Intra- and interspecies competition of blackgrass and wheat in the context of herbicidal resistance and environmental conditions in Poland

Blackgrass (Alopecurus myosuroides Huds.), one of the most aggressive grass weeds in Europe, is also a strong competitor of crops. This study aimed to assess the impact of environmental conditions on the competition between (1) ACC-ase and ALS herbicide-resistant (BR) and herbicide-susceptible (BS) blackgrass biotypes, (2) BR and winter wheat cv. Arkadia (W), and (3) BS and W. In the replacement series model, the experiment was conducted at seven sites across Poland during two seasons (2018/19 and 2019/20). In the BR-BS experiment, the BS biotype was in majority more competitive toward the BR biotype. However, in the regime of optimal hydrothermal conditions and at a higher sand content in the soil we observed a higher competitiveness of BR towards BS. The combined interactions between W and BR or BS were also affected by environmental conditions, i.e., soil texture and hydrothermal coefficient, as explained by PCA and k-means cluster analysis. At most sites, W was more competitive toward both BS and BR, which could result from earlier emergence of W in relation to B in majority of sites. Except for two cases, located on heavy, clay soils, during humid seasons, where B was more competitive toward W. We summarize that blackgrass competitiveness towards other biotypes and wheat depends to some extent on environmental conditions; however, the phenomenon should be explored in more detail.


Confirmation of blackgrass biotype resistance to ALS and ACCase inhibitors. The amplification
of the blackgrass ALS sequence resulted in the generation of a 1919 bp nucleotide sequence. Mutations were identified in the analyzed nucleotide sequence fragment, where 22 changes were synonymous and seven were nonsynonymous. The analysis of the amino acid sequence to find mutations known to confer resistance to ALS inhibitors revealed the presence of an amino acid substitution at position P197. Two mutations were found: P197A and P197T (Fig. 1A).
The amplification of the blackgrass ACCase sequence fragment resulted in the generation of 2879 bp nucleotide sequences comprising nucleotides from 9223 to 120,101 in the ACCase sequence. The analyzed fragment contained a DNA sequence encoding the carboxyl-transferase (CT) domain (nucleotides from 1117 to 2765 of the analyzed PCR product), which is known to be the site of action for the active substance of the herbicides belonging to group 1. Four mutations were identified in the analyzed nucleotide sequence of the CT domain, where three changes were synonymous and one was nonsynonymous. The nonsynonymous mutation resulted in the substitution of Trp to Leu at position 1999 in the amino acid sequence (Fig. 1B), which is a mutation known for conferring resistance to ACCase inhibitors.
Competition between resistant and susceptible biotypes of blackgrass. ANOVA effects for the relative plant biomass (BCRb) and the relative grain number (BCRgn) of S and R blackgrass biotypes are presented in Table 1. The single effect for site and the interaction effect for year × site were significant for both BCRb and BCRgn, while BCRgn was also significantly affected by the single effect of year.
The values of BCRb (BCRgn) < 1 indicate that the R biotype is more competitive toward the S biotype in the case of biomass (grain) production. In contrast, if BCRb (BCRgn) > 1, then the S biotype is more competitive toward R.
A detailed site × year comparison showed that BR and BS biotypes competed differently. During 2018/19, the BR biotype was more competitive in BCRb at two sites, i.e., Lipnik and Swojczyce, and in BCRgn at one site, i.e., Swojczyce ( Fig. 2A). The BS biotype was also more competitive toward BR in BCRb in two sites, i.e., Wrocław and Mochełek, and in BCRgn in all the sites, except for Swojczyce. There was a positive correlation (r = 0.69) between BCRb and BCRgn in 2018/19, but the correlation was nonsignificant (Fig. 3A). In terms of the correlation of each www.nature.com/scientificreports/ competitive ratio in 2018/2019 with the hydrothermal coefficient K, a negative correlation was found, especially for BCRgn (r = − 0.50) (Fig. 3B,C), indicating that the drier the weather was, the more competitive the BS biotype; however, the trends were not significant. There was no correlation found for the competitive ratios and the soil conditions, expressed here as a percentage content of sand in the soil (Fig. 3D,E).
In the 2019/20 season, the average BCRgn value was significantly, by 38%, lower than that in 2018/19, whereas the value of BCRb was similar (Fig. 2B). During that season, the BR biotype was more competitive toward BS again in Swojczyce but also in Mochełek and in terms of BCRb in Wrocław and BCRgn in Lipnik. In this season, both the competitive ratios BCRb and BCRgn were positively (r = 0.87) and significantly correlated (Fig. 3F). The hydrothermal conditions K indicated mostly humid conditions in Poland in the 2019/20 season (Table 9, Materials and methods section). The r-Pearson correlations between the K coefficient and the BCRb (BCRgn) indexes had the opposite direction (r = 0.7 and 0.59, respectively) as in the dry season of 2018/2019. The increasing K provoked an increase in the competitiveness of the S biotype, both in biomass and grain production (Fig. 3G,H). On the other hand, a negative correlation occurred in the 2019/2020 season between the sand content [%] in the soil and BCRb (BCRgn), r = − 0.70 and r = − 0.85 (significant), respectively (Fig. 3 I,J). The higher the sand content in the soil was, the higher the competitiveness of the BR biotype towards BS.
Competition between winter wheat and resistant/susceptible biotypes of blackgrass. The emergence of winter wheat (W) was noted 8-16 and 6-19 days after sowing in 2018 and 2019, respectively ( Table 2). The fastest emergence of W was noted in 2018 in Mydlniki and Lipnik and the longest in Mochełek and Czesławice. Even though the dates of sowing for both W and B were the same, B emerged, 0-15 days later than W at most sites. In 2018, the emergence of B on the same day as W occurred in Lipnik and Mochełek, while in Winna Góra and Mydlniki, B emerged 6 days after W. In autumn 2019, the emergence of W and B was at the same date in Mochełek and Mydlniki (Table 2).  Table 1. Mean squares from the three-way analysis of variance for the competitive ratio of susceptible (S) and herbicide-resistant blackgrass (R) calculated for the relative plant biomass (BCRb) and relative grain number (BCRgn). df degrees of freedom. *P < 0.05; **P < 0.01; ***P < 0.001.  (Table 3). Regarding WBCRgn, a significant effect was also detected in the interaction site × biotype. For WBCRb, the interaction of year × biotype × site was also significant.

Source of variation
The individual model of competition between blackgrass and winter wheat for each site, both R and S biotypes, and during two seasons was confirmed by the χ 2 goodness-of-fit. They are presented in the figures in the supplemental materials (Figs. S1-S7). A typical model for the competition between winter wheat (W) and blackgrass (B) occurred in two variants: model IIa, which represents the higher relative yield of B, and model IIb, when the higher relative yield was attributed to W (Fig. 4).
The average competitive ratios for both biomass (WBCRb) and grain number (WBCRgn) were similar for the susceptible and resistant B biotypes in the competition with W (Figs. 5 and 6). However, in terms of WBCRb, there was a significant difference between seasons, with the R biotype being less competitive in 2018/2019 than in 2019/2020. Stronger competitiveness in biomass production was attributed to BR in the five locations in 2019/2020 (Mydlniki, Lipnik, Czesławice, Mochełek and Swojczyce) (Fig. 5C), while in 2018/2019, it was only at two locations, i.e. Swojczyce and Mydlniki (Fig. 5A). In 24 cases across Poland, W was more competitive than B in the mixture, as the RY (relative yield) of W was higher in the mixture (convex line) than the RY of B (concave line)-Model IIb (Fig. 4). Mydlniki was the only site where B, both R and S biotypes, was more competitive in the mixture with W, or no competitiveness occurred between both species in both seasons (Fig. S7) The competitive ratios for Mydlniki for the WBCRgn and WBCRb were low, 0.7-1.4 and 0.9-1.0, respectively (Figs. 5 and 6). Additionally, in the 2018/19 season, both the S and R biotypes in Swojczyce were more competitive toward W in the WBCRgn (Fig. 6A,C).

Principle component analysis and cluster analysis.
The relationships between all traits of winter wheat W grown with BR and BS, as well as the environmental parameters, were calculated using principal multidimensional component (PCA) and cluster (CA) analyses separately for the 2018/19 (Fig. 7A,C) and 2019/20 (Fig. 7B,D) seasons, as the variation in years was in interaction with sites or biotype. Three characteristics of   www.nature.com/scientificreports/ soil (sand, silt, and clay contents) were uniform in the two seasons, while the hydrothermal coefficients K and the W parameters varied in both seasons. In 2018/19, the first component (Y1) explained 53.1% of the total variance. It was significant for the silt content (× 2) and plant biomass (× 8), seed yield (× 6), and number of seeds (× 9) in the same direction, while the content of sand (× 1) was significant in the opposite direction: (× 1 − × 9 correspond to the characteristics in Fig. 7A). The second principal component (Y2) explained 23.8% of the total variance (Y 2 = − 0.39x 1 + 0.29x 2 + 0.66x 3 + 0.23x 4 − 0.15x 5 − 0.16x 6 − 0.91x 7 − 0.62x 8 + 0.06x 9 ), and it was significant for the thousand grain weight (TGW, × 7) (Fig. 7A). The projection of seven locations across Poland, where the BS/BR mixtures with W were planted, displayed two distinctive groups, i.e., Lipnik, Mochełek, Winna Góra, and Swojczyce, highly related to the sandy soil (mean = 76.63%) and low moisture (K = 1.1) in 2018/19 (Fig. 7C). Therefore, the alternation in significantly lower seed number (111.4) and seed yield (5.06) was found ( Table 4). The second group consisted of sites, e.g., Czesławice, Wrocław, and Modlniki, where the W yield was significantly higher due to heavier soils (silt content = 49.25) and a higher hydrothermal coefficient (K = 1.97), so the seed number amounted to 245.1 and the yield amounted to 10.19 (Table 4).

Discussion
In Poland, blackgrass herbicide resistance has been recorded mainly in the northern and northeastern parts (Pomerania, Warmia-Mazury) and the Lower Silesia region 24,43 . The resistant biotype tested in our study had mutations at single positions within the ALS and ACCase sequences. In the ALS amino acid the changes at position P197 were identified. P197T was previously reported in A. myosuroides plants 44,45 , while P197A was detected in the Lolium rigidum and Conyza canadensis populations 46,47 . The W1999L mutation found in the ACCase sequence was also acknowledged to confer resistance to the ACCase-inhibiting herbicide fenoxaprop in Alopecurus japonicus 48 .
Various modeling approaches might be employed to examine the effectiveness of integrated weed management, reducing the need for more expensive and cumbersome long-term in situ experimentation 49 . We conducted  Table 3. Mean squares from the three-way analysis of variance for the competitive ratio of winter wheat (W) and herbicide-resistant or herbicide-susceptible blackgrass (B) calculated for the relative grain number (WBCRgn) and plant biomass (WBCRb). df degrees of freedom. *P < 0.05; **P < 0.01; ***P < 0.001. www.nature.com/scientificreports/ short-term (two growing seasons) experiments in situ in the replacement series model as a tool to investigate the competitiveness between blackgrass-blackgrass and wheat-blackgrass with environmental factors and blackgrass herbicide resistance. Within two years of study, the tested blackgrass biotypes, herbicide-susceptible (BS) and herbicide-resistant (BR), displayed different competitiveness with each other and with winter wheat. Based on two competitive ratios, i.e., plant biomass and the number of grains produced by the tested plants, we could draw main competitive relations, significant from biological and agronomic points of view 50 , in different sites across Poland. The sites in this experiment were diversified in terms of soil conditions, from the lightest loamy sands (Lipnik, Mochełek), sandy loams (Winna Góra, Swojczyce), medium-heavy sandy clay loam (Wrocław), to heavy and compacted silt loams (Mydlniki, Czesławice). Additionally, two vegetative seasons during the studies displayed contrasting weather conditions; the first season was dry in most of the sites, whereas the other season was humid.

Source of variation
In the BR-BS experiment, on average, the BS biotype was more competitive toward the BR biotype, which was expressed by the values of the competitive ratio greater than one. This is consistent with other authors' findings, who point out that herbicide-susceptible weed biotypes perform better than resistant ones in the absence of herbicide selective pressure. For example, under drought stress, the nontarget site-resistant biotype of blackgrass germinated less than the susceptible biotype 51 . In other studies, the blackgrass biotype with enhanced metabolism resistance to ALS inhibitors produced longer tillers than the susceptible plants during vegetative growth. However, there was a reproductive cost of resistance, i.e., 27% fewer seed heads per plant and a 23% reduction in total seed head length 52 . The worse performance of herbicide-resistant biotypes, the so-called resistance penalty, was also observed in other herbicide-resistant weed species [53][54][55] . Our study also showed site-specific correlations between both competitive ratios of BR and BS biotypes, especially in the humid season. We found that the content of sand affected the competitive relations between the BS-BR biotypes during the humid season. Specifically, the more sand there was in the soil, the higher the competitiveness of the BR biotype toward the BS biotype, and vice versa. In our opinion, this finding suggests the existence of internal plasticity of blackgrass biotypes that change their intraspecies competitive efforts depending on the site-specific conditions. www.nature.com/scientificreports/ The diversified soil and weather conditions also affected the studied parameters of winter wheat (seed number and biomass), as was observed in the wheat (W) -barnyardgrass (BS/BR) experiment. The multivariate data mining at each site and BS/BR mixtures with winter wheat explained the division into two distinctive in situ groups. The first group was highly related to sandy soil and low moisture. In contrast, the second group had heavier soils of higher moisture where wheat productivity was significantly higher, as confirmed in the studies by 56 . In our study, wheat cv. Arkadia has been shown to suppress blackgrass at most study sites, particularly in the humid season of 2019/20 and regardless of the herbicide-susceptibility level of B. Increasing wheat competitiveness against weeds through a combination of breeding and planting design (planting density, row spacing, and orientation) has a strong potential to reduce weed-induced yield losses in wheat 30,57 . Sardana 58 declared that cultivars possessing traits such as fast germination and growth, high biomass, and large leaf area have a competitive advantage over weeds. Lazzaro 59 listed four traits linked to the competitive ability of wheat against weeds, i.e., aboveground biomass before stem elongation, tillering index, plant height, flag leaf morphology, and two production-related traits (grain yield and thousand-kernel weight). The combined effect on wheat-blackgrass interactions was also studied by Andrew and Storkey 60 using the INTERCOM model parameterized for two wheat cultivars with contrasting competitive ability and simulations run across 10 years at different crop densities and two sowing dates. The authors found that sowing date, sowing density, and cultivar choice largely work in a complementary fashion, allowing enhanced competitive ability against weeds when used in combination. However, the relative benefit of choosing a more competitive cultivar decreases at later sowing dates and higher crop densities. Perhaps another important element in the competitiveness of W against B in our study that supported W competitiveness was the time of W emergence, which was in the vast majority of sites earlier than the emergence of B, or at least both species emerged on the same day. As other studies confirm, a delay in weed emergence to wheat emergence enhances wheat competitiveness toward weeds and a final wheat grain yield 42,61-63 .
The observed W-B competition pattern exception was noted only in two sites, namely, Mydlniki and Czesławice (south and south-eastern Poland), located on heavy, clay soils, where B was more competitive toward W during humid seasons. Blackgrass also takes advantage of environmental conditions, such as clay content and moisture in the soil 6,10 , which could support its competitiveness with W at both sites.  www.nature.com/scientificreports/

Conclusions
Our study showed all the competitive replacement series models between blackgrass-blackgrass and winter wheat-blackgrass of different herbicide susceptibilities (R-resistant, S-susceptible). There were site-specific changes in the blackgrass R and S competitive abilities, depending on the sand soil content, supporting the thesis about blackgrass plasticity depending on environmental conditions. Winter wheat in our study was a strong competitor for blackgrass; only in three out of a total of 28 cases was blackgrass more competitive toward wheat. One of the winter wheat characteristics that could affect this relationship was that wheat emerged 15-0 days earlier than blackgrass. A more detailed understanding of belowground competition may also be required to increase the robustness of the predictions when water or nutrients are limiting.

Materials and methods
Plant material. Ethics approval and consent to participate. The collection of biotypes of blackgrass samples from plantations was permitted by local famers on survey. The plant collection and the study complied with local and national (Poland) regulations. The study was approved by the institutional research ethics committee of Institute of Plant Protection in Poznań, and written informed consent was obtained from each participant. All the methods in this manuscript were carried out in accordance with relevant guidelines and regulations. Two biotypes of blackgrass (B) with contrasting susceptibility to herbicides were used in the two-year-long pot experiments at seven sites across Poland. Both biotypes were collected in Poland in 2017. The herbicide multiresistant (BR) biotype was from western Poland, voivodship Lubuskie, from winter wheat. The susceptible biotype (BS) was from eastern Poland from the ruderal site. An initial level of herbicide resistance of both BR and BS biotypes was confirmed in two sets of greenhouse dose-response tests carried out in 2017 and is presented in Table 5.   Table, and 30 s or 45 s at 72 °C for the amplification of ALS and ACCase sequences, respectively. The final elongation step was set to 2 min at 72 °C. The reaction products were separated with 1% gel electrophoresis followed by purification from the gel with a Wizard SV Gel and PCR Clean-Up System (Promega, Madison, WI, USA). ALS and ACCase sequence fragments were ligated to the pJET1.2 plasmid using a CloneJET PCR Cloning Kit (Thermo Fisher Scientific, Waltham, MA, USA) and cloned into DH10B Escherichia coli competent cells (Table 6). Three plasmids were isolated from E. coli cells using the NucleoSpin Plasmid (Mecherey-Nagel). Three plasmids containing inserts synthesized in each PCR were digested with BglII to confirm the presence of the insert and sequenced by Genomed (Warsaw, Poland). Sequencing data were analyzed using the BioEdit Sequence Alignment Editor 7.5.5 64 .

Molecular analysis of the herbicide-resistant blackgrass biotype.
Description of the replacement-series pot experiments. Following the replacement series competition model 65 , the pot experiments in field conditions were designed and set across Poland at six or seven sites (Fig. 8, Table 7). The experimental design was based on randomized blocks with three replications and took place in the next two growing seasons, 2018/19 and 2019/20. At each of the sites, pots (7 l vol., 0,0075 m 2 area) were dug into the soil at a distance of 0.5 m, and ca. A total of 2.5 cm of pots remained above the soil surface. Subsequently, they were filled with local arable soil from the 0-30 cm layer. Prior to placing it in the pots, the soil was sieved through a metal mesh to remove stones and www.nature.com/scientificreports/ other larger impurities. Soil samples were taken for detailed texture and chemical analyses (Table 7). Black foil was placed in the area between pots, and the whole experiment was protected from insects and birds with a fine plastic net. The layout of pots and the distribution of plants in a pot remained the same at every site. The competitive effects of BR and BS biotypes were studied in two proportions of plants R4:S6 and R6:S4 at six sites, i.e., Lipnik, Mochełek, Czesławice, Swojczyce, Wrocław, and Mydlniki. The competitive effects of BR or BS were tested against winter wheat (W) cv. Arkadia (breeder: HR Danko PL) at all seven sites. The proportions of winter wheat W and BR were in the ratios of W10:R0, W8:R2, W6:R4, W4:R6, W2:R8, and W0:R10. The same ratios were used for the W and BS plants: W10:S0, W8:S2, W6:S4, W4:S6, W2:S8, and W0:S10.
Seeds of the W, BR, and BS biotypes were sown on the same day, optimal for W ( Table 8). The grains of B were sown at a depth of 0.5 cm, and W was sown at a depth of 2 cm, with a few additional grains per spot. The sowing spot of each plant was marked.
In the spring, the density of W and B plants was regulated according to the layout. During that time, N fertilization was applied in 30.0 g NH 4 NO 3 m −2 ; 50% of the dose was applied at the beginning of vegetation, and the remaining dose was applied at wheat shooting (BBCH 31-33). Wheat and blackgrass were harvested at full-grain      Weather conditions. Weather data, i.e., precipitation and air temperature, were collected during the study from the nearest weather stations (Tables S1 and S2). The weather data for the months from winter wheat sowing till harvesting, with temperatures > 0 °C, namely, October-November 2018 and April-July 2019, October-November 2019, and April-July 2020, were valorized using the hydrothermal coefficient (K) according to the equation:   where RY is the relative yield and p is the proportion of biotype. If the competitive ratio BCR < 1, the R biotype is more competitive toward the S biotype; if BCR > 1, the S biotype is more competitive toward the R biotype.
where RY is the relative yield and p is the proportion of species.
If the competitive ratio WBCR < 1, B is more competitive toward W; if WBCR > 1, W is more competitive toward B.
The RY and the total relative yield (TRY) for all the measured parameters of B and W were calculated according to the formulas: The analysis was performed for the replacement series experiment 63 . where RY W is the relative yield of W; RY B is the relative yield of B; p is the proportion of species; W mix is the value of the W parameter analyzed for the mixture; W mon is the value of the W parameter analyzed for the monoculture; B mix is the value of the B parameter analyzed for the mixture; B mon is the value of the B parameter analyzed for the monoculture; and TRY is the total relative yield. The RY values were measured for both W and B and were averaged for a single plant.
The biomass and grain yield or seed number, calculated into RY and TRY, were presented as graphs and fitted into one of the five competition models according to Radosevich 68 . The most explanatory models were determined by using the χ 2 goodness-of-fit 69 , giving the best fit for all tested sites and biotypes. If RY is a straight line, it denotes no competition; a convex line shows a benefit to species; a concave line represents a loss to species. If TRY equals 1 (straight line), there is competition for the same resources between both species/biotypes. If TRY is greater than 1 (convex), there is no competition since the demand does not exceed the resources. If TRY is less than 1 (concave), an antagonism resulting in a mutual loss to the species involved becomes apparent 67 .
ANOVA of two models was performed, i.e. the two-way model for the year × site of plants to calculate the BCRgn and BCRb in R/S blackgrass and the three-way model for the year × site × biotype to calculate the WBCRgn and WBCRb in wheat/blackgrass. Data from the indexes were transformed to a normal distribution. The post hoc HSD Tukey test was used for the means separation. The r-Pearson correlation coefficient was calculated for relations between BCRgn/WBCRgn and BCRb/WBCRb, the relations between the hydrothermal K coefficient and BCRgn/WBCRgn and BCRb/WBCRb and the relations between sand [%] and BCRgn/WBCRgn and BCRb/WBCRb. The exploration technique of principal component analysis (PCA) was used to explain the multidimensional diversity of W and environmental variables in terms of the first two components. Five biometrical parameters of W, i.e. mean length and fresh biomass of the plant, number of seeds per plant, yield of seeds per plant, and TGW (1000 grain weight), and four environmental parameters, i.e. The hydrothermal coefficient K and soil texture were analyzed using PCA and CA (k-means procedure).