing primers (Scheen et al. Reference Scheen, Pfeil, Petri, Heidari, Nylinder and Oxelman2012) or diagnostic primers for environmental DNA (eDNA) analysis (e.g., Kuehne et al. Reference Kuehne, Ostberg, Chase, Duda and Olden2020).
The invasive Reynoutria are now well represented by LEAFY sequence data that include the first intron and second exon. However, there are other Reynoutria taxa, as well as species in related genera, for which no LEAFY sequences have been published. The methods employed herein should be applied to other taxa to assess phylogenetic relationships and investigate recent hybridization or polyploid species origins.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/inp.2024.36
Data availability
Newly generated DNA sequences were deposited to GenBank, under the accession numbers that are referenced in Supplementary Table S1.
Acknowledgments
We are grateful for the assistance of plant collectors and the coordinators of the plant collecting efforts, in particular Maureen Kalscheur (Wisconsin Department of Natural Resources) and Matt Wallrath (University of Wisconsin Extension). We appreciate the comments and suggestions of two anonymous reviewers that helped us to improve the article.
Funding statement
Funding was provided by the University of Wisconsin–Whitewater Undergraduate Research Program and the University of Wisconsin–Whitewater Department of Biology.
Competing interests
The authors declare no conflict of interest.
Open access
