Antibody-based serological methods, including enzyme-linked immunosorbent assay (ELISA), dot immunobinding assay (DIBA) and Western blot, play an essential role in the detection of plant viruses. In general, polyclonal antibodies are the first choice for the detection of many plant viruses due to their ease of preparation; however, the coat proteins (CPs) of some viruses have the same or similar epitopes, and the polyclonal antibody of one virus may produce cross-reactions with other viruses of the same genus in routine serological detection (Mrkvová et al., 2022). Monoclonal antibodies (MAbs) have the advantage of higher specificity than polyclonal antibodies, although some MAbs can only detect partial isolates of a virus (Tian et al., 2014).

Potato virus Y (PVY) is the type species of the largest plant RNA virus genus Potyvirus and causes huge economic losses to the production of several solanaceous crops including potato, pepper and tobacco (Wylie et al., 2018). Based on the recognition of different resistance genes in potatoes, PVY isolates can be divided into three main strains including N, O and C. Three MAbs for PVY, MAbs 1128, 1129 and 1130, are widely used to determine the overall incidence of PVY. MAb1128 is claimed to detect PVY^N specifically and MAb1129 detects both PVY^O and PVY^C, whereas MAb1130 is believed to detect the above three strains of PVY. The minimal epitopes recognized by MAbs 1128, 1129 and 1130 are amino acids 25NLNKEK30, 16RPEQGSIQSNP26 and 5IDAGGS10, respectively, of PVY CP; however, 20.10%, 48.57% and 12.73% of the 369 PVY isolates analysed do not contain these epitopes (Tian et al., 2014), which will significantly influence the detection accuracy.

To establish an effective detection system covering all PVY isolates, 54 MAbs for PVY were prepared and used for the identification of their minimal recognition epitopes. First, we constructed the expression vector pEHISTEV-PVY-CP to express PVY CP in Escherichia coli cells. Then, we introduced deletion to the expression vector and obtained a series of deletion mutants, Δ2–15, Δ2–30, Δ16–45, Δ31–60, Δ46–75, Δ61–90, Δ61–255, Δ226–255, Δ241–267 and Δ256–267. The recognition epitope of a PVY Mab can be deduced according to its recognition with different CP mutants. After the first round of screening, the recognition epitopes of MAbs N1, M1, M2 and C1 were mapped to the region of amino acids 2–15, 31–45, 85–98 and 256–267 of PVY CP (Figure 1a; Figures S1 and S2). To elucidate the minimal recognition epitopes of each MAb, we further deleted the amino acids one by one from both N and C termini of the region identified above. The results showed that the minimal recognition epitopes for N1, M1, M2 and C1 were mapped to 4TIDAGGSTK12, 37GTSGTHTVP45, 89QFDTWYE95 and 261LLGVKN266, respectively, of PVY CP (Figure 1b; Figure S3). Substitutions of the epitope amino acids to alanine abolished recognition of the corresponding MAbs with PVY CP (Figure S4). The recognition epitopes of N1, M1 and C1 were predicted to be located in random coils and that of M2 was located in an α-helix (Figure S5). The isoforms and subclasses of N1, M1 and M2 were identified as IgG1 with the kappa light chain, while that of Mab C1 was IgG2b with the kappa light chain (Table S1). The titers of these four MAbs were all 1 : 243 000 as detected with an indirect ELISA (Table S1).

To evaluate the performance of these four MAbs, we synthesized the CP of 10 PVY isolates deposited in the GenBank. These 10 PVY isolates had one to four amino acid differences from the epitope recognized by MAb N1 but contained the epitope recognized by M1 and M2. Isolate PVY8 had one amino acid difference with the epitopes recognized by MAb C1 (Figure S6). The results showed that MAb N1 could detect six isolates excluding PVY1, PVY2, PVY5 or PVY6; M1 and M2 could detect all these 10 isolates, while MAb C1 could detect nine isolates except PVY8 (Figure 1c).

We then tested the specificity of these four MAbs by using DIBA, ELISA and Western blot. These four MAbs showed positive reactions only with PVY, not with other viruses tested, including seven viruses of the genus Potyvirus, one virus of the genus Potexvirus, two viruses of the genus Carlavirus and four viruses of the genus Tobamovirus (Figure 1d–f). Sequence analysis of eight potyviruses showed that the epitopes recognized by the four MAbs were PVY-specific (Figure S7).

MAbs N1, M1, M2 and C1 could detect PVY from leaf extracts diluted for 10 240, 5120, 640 and 5120 times, respectively, which meant these four MAbs had high sensitivity (Figure 1g).

There are 1885 full-length PVY CP sequences available in the GenBank till 12 April 2022. We numbered these PVY isolates from 1 to 1885. A total of 739, 647 and 1639 isolates, respectively, contained the recognition epitopes of MAbs 1128, 1129 and 1130 (Tables S2 and S3). A total of 121 isolates did not contain the epitopes recognized by either of these three MAbs (Tables S2 and S3); 694, 1853, 1843 and 1851 isolates contained the recognition epitope of N1, M1, M2 and C1, respectively (Tables S2 and S4). The sequence conservation analysis of 1885 PVY CPs also showed that the recognition epitopes of M1, M2 and C1 were more conserved than that of N1 and MAb1130 (Figure S8). The number of PVY isolates that did not contain the recognition epitopes of N1, M1, M2 and C1 was 1191, 32, 42 and 34, respectively (Table S2). Because these four MAbs had different and conserved recognition epitopes, a mixture of two or three MAbs will detect more PVY isolates. The mixture of M1 and M2, M1 and C1 or M2 and C1 could separately detect 1885, 1883 or 1876 PVY isolates (Figure 1h and Table S4). MAb N1 had a narrow coverage and a mixture of N1 with M1, M2 or C1 could separately detect 1867, 1851 or 1856 isolates (Table S4), which were fewer than that detected by other MAb mixtures; therefore, N1 was not included for further consideration. The specificity analysis also showed that mixtures of M1 and M2, M1 and C1 and M2 and C1 were positive only for PVY (Figure S9).

To sum up, we have mapped the recognition epitopes of four newly prepared MAbs and shown that the cocktail of M1 and M2 can accurately detect all the 1885 PVY isolates deposited in the GenBank. Besides, by analysing the amino acid sequence of its CP, we can determine whether a new PVY isolate can be detected with these four MAbs. Therefore, the cocktail of M1 and M2 will play a great role in the future quarantine and routine detection of PVY. This study has established a novel approach to improve the detection coverage of MAbs by combining MAbs that recognize different and conserved epitopes. This method can be further applied to the precise detection of other plant viruses and even animal viruses including the current pandemic and rapidly mutating SARS-CoV-2.

基于抗体的血清学方法，包括酶联免疫吸附测定（ELISA）、斑点免疫结合测定（DIBA）和蛋白质印迹，在植物病毒的检测中发挥着重要作用。一般来说，多克隆抗体因其易于制备而成为许多植物病毒检测的首选；然而，某些病毒的外壳蛋白（CP）具有相同或相似的表位，一种病毒的多克隆抗体在常规血清学检测中可能会与同属的其他病毒产生交叉反应（Mrkvová et al. , 2022  ）。尽管有些MAb只能检测病毒的部分分离株，但单克隆抗体（MAb）具有比多克隆抗体更高特异性的优点（Tian等，  2014）。

马铃薯病毒Y（PVY）是最大的植物RNA病毒马铃薯病毒属的模式种，给马铃薯、辣椒和烟草等多种茄科作物的生产造成巨大的经济损失（Wylie等，  2018）。基于对马铃薯中不同抗性基因的识别，PVY分离株可分为三个主要菌株，包括N、O和C。PVY的三种MAb，MAb 1128、1129和1130，被广泛用于确定PVY的总体发病率。MAb1128 据称可特异性检测 PVY ^N，MAb1129 可检测 PVY ^O和 PVY ^C，而 MAb1130 被认为可检测上述三种 PVY 菌株。MAb 1128、1129和1130识别的最小表位分别是PVY CP的氨基酸25NLNKEK30、16RPEQGSIQSNP26和5IDAGGS10；然而，分析的369个PVY分离株中，有20.10%、48.57%和12.73%不包含这些表位(Tian et al .,  2014 )，这将显着影响检测准确性。

为了建立覆盖所有PVY分离株的有效检测系统，制备了54种PVY单克隆抗体并用于鉴定其最小识别表位。首先，我们构建了表达载体pEHISTEV-PVY-CP以在大肠杆菌细胞中表达PVY CP。然后，我们在表达载体中引入缺失，得到了一系列缺失突变体：Δ2–15、Δ2–30、Δ16–45、Δ31–60、Δ46–75、Δ61–90、Δ61–255、Δ226–255、Δ241 –267 和 Δ256–267。PVY Mab的识别表位可以根据其与不同CP突变体的识别来推断。经过第一轮筛选后，MAb N1、M1、M2 和 C1 的识别表位被定位到 PVY CP 的氨基酸 2-15、31-45、85-98 和 256-267 区域（图 1a；图 1a）。 S1 和 S2)。为了阐明每种 MAb 的最小识别表位，我们进一步从上述确定的区域的 N 和 C 末端一一删除了氨基酸。结果显示，N1、M1、M2 和 C1 的最小识别表位分别映射到 PVY CP 的 4TIDAGGSTK12、37GTSGTHTVP45、89QFDTWYE95 和 261LLGVKN266（图 1b；图 S3）。将表位氨基酸替换为丙氨酸，消除了 PVY CP 对相应 MAb 的识别（图 S4）。N1、M1 和 C1 的识别表位预计位于随机卷曲中，M2 的识别表位位于 α 螺旋中（图 S5）。N1、M1 和 M2 的亚型和亚类被鉴定为具有 kappa 轻链的 IgG1，而 Mab C1 的亚型和亚类被鉴定为具有 kappa 轻链的 IgG2b（表 S1）。通过间接ELISA检测，这四种MAb的效价均为1:243 000（表S1）。

为了评估这四种 MAb 的性能，我们合成了 GenBank 中存放的 10 个 PVY 分离株的 CP。这 10 个 PVY 分离株与 MAb N1 识别的表位有 1 到 4 个氨基酸差异，但包含 M1 和 M2 识别的表位。分离株 PVY8 与 MAb C1 识别的表位有一个氨基酸差异（图 S6）。结果表明，MAb N1可以检测除PVY1、PVY2、PVY5和PVY6之外的6个分离株；M1和M2可以检测所有这10个分离株，而MAb C1可以检测除PVY8之外的9个分离株（图1c）。

然后我们使用 DIBA、ELISA 和 Western blot 测试了这四种 MAb 的特异性。这四种 MAb 仅对 PVY 表现出阳性反应，对其他测试的病毒没有表现出阳性反应，包括七种马铃薯病毒属病毒、一种马铃薯病毒属病毒、两种卡拉病毒属病毒和四种烟草花叶病毒属病毒（图 1d-f） 。八种马铃薯Y病毒的序列分析表明，四种MAb识别的表位是PVY特异性的（图S7）。

单克隆抗体N1、M1、M2和C1可以分别从稀释10 240、5120、640和5120倍的叶提取物中检测PVY，这意味着这四种单克隆抗体具有较高的灵敏度（图1g）。

截至 2022 年 4 月 12 日，GenBank 中有 1885 个全长 PVY CP 序列。我们将这些 PVY 分离株编号为 1 至 1885。总共 739 个、647 个和 1639 个分离株分别包含 MAb 1128、1129 和 1129 的识别表位。 1130（表S2和S3）。总共 121 个分离株不包含这三种 MAb 所识别的表位（表 S2 和 S3）；694、1853、1843和1851分离株分别含有N1、M1、M2和C1的识别表位（表S2和S4）。对1885个PVY CP的序列保守性分析也表明，M1、M2和C1的识别表位比N1和MAb1130的识别表位更加保守（图S8）。不包含N1、M1、M2和C1识别表位的PVY分离株的数量分别为1191、32、42和34（表S2）。由于这四种 MAb 具有不同且保守的识别表位，因此两种或三种 MAb 的混合物将检测更多的 PVY 分离株。M1和M2、M1和C1或M2和C1的混合物可以分别检测1885、1883或1876 PVY分离株（图1h和表S4）。MAb N1覆盖范围窄，N1与M1、M2或C1的混合物可分别检测到1867、1851或1856个分离株（表S4），比其他MAb混合物检测到的要少；因此，N1 没有被进一步考虑。特异性分析还表明，M1 和 M2、M1 和 C1 以及 M2 和 C1 的混合物仅对 PVY 呈阳性（图 S9）。

综上所述，我们绘制了四种新制备的 MAb 的识别表位，并表明 M1 和 M2 的混合物可以准确检测 GenBank 中保藏的所有 1885 个 PVY 分离株。此外，通过分析其CP的氨基酸序列，我们可以确定这四种MAb是否可以检测到新的PVY分离株。因此，M1和M2的混合物将在今后PVY的检疫和常规检测中发挥巨大作用。这项研究建立了一种新方法，通过结合识别不同和保守表位的单克隆抗体来提高单克隆抗体的检测覆盖率。该方法可以进一步应用于其他植物病毒甚至动物病毒的精确检测，包括当前大流行和快速变异的SARS-CoV-2。