Weed species have increasingly emerged with resistance against previously effective herbicides, such as glyphosate and inhibitors of acetyl coenzyme A carboxylase (ACCase) and acetolactate synthase (ALS) (Heap, 2024). Owing to its novel mode of action, 4-hydroxyphenylpyruvate dioxygenase (HPPD)-inhibitors are effective in controlling herbicide-resistant weeds and recently attracted much attention. Resistance to HPPD-inhibitors has been slow to evolve in weeds, and only a few cases of resistant events have been reported and most of these are associated with enhanced herbicide metabolism (Heap, 2024; Lu et al., 2023). Since resistant sites in the entire target gene are largely unknown, we in vivo mutagenized the HPPD gene in Arabidopsis and rice using base editing libraries to uncover potential target-site resistant mutations.

Arabidopsis are highly sensitive to mesotrione. After large-scale transformation of base editing pools (Figure S1), we found one T1 seedling growing normally in medium containing 100 nM mesotrione (Figure S2a). Genotyping showed that a heterozygous T-to-C substitution occurred, causing the Y342H mutation in the HPPD protein sequence (Figure S2b, c). Homozygous progenies of HPPD^Y342H (Figure S2d) were tolerant up to 200 nM mesotrione and 50 nM isoxaflutole in medium (Figures 1a and S3a) and had a much higher survival rate than wild type (WT) when sprayed with ≥1 μM mesotrione or isoxaflutole in pots (Figure 1b, c). AtHPPD^Y342H mutants exhibited similar phenotype and showed no significant differences in plant height and seed yield with the WT plants (Figure S4).

We transformed the AtHPPD^WT and AtHPPD^Y342H genes under the native promoter into Arabidopsis. As expected, the AtHPPD^Y342H transgenic lines exhibited higher mesotrione tolerance than the AtHPPD^WT transgenic plants (Figure S3a), even though the transgenic AtHPPD^Y342H had equal or lower expression levels compared with AtHPPD^WT (Figure S3b). Together, these results showed that the AtHPPD^Y342H mutation significantly increases the target-site resistance to HPPD-inhibitors.

We next determined whether this mutation also increased tolerance in soybean, a crop highly sensitive to HPPD-inhibitors. Alignment of protein sequences between Arabidopsis and soybean showed that the Y388 of GmHPPD corresponds to the Y342 of AtHPPD. We generated the GmHPPD^Y388H lines by base editing, and the resulting homozygous and T-DNA-free GmHPPD^Y388H offsprings in the T3 generation were treated with mesotrione and isoxaflutole. Our results showed that WT plants were suppressed by 2.9 μM mesotrione, whereas the GmHPPD^Y388H mutants were tolerant up to 58.9 μM of mesotrione (Figure 1d). WT and GmHPPD^Y388H plants were seriously affected by 2.8 μM and 55.7 μM of isoxaflutole, respectively (Figure 1e). The GmHPPD^Y388H mutants were estimated to increase 20-fold resistance to HPPD-inhibitors than WT plants. Importantly, consistent with the results obtained from Arabidopsis, the seed amount and yield did not significantly differ between GmHPPD^Y388H mutants and WT plants (Figure 1f), indicating that there was no fitness cost for the herbicide tolerance.

We then overexpressed the GmHPPD^Y388H in soybean to further improve herbicide resistance (Figure S5). Owing to the combined effect of increased expression level and target-site mutation, transgenic offsprings showed a 100% survival rate albeit with ~10% yield reduction under 590 μM mesotrione or 278 μM isoxaflutole treatment (Figure 1g), which are in the range of concentrations required for weed control in the field (300–600 μM).

Unlike Arabidopsis and soybean, Japonica rice is resistant to triketones due to an endogenous HIS1 gene that detoxifies these herbicides (Maeda et al., 2019). To further increase their resistance by target-site mutations, we pool-edited the HPPD gene in Japonica rice using base editors. Rice exhibits high sensitivity to HPPD-inhibitors at seed germination stage, so we screened for resistant mutants using T1 seeds. This strategy also enabled us to simultaneously test different HPPD-inhibitors (Figure S6). Hundreds of amino acid mutations were identified from randomly examined transgenic lines (Figure S7; Table S1). Although most of these mutants remained sensitive, several mutants exhibited increased resistance to one or more of the tested herbicides (Data S1). The N338D mutation resulted in a slight tolerance to mesotrione, tembotrione and isoxaflutole at the germination stage, but only isoxaflutole-resistance was confirmed at the seedling stage (Figure 1i, j). The P336L mutation increased resistance to tembotrione but not mesotrione or isoxaflutole (Figure 1i, j). Both the N338D and P336L mutants showed decreased plant height and/or lower seed setting rate (Figure S8), indicating that there were fitness cost for the herbicide tolerance. The Y339H mutation, which corresponds to the AtHPPD-Y342H, did not improve tolerance to any of the tested HPPD-inhibitors (Data S1). Since the HIS1 gene makes a major contribution to the resistance, it is not surprising that mutations in OsHPPD only slightly increase the overall tolerance to HPPD-inhibitors.

Very recently, the HPPD coding sequences of cotton and Arabidopsis were also evolved in vitro or in vivo (Qian and Shi, 2024; Wang et al., 2024). Most of the resistant mutations are located within or near the helix gate, which may alter the binding site conformation and affect herbicide accessibility. Nevertheless, the increased resistance endowed by these mutations alone might not be sufficient for application in the field, which might explain why almost no target-site resistant weeds have been reported thus far. With the long-term application of HPPD herbicides, emerging resistant weeds likely have enhanced metabolism of the specific types of herbicides, which implies that rotating use of different types of HPPD herbicides may help to delay the emergence of resistant weeds.

Recommendable strategies for improving crop tolerance include combined the target-site mutations with a higher background-tolerance HPPD gene such as that of maize (Siehl et al., 2014), enhanced expression by ectopically expressing HPPD gene in plastids (Dufourmantel et al., 2007) or directly knocking-up the endogenous HPPD gene (Lu et al., 2021), and introduced the HPPD-inhibitor metabolism gene HIS1 (Maeda et al., 2019). Variations in 3’-UTR of the OsHPPD, which may affect the regulation of mRNA stability or protein translation, have also been reported to improve resistance (Wu et al., 2023).

越来越多的杂草物种对以前有效的除草剂产生了抗药性，例如草甘膦、乙酰辅酶 A 羧化酶 (ACCase) 和乙酰乳酸合酶 (ALS) 抑制剂 (Heap, 2024 )。 4-羟基苯基丙酮酸双加氧酶（HPPD）抑制剂由于其新颖的作用方式，可有效控制除草剂抗性杂草，近年来引起了广泛关注。杂草对 HPPD 抑制剂的抗性进化缓慢，仅报道了少数抗性事件，其中大多数与除草剂代谢增强有关（Heap， 2024 ；Lu等， 2023 ）。由于整个靶基因中的抗性位点在很大程度上是未知的，我们使用碱基编辑文库对拟南芥和水稻中的HPPD基因进行体内诱变，以发现潜在的靶位点抗性突变。

拟南芥对甲基磺草酮高度敏感。对碱基编辑池进行大规模改造后（图S1），我们发现一株T1幼苗在含有100 nM甲基磺草酮的培养基中正常生长（图S2a）。基因分型显示，发生了杂合的 T 到 C 替换，导致 HPPD 蛋白序列中出现 Y342H 突变（图 S2b、c）。 HPPD ^Y342H的纯合后代（图 S2d）在培养基中耐受高达 200 nM 甲基磺草酮和 50 nM 异恶唑草酮（图 1a 和 S3a），并且当喷洒 ≥1 μM 甲基磺草酮或异恶唑草酮时，其存活率比野生型 (WT) 高得多在盆中（图1b，c）。 AtHPPD ^Y342H突变体表现出相似的表型，并且在植物高度和种子产量方面与 WT 植物没有显着差异（图 S4）。

我们将天然启动子下的AtHPPD ^WT和AtHPPD ^Y342H基因转化到拟南芥中。正如预期的那样， AtHPPD ^Y342H转基因株系表现出比AtHPPD ^WT转基因植物更高的甲基磺草酮耐受性（图 S3a），尽管转基因AtHPPD ^Y342H与AtHPPD ^WT相比具有相同或更低的表达水平（图 S3b）。总之，这些结果表明 AtHPPD ^Y342H突变显着增加了靶位点对 HPPD 抑制剂的抗性。

接下来我们确定这种突变是否也增加了大豆（一种对 HPPD 抑制剂高度敏感的作物）的耐受性。拟南芥和大豆之间的蛋白质序列比对表明，GmHPPD 的 Y388 对应于 AtHPPD 的 Y342。我们通过碱基编辑生成了GmHPPD ^Y388H系，并用硝磺草酮和异恶唑草酮处理 T3 代中所得的纯合且不含 T-DNA 的GmHPPD ^Y388H后代。我们的结果表明，WT 植物被 2.9 μM 甲基磺草酮抑制，而GmHPPD ^Y388H突变体能够耐受高达 58.9 μM 甲基磺草酮（图 1d）。 WT 和GmHPPD ^Y388H植物分别受到 2.8 μM 和 55.7 μM 异恶唑草酮的严重影响（图 1e）。据估计， GmHPPD ^Y388H突变体对 HPPD 抑制剂的抗性比 WT 植物增加了 20 倍。重要的是，与拟南芥获得的结果一致， GmHPPD ^Y388H突变体和 WT 植物之间的种子数量和产量没有显着差异（图 1f），表明除草剂耐受性没有适应性成本。

然后，我们在大豆中过表达GmHPPD ^Y388H，以进一步提高除草剂抗性（图 S5）。由于表达水平增加和靶位点突变的综合作用，转基因后代在 590 μM 甲基磺草酮或 278 μM 异恶唑草酮处理下显示出 100% 的存活率，尽管产量降低了约 10%（图 1g），其范围在田间杂草控制所需的浓度（300–600 μM）。

与拟南芥和大豆不同，粳稻对三酮具有抗性，因为内源性HIS1基因可以解毒这些除草剂（Maeda等， 2019 ）。为了通过靶位点突变进一步提高其抗性，我们使用碱基编辑器对粳稻中的HPPD基因进行了联合编辑。水稻在种子萌发阶段对 HPPD 抑制剂表现出高度敏感性，因此我们使用 T1 种子筛选抗性突变体。该策略还使我们能够同时测试不同的 HPPD 抑制剂（图 S6）。从随机检查的转基因品系中鉴定出数百个氨基酸突变（图 S7；表 S1）。尽管大多数突变体仍然敏感，但一些突变体对一种或多种测试的除草剂表现出增强的抗性（数据 S1）。 N338D突变导致发芽阶段对甲基磺草酮、环磺草酮和异恶唑草酮有轻微耐受性，但在苗期仅证实对异恶草酮具有抗性（图1i，j）。 P336L 突变增加了对环磺草酮的耐药性，但不增加硝磺草酮或异恶唑草酮的耐药性（图 1i、j）。 N338D 和 P336L 突变体均表现出株高降低和/或结实率降低（图 S8），表明除草剂耐受性存在适应性成本。对应于 AtHPPD-Y342H 的 Y339H 突变并未提高对任何测试的 HPPD 抑制剂的耐受性（数据 S1）。由于HIS1基因对耐药性做出了重大贡献，因此OsHPPD突变仅略微增加对 HPPD 抑制剂的总体耐受性也就不足为奇了。

最近，棉花和拟南芥的HPPD编码序列也在体外或体内进化（Qian 和 Shi， 2024 ；Wang等， 2024 ）。大多数抗性突变位于螺旋门内或附近，这可能会改变结合位点构象并影响除草剂的可及性。然而，仅由这些突变赋予的抗性增加可能不足以在田间应用，这可能解释了为什么迄今为止几乎没有报道靶点抗性杂草。随着HPPD除草剂的长期使用，新出现的抗性杂草可能会增强特定类型除草剂的代谢，这意味着轮换使用不同类型的HPPD除草剂可能有助于延缓抗性杂草的出现。

提高作物耐受性的推荐策略包括将靶位点突变与更高背景耐受性的HPPD基因（例如玉米）相结合（Siehl等人， 2014 ），通过在质体中异位表达HPPD基因来增强表达（Dufourmantel等人，2014）。 2007 ）或直接敲除内源性HPPD基因（Lu et al ., 2021 ），并引入HPPD抑制剂代谢基因HIS1 （Maeda et al ., 2019 ）。据报道， OsHPPD 3'-UTR 的变化可能会影响 mRNA 稳定性或蛋白质翻译的调节，也可改善耐药性 (Wu et al ., 2023 )。