Light traps make important contributions to integrated pest management strategies by helping to reduce the impact of humanity on the environment. There is considerable interest in developing light sources that utilize less energy and that are more eco-friendly. Insects can perceive wavelengths of light that range from 650 to 300 nanometres. Recently, ultraviolet lights have been used as an effective tool for attracting insect pests (Cook et al., 2007). When GFP is excited at 475 nm light, the subsequent emission maximum is 503 nm (Verkhusha and Lukyanov, 2004). It was suggested the fluorescence from chemical compounds attracts insects to carnivorous plants (Kurup et al., 2013). To date, no reports have tested whether fluorescence influences insect behaviour. There are no reports on whether it is feasible to use the fluorescence emitted by GFP to develop pest control technology. Spodoptera litura (Fabricius) has phototactic responses and damages lots of vegetable and field crops (Yang et al., 2012). In our present study, firstly varied transgenic tobaccos were developed by chloroplast and nuclear transformation (Figure S1) and fluorescein isothiocyanate (FITC) treated tobacco was also integrated. Finally, we performed a series of experiments to evaluate interactions between insects and various types of tobacco.

We produced different fluorescent tobacco lines. We observed that substantial GFP fluorescence was emitted from transplastomic tobacco lines possibly because the large number of chloroplasts in leaf cells facilitate high-level expression of exogenous protein. Indeed, most leaves emitted green light and single cells strongly fluoresced (Figure 1d–f and d'–f'), which was not the case for the transgenic tobacco plants that harboured a nuclear-localized transgene that expresses GFP (Figure 1a–c and a'–c'). The wavelengths of light emitted from leaves of the transplastomic tobacco that were excited at 475 nm were different from the leaves of wild-type tobacco (Figure 1f,f').

We found that S. litura larvae had a significant preference for transplastomic tobacco leaves that were exposed to continuous light conditions (Figure 1g–i; 6 h: χ² = 8.53, P = 0.003; 12 h: χ² = 11.84, P = 0.001; 24 h: χ² = 11.84, P = 0.001; Chi-squared test) and that located in the darkness (Figure 1j; 6 h: χ² = 4.11, P = 0.043;12 h: χ² = 6.36, P = 0.012; 24 h: χ² = 10.67, P = 0.001; Chi-squared test). A video record also showed that the larvae had a feeding preference for transplastomic tobacco leaves (Video S1). In addition, all of the S. litura larvae were trapped and died after they were released into petri dishes for 15 min (Video S2). However, the larvae showed no feeding preference for the nuclear transgenic tobacco leaves with the exception of the 6 h time point after they were transferred to the dark (χ² = 5.06, P = 0.025; Chi-squared test) (Figure 1p,q). The larvae also had no feeding preference for tobacco leaves coated with FITC (Figures 1u and S4). The transplastomic tobacco plants were severely damaged by the S. litura larvae during a 48 h feeding experiment. Indeed, these plants were almost completely defoliated after 72 h (Figure 1v–x).

In the Y-tube olfactometer test (Figures 1k and S3), we observed no chemotaxis when the S. litura larvae were exposed only to fresh air from both arms (Figure 1k). When either the transplastomic or control tobacco odour sources were utilized in one arm of the Y-tube olfactometer, we observed significant chemotaxis among the larvae to the tobacco plant (pGFP: χ² = 3.99, P = 0.046; Control: χ² = 4.44, P = 0.035). However, we observed no significant difference in chemotaxis when we used the transplastomic tobacco in one arm and wild-type tobacco in the other arm (Figure 1k). These data indicate that there is no significant chemotactic difference between the transplastomic tobacco and the non-transgenic tobacco.

In the choice test, S. litura females oviposited a greater mass of eggs on the transplastomic tobacco plants relative to the control plants (Figure 1l–n; χ² = 40.50, P = 0.006). Simultaneously, a larger number of eggs were laid on the transplastomic tobacco plants than on the non-transgenic control plants (Figure 1o; t = 5.97, P = 0.002). In contrast, no significant oviposition preference was observed for the transgenic tobacco harbouring a nuclear-localized transgene that expresses GFP relative to the control plants (Figure 1r–t).

For the investigation in plastic pots, the transplastomic tobacco plants were fed to lepidopterous insect pests in a greenhouse. The damage was severe after 15 days (Figure 1y). At the same time, more than 90% of the transplastomic tobacco plants were severely damaged (Figures 1z and S2a). However, the adjacent non-transgenic control plants were damaged at a rate that was <2% (Figure 1z). In addition to S. litura, pests collected from the damaged transplastomic tobacco plants in the greenhouse included a small number of the Spodoptera exigua Hübner (Lepidoptera: Noctuidae) (Figure S2b,c). GFP fluorescence present in the frass and midguts of the collected lepidopterous larvae (Figure S2d,e) indicated that those pests mainly fed on the transplastomic tobacco.

Although the crystal jelly uses fluorescence to attract prey (Steven et al., 2015), no one knows whether GFP provides other biological or ecological functions to jellyfish. Our findings support the idea that fluorescence from jellyfish can attract insects when they accumulate in plants. We found that S. litura were attracted to transplastomic tobacco that emitted green fluorescence at approximately 500 nm. Thus, combining this technology with pesticides could lead to an effective push-pull strategy to manage and kill S. litura. Particular amino acid substitutions in GFP can alter the excitation and emission maxima of GFP (Verkhusha and Lukyanov, 2004). Photoresponse is widespread among insects (Casper et al., 2021). Thus, we can engineer fluorescent proteins that are compatible with the spectral sensitivity of specific insect photoreceptors and lure particular insects to particular plants. Recently bioluminescent plants were generated (Mitiouchkina et al., 2020). These approaches will lead to the development of bioluminescent and fluorescent plants that lure and trap insect pests. In summary, our findings indicate that light emitted from plants as either bioluminescence or fluorescence can contribute to ecological interactions between animals and plants and provide a novel means for monitoring and managing insects and pests. Certainly, all ethical and environmental concerns should be resolved before using this sort of biotechnology outside of the laboratory.

光阱通过帮助减少人类对环境的影响，为综合虫害管理战略做出了重要贡献。人们对开发使用更少能源且更环保的光源产生了相当大的兴趣。昆虫可以感知 650 到 300 纳米范围内的光波长。最近，紫外线已被用作吸引害虫的有效工具（Cook等人，  2007 年）。当 GFP 在 475 nm 光下被激发时，随后的发射最大值为 503 nm（Verkhusha 和 Lukyanov，  2004 年）。有人提出化学化合物的荧光将昆虫吸引到食肉植物上（Kurup等，  2013）。迄今为止，没有报告测试荧光是否影响昆虫行为。没有关于利用 GFP 发出的荧光开发害虫防治技术是否可行的报道。Spodoptera litura (Fabricius) 具有趋光反应并损害大量蔬菜和大田作物 (Yang et al .,  2012 )。在我们目前的研究中，首先通过叶绿体和核转化开发了多种转基因烟草（图S1），并且还整合了异硫氰酸荧光素（FITC）处理的烟草。最后，我们进行了一系列实验来评估昆虫与各种烟草之间的相互作用。

我们生产了不同的荧光烟草线。我们观察到大量的 GFP 荧光从转质体烟草系中发出，这可能是因为叶细胞中大量的叶绿体促进了外源蛋白的高水平表达。事实上，大多数叶子发出绿光，单细胞发出强烈的荧光（图 1d-f 和 d'-f'），但对于含有表达 GFP 的核定位转基因的转基因烟草植物而言，情况并非如此（图 1a-c和a'-c'）。在 475 nm 激发的转质体烟草叶子发出的光波长与野生型烟草的叶子不同（图 1f，f'）。

我们发现S. litura幼虫对暴露于连续光照条件下的转质体烟叶有显着偏好（图 1g-i；6 小时：χ ²  = 8.53，P  = 0.003；12 小时：χ ²  = 11.84，P  = 0.001；24 h：χ ²  = 11.84，P  = 0.001；卡方检验）和位于黑暗中的（图 1j；6 h：χ ²  = 4.11，P  = 0.043；12 h：χ ²  = 6.36，P  = 0.012；24 小时：χ ²  = 10.67，P = 0.001; 卡方检验）。视频记录还显示，幼虫对转基因烟叶有喂养偏好（视频 S1）。此外，所有的S. litura幼虫在被释放到培养皿中 15 分钟后被困并死亡（视频 S2）。然而，幼虫对核转基因烟叶没有表现出取食偏好，除了转移到黑暗后的6小时时间点（^χ2 =  5.06，P  = 0.025；卡方检验）（图1p，q ）。幼虫对涂有 FITC 的烟叶也没有摄食偏好（图 1u 和 S4）。转质体烟草植物受到S. litura的严重破坏幼虫在 48 小时的喂养实验中。事实上，这些植物在 72 小时后几乎完全落叶（图 1v-x）。

在 Y 型管嗅觉仪测试中（图 1k 和 S3），当斜纹夜蛾幼虫仅暴露于来自双臂的新鲜空气时，我们没有观察到趋化性（图 1k）。当在 Y 型管嗅觉仪的一个臂中使用转质体或对照烟草气味源时，我们观察到幼虫对烟草植物的显着趋化性（pGFP：χ ²  = 3.99，P  = 0.046；对照：χ ²  = 4.44 ,磷 = 0.035)。然而，当我们在一只手臂中使用转基因烟草而在另一只手臂中使用野生型烟草时，我们观察到趋化性没有显着差异（图 1k）。这些数据表明转基因烟草和非转基因烟草之间没有显着的趋化性差异。

在选择测试中，相对于对照植物，斜纹夜蛾雌性在转质体烟草植物上产卵量更大（图 1l-n；^χ2 =  40.50，P  = 0.006）。同时，与非转基因对照植物相比，在转基因烟草植物上产卵的数量更多（图 1o；t  = 5.97，P  = 0.002）。相反，对于含有相对于对照植物表达GFP的核定位转基因的转基因烟草，没有观察到显着的产卵偏好（图1r-t）。

为了在塑料盆中进行调查，将转基因烟草植物喂给温室中的鳞翅目害虫。15 天后损坏严重（图 1y）。同时，超过 90% 的转基因烟草植物受到严重破坏（图 1z 和 S2a）。然而，相邻的非转基因对照植物以 <2% 的速率受损（图 1z）。除了S. litura，从温室中受损的转基因烟草植物中收集的害虫还包括少数Spodoptera exigua Hübner（鳞翅目：夜蛾科）（图 S2b，c）。收集到的鳞翅目幼虫的粗肠和中肠中存在的 GFP 荧光（图 S2d，e）表明这些害虫主要以转基因烟草为食。

虽然水晶果冻使用荧光来吸引猎物（Steven et al .,  2015），但没有人知道 GFP 是否为水母提供了其他生物或生态功能。我们的研究结果支持这样一种观点，即水母发出的荧光可以在昆虫在植物中积累时吸引它们。我们发现S. litura被在大约 500 nm 处发出绿色荧光的转基因烟草所吸引。因此，将这项技术与杀虫剂相结合，可以形成一种有效的推拉策略来管理和杀死斜纹夜蛾。GFP 中的特定氨基酸取代可以改变 GFP 的激发和发射最大值（Verkhusha 和 Lukyanov，  2004）。光反应在昆虫中很普遍（Casper等人，  2021 年）。因此，我们可以设计与特定昆虫感光器的光谱敏感性兼容的荧光蛋白，并将特定昆虫引诱到特定植物。最近产生了生物发光植物（Mitiouchkina等，  2020）。这些方法将导致引诱和诱捕害虫的生物发光和荧光植物的发展。总之，我们的研究结果表明，植物发出的光作为生物发光或荧光可以促进动植物之间的生态相互作用，并为监测和管理昆虫和害虫提供一种新的手段。当然，在实验室外使用这种生物技术之前，应该解决所有伦理和环境问题。