Amaranthus retroflexus L., native to America, is now widely distributed in many cultivation areas of China. It occurs mainly in soybean and corn fields, and also infests wheat, sweet potato, cotton and other fields [1]. Due to its strong adaptability and reproductive ability, A. retroflexus has become a malignant weed that competes with crops for nutrients and water, resulting in a significant decrease in grain yield and quality of crops [2].

With the introduction of herbicides in the 1940s, which replaced the labor-intensive weed control measures of early agriculture, herbicides have been the most effective control tool for weed management to date. Nicosulfuron, a sulfonylurea herbicide, has been widely used due to its broad weed spectrum, low application amount and high crop safety [3]. However, intense selection pressure from long-term single application results in the evolution of herbicide-resistant weed biotypes, and now resistant A. retroflexus populations have been reported worldwide [4]. The existence of resistant weeds reduces the efficacy of corresponding herbicides, makes it difficult to select herbicides, and increases herbicide doses. These problems not only increase the cost of weed control and cause crop yield loss, but also aggravate the risk of herbicide damage and environmental pollution. At present, resistant weeds have become a serious problem in weed control.

Fitness is considered to be an ability that helps an organism establish, survive, and reproduce successfully [5]. The effects of resistance evolution involve multiple traits in several processes of plant growth and ultimately change fitness in different ways. Plant functional traits such as survival, growth, and reproduction can be measured at individual level and determine the fitness components of the plant [6]. The fitness cost in young plants dictates whether they have benefits or drawbacks in competitive settings. Therefore, it’s essential to investigate the manifestation of these fitness costs and the alterations in life history linked to characteristics and events in the plant’s early stages [7].

Seed germination which signifies the beginning of life cycle provides opportunity to judge the evolutionary dynamics of seedling competitive ability. Due to the heterogeneity of environmental gradients, seed germination and relative dormancy after shedding are influenced by diverse environmental factors that modify seed physiology and behavior [8, 9]. In the study of fitness cost of resistant populations, the difference in germination characteristics has been reported as one of the components [10]. There was no major difference in seed germination under diverse temperatures, pH ranges, and light conditions. However, this study showed that the seeds tended to respond differently to a gradient of osmotic and salt stress in resistant and susceptible Polypogon fugax populations [10]. Amaranthus powellii with Trp574Leu acetylhydroxylate synthase (AHAS) mutations appears to have multiple pleiotropic effects on the initial stages of the plant life cycle, while Polypogon fugax had a decreased fitness due to its evolved resistance to aryloxyphenoxy propanoate herbicides [11, 12]. Up to now, however, there is still little research on the different germination responses of resistant A. retroflexus populations under various stress treatments. The purpose of this study is to: (a) explore the impacts of different environmental gradients such as temperature, light, pH, salt stress, osmotic potential energy, and burial depth on the germination of nicosulfuron-resistant and -sensitive A. retroflexus populations; (b) compare the germination responses of the nicosulfuron-resistant and -sensitive populations to these treatments; (c) search for effective tools to assist resistant weed management.

Stress often has negative effects on plants, especially in terms of growth, development and reproduction [14]. It is able to determine the distribution of species, while providing a selective evolutionary pressure on a particular population [15]. As a result of weed resistance evolution, fitness cost can predict the evolutionary trend of resistant and sensitive populations, and seed germination may be the most important phenological event determining the successful establishment of annual plants [16]. High germination ability mean that space is captured more efficiently, thus ensuring their survival and distribution in the population and the preservation of evolutionary traits. This study mainly took into account that A. retroflexus is an invasive organism with a wide range of habitats, so different stress gradient factors were set up to compare the responses of nicosulfuron-resistant and sensitive populations to temperature, light, salt, pH, osmotic potential and burial depth.

Germination rate of C4 species are reported to increase with temperature. The results of this study are consistant with this characteristic. The experimental data show that the resistant and sensitive populations of A. retroflexus can germinate in a broad spectrum of temperatures, and the germination rate of them is more than 90% within 24 h under suitable temperature conditions. Light can stimulate the germination of A. retroflexus to a certain extent. The interaction between temperature and light affected the germination of A. retroflexus. The germination rate in A. retroflexus S population was slightly higher than that of R population in the higher temperature, while the difference was not significant, which was consistent with previous study [17]. Despite the difference being relatively small, even a small increase in germination rates could provide a competitive advantage, especially for invasive species [18].

In darkness, the germination rate in R population decreased linearly with the decrease of temperature, and the germination index in S population was significantly higher than that in R population at temperature of 40/35°C and 35/30°C, indicating that the seed vitality of R population was worse than that of S population. At low temperatures, the catabolism of stored proteins slows down and the available levels of free amino acids are reduced [5]. The germination index in R population was higher than that in S population under the relatively low temperature of 25°C/20°C. A similar situation has been found in Lactuca sativa L. and Kochia scoparia when temperatures are low in late autumn or winter [5, 19]. It was suggested that the rapid germination of R populations at low temperature was related to the increase of the level of branched-chain amino acids [18, 20]. In this study, the minimum temperature was set at 20°C, and this characteristic may be more significant at lower temperatures.

Extreme salinity and osmotic potential levels are usually the main abiotic limiting factors for plants and are not conducive to the germination of many weeds [21–24]. Sodium content modifies soil composition and replaces calcium and magnesium through an anionic exchange process, resulting in water and nutrient deficiencies [25]. It was found that the R and S populations had a certain tolerance to sodium chloride, and the germination rate at 160 mM was still higher than 50%. Although osmotic potential stress increased, the germination rates in R and S populations also gradually decreased, the germination rates remained higher than 80% from 0 Mpa to -0.6 Mpa, indicating that osmotic potential and salt stress were not the main factors affecting the germination of A. retroflexus.

Under the salt stress and osmotic potential influences, the germination rate in S population was higher than in R population. Similar results were found in other studies for Polypogon fugax, Alopecurus japonicus [12, 26]. However, for glyphosate-resistant and -susceptible Italian ryegrass, the germination ability of resistant population was reported to be higher than that of S population [27]. When salt stress was 80 mM and water potential was -0.1 Mpa~-0.4 Mpa in the current study, the germination index in S population was obviously higher than that in R population. Compared with the S population, R population has a lower germination rate and germination delay under salt stress and osmotic potential stress, which was expected to be enhanced by a higher probability of escaping pre-planting management measures [28–30].

The seeds had good adaptability and maintained a high germination rate in the range of pH 4~10, suggesting that pH value was not a limiting factor for the germination of A. retroflexus. Similar conclusions were also obtained in studies on R and S Alopecurus japonicus biotypes [26]. Many weeds have been studied and found to have wide germination range on pH, such as Lolium rigidum [24], Polypogon fugax [31] and Beckmannia syzigachne [32]. Based on the data from this experiment, it was found that the resistant population evolved a weak sensitivity to acidic environment compared with the sensitive population. In addition, the vitality of resistant seeds under pH conditions was significantly lower than that of sensitive populations except for weak alkali conditions. The germination of early weed seeds could be controlled by preemergence herbicide, while resistant weeds with irregular emergence would be exposed to herbicides with a high probability and have more opportunities for selective evolution [30].

In this study, the optimal burial depth for germination in R and S populations was 2 cm, and the germination rate decreased when depth of burial increased. The same has been reported for some species of weeds, such as Brunnichia ovata [21, 33]. Seeds of A. retroflexus can not germinate with exposure at a depth of 6 cm or above. Due to its small seed size, the nutrients contained in the germ are not enough to support the cotyledon to emerge from the surface in the absence of light [34]. Therefore, effective weed management measures can be considered in combination with changing tillage systems and using deep tillage operations. From the results of this experiment, we can see that the seed germination rate placed on the soil surface was lower than that placed in the petri dish, and similar results were found in other studies of Lolium rigidum [24, 35]. This difference may be caused by poor seed-soil contact, low water utilization, lack of light, fluctuation in moisture at soil surface or possibly predation and etc [24].

The studies above have shown that weed seed germination is determined by the interaction of various factors, which are both influenced by genetics and driven by environment [36]. Other studies suggest that alleles equipped with herbicide resistance have pleiotropic impacts on the life cycle of weeds [37, 38]. The results of this study showed that in response to temperature, light, salt, pH, osmotic potential and buried depth, the germination ability of resistant species was almost all lower than that of sensitive species. The significant germination delay in the adverse environment such as strong acid and alkali was considered to be an evolutionary strategy of resistant species, and the germination characteristics showed a decline in fitness. Reduced fitness of resistant plants further slows the evolution of resistance [39]. Enhanced seed dormancy can impact management strategies for preventing or controlling herbicide resistance, including the rotation of herbicides [40]. Results from these experiments may be helpful for the design of weed management strategies, such as deep tillage, cover crop, crop rotation, etc.

As a malignant weed in agricultural systems, A. retroflexus has been reported to have evolved varying degrees of resistance to the commonly used herbicide nicosulfuron. The results showed that both the resistant and sensitive populations could germinate in a wide range of osmotic potential, salt and pH values, which were not the main abiotic limiting factors for the germination of this species. The extraordinary ability of A. retroflexus to germinate under a variety of abiotic stress enables it to be a widespread and problematic weed in different areas. Our findings also demonstrated that there were variations in germination characteristics between nicosulfuron-resistant and -sensitive populations, which possibly depend on the resistance genes and their different genetic background. Germination of A. retroflexus had a strong relationship with temperature and burial depth. At lower temperature of 25°C/20°C, its germination would reduce greatly, and when the burial depth was ≥ 6 cm, both the R and S populations can not germinate. Thus, proper early planting time, optimum tillage depth and cover crop residue cover on the soil surface could be studied further for its management.