Foxtail millet (Setaria italica L.), a crop rich in nutrients, originated in China [1, 2]. Presently, foxtail millet is extensively cultivated as a food and fodder crop throughout Eurasia and the Far East [3]. Meanwhile, foxtail millet is being promoted as a model crop for cereal crops [4–8]. However, the yield and quality of crops are seriously affected by weed growth [9]. The competitions from weeds have been reported to reduce foxtail millet grain yield by as much as 55.56% [10]. Manual control with intensive labour is the main method to control weeds for the crop. Compared with conventional manual control of weeds, chemical control is less labor-intensive and more efficient. However, there is no suitable herbicide used for weed control in foxtail millet fields during the post-emergence stage.

Tribenuron-methyl (TBM) is a highly selective post-emergence sulfonylurea herbicide developed by the DuPont Company, USA. TBM belongs to the acetyl lactic acid synthase inhibitor, and can effectively control the vast majority of annual broadleaf weeds [11, 12]. There are significant differences in the sensitivity of wheat varieties to TBM [13]. Wheat varieties with high activity of glutathione S-transferase are more tolerant to TBM [14, 15]. At three-leaf stage of weeds, the recommended dosage of 75% TBM treatment shows better effect in broadleaf weed control and is relatively safe to hulless oat [16]. Rational application of TBM effectively inhibits weed growth, increases light penetration rate in fields, and enhances the yield of wheat; otherwise, it can seriously inhibit the yield and grain quality of wheat [17]. However, there is little evidence whether TBM is safe to foxtail millet crops or not.

Moreover, studies on chemical weed control in foxtail millet fields are rare. Application of 1800 g ha⁻¹ of 44% monosulfuron plus propazine Wettable Powder (WP) at the pre-emergence stage can control the vast majority of broadleaf weeds and certain gramineae weeds in foxtail millet fields [18]. Sethoxydim is a safe post-emergence herbicide for gramineae weeds control in Zhangzagu 5, which has a sethoxydim resistant gene [19]. Additionally, with increasing monosulfuron concentrations, grain content of fat decreases while starch and amino acid contents increase in foxtail millet [20].

Grain yield and quality determine much of the value of cereal crops to the producers [21, 22]. Weeds can be controlled using herbicides [23], but certain crop damages are likely as the yield and quality affected by herbicides [24, 25]. Therefore, there is a need to study the effect of herbicides on crop grain yield and quality. Presently, there is little knowledge about the effect of the post-emergence herbicide TBM on the grain yield and quality of foxtail millet.

To address the above issue, the present study was carried out to investigate the effect and safety of the post-emergence herbicide TBM on the performance, yield, and grain quality of Zhangzagu 10 and Jingu 21. The variety Jingu 21, one of the main varieties of foxtail millet in China, which has been widely grown in the major foxtail millet production regions of northern China for 30 years, is a high-quality conventional variety. The variety Zhangzagu 10 is a new hybrid variety, whose one parent plant was obtained from improved Jingu 21, which showed enormous promotion potential in the world. The results will have certain guiding significance for the production and promotion of Zhangzagu 10.

The changes in foxtail millet grain yield and quality were remarkable under herbicide TBM stress. The safety of herbicides on crops can be assessed by measuring grain yield, quality, and agronomic traits [28, 29]. In this study, TBM treatment inhibited the growth of foxtail millet, which was reflected by reduced root and shoot biomass (Figs 1 and 2). Additionally, a decline in grain yield of foxtail millet was obtained following TBM treatment (Tables 2 and 3). Wang et al. [17] reported that increasing dosages of TBM negatively impact yield and yield compositions of wheat under artificial dislodgement conditions. In this study, yield reduction varied between years and varieties. Differences in yield reduction might be attributed to varietal properties of herbicide tolerance and climatic conditions over the 2-year study period. The annual average precipitation during the filling stage (August through September) was higher in 2014 versus 2013. Consequently, foxtail millet yield and 1000-grain weight were lower in 2014 than those in 2013.

Foxtail millet yield was determined by grains per panicle, 1000-grain weight, and panicles ha⁻¹. Grain yield cannot be determined by a single yield component, because several yield components affect yield to varying degrees at different dosages of TBM. The herbicide directly affected the principal yield components, so as to the yield [28, 30, 31]. Grain yield was positively correlated with grains per panicle and 1000-grain weight, but not with panicles ha⁻¹ (Table 4). Panicle length and diameter may represent other good ways of quickly estimating potential yield reduction caused by TBM exposure. A reduction in panicle length and diameter is likely associated with fewer grains per panicle because of a reduced sink. Furthermore, panicle length and diameter were positively correlated with grains per panicle, 1000-grain weight, and yield, (Table 4). Therefore, a reduction in panicle length and diameter would reduce yield components causing a reduction in grain yield. These results were consistent with those obtained by Robinson et al. [28], who reported that several yield components were reduced by decreased agronomic characteristics after exposure to herbicides, leading to reduced yield. In this study, foxtail millet was damaged by high dosages of TBM. As a result, panicle length, panicle diameter, grains per panicle, and 1000-grain weight were significantly reduced, thereby decreasing grain yield. In Zhangzagu 10, the damage was caused by TBM at dosages ≥45 g ai ha⁻¹. However, in Jingu 21, the damage was caused at dosages ≥22.5 g ai ha⁻¹. Our results revealed that compared with Jingu 21, the hybrid Zhangzagu 10 had higher grain yield and greater herbicide tolerance.

Grain protein content is an important attribute in the nutritional quality of foxtail millet, which contains the highest protein quality among the cereal crops [32]. However, improper application of herbicide TBM affects grain protein content [33]. In this study, the grain protein content of foxtail millet decreased at high dosages of TBM, which may be attributed to the inhibitory action of the herbicide on acetolactate synthase, an enzyme responsible for the biosynthesis of branched chain amino acids (e.g., isoleucine, leucine, and valine) [34]. The significant reduction in grain protein content was observed at dosages ≥45 g ai ha⁻¹ in Zhangzagu 10 and at dosages ≥22.5 g ai ha⁻¹ in Jingu 21 (Fig 3). This result suggests that the hybrid Zhangzagu 10 has a stronger tolerance to TBM than the conventional Jingu 21. Additionally, TBM at dosages <45 g ai ha⁻¹ may not affect protein synthesis in Zhangzagu 10, possibly because of the ability of this variety to metabolize or breakdown the herbicide into non-phytotoxic products [35]. Lum et al. [24] reported that nicosulfuron (a sulfonylurea herbicide) at dosages <105 g ai ha⁻¹ caused little inhibition of protein synthesis in maize.

Grain mineral content is affected by a number of factors, such as varietal properties, soil mineral levels, efficiency of mineral uptake, and weed control methods [36, 37]. In China, Jingu 21 is a high-quality foxtail millet variety. Compared with Jingu 21, Zhangzagu 10 has higher grain yield but lower quality [38]. In this study, grain mineral concentrations were consistently higher in Jingu 21 than in Zhangzagu 10 without TBM treatment, reflecting the varietal properties. As shown in Table 1, there were significant differences among the soil mineral levels, which were all within the normal range. All treatments were conducted under similar soil conditions each year, therefore, the variances in the results were attributed to the herbicide. Treatment with TBM ≤22.5 g ai ha⁻¹ did not affect grain Fe, Mn, Cu, and Zn contents in Zhangzagu 10. However, grain mineral contents in Jingu 21 significantly decreased at dosages ≥11.25 g ai ha⁻¹. Gugala et al. [37, 39] reported that herbicides affect grain mineral concentrations. Zn, Mn, Fe, and Cu concentrations in foxtail millet grain decreased in response to increased TBM dosages, probably due to reduced root activity [40]. The herbicide tolerance of hybrid Zhangzagu 10 probably helped to protect the root against damage at dosages ≤22.5 g ai ha⁻¹. Future studies should investigate the effect of TBM on more genotypes of foxtail millet, and the molecular and biochemical mechanisms responsible for TBM tolerance.

The recommended dosage of TBM (22.5 g ai ha⁻¹) was not safe for the conventional variety Jingu 21, but relatively safe for the hybrid variety Zhangzagu 10. At this dosage, the herbicide significantly decreased grain yield and biomass, grain protein content, and grain Mn, Cu, Fe, and Zn concentrations in Jingu 21, but not in Zhangzagu 10. The hybrid variety Zhangzagu 10 showed a stronger tolerance to TBM than the conventional variety Jingu 21.