Since the promotion of genetically modified (GM) crops in the United States in 1996, the global agricultural production landscape has undergone a profound transformation. Insect-resistant and herbicide-tolerant GM corn, soybeans, canola, and cotton have rapidly spread across North and South America, significantly boosting crop yields and production efficiency. This has not only reduced pesticide use but also lowered production costs and alleviated environmental pressure.

The Controversy and Challenges of the "Refuge" Strategy

To effectively delay the evolution of resistance in target pests, multinational corporations led by Monsanto introduced the "High-Dose/Refuge" strategy. "High-dose" aimed to achieve up to 99.99% efficacy against the naturally susceptible population of target pests, thoroughly eliminating potential resistant individuals. "Refuge" required planting a certain proportion of non-GM crops in GM crop fields to provide a breeding space for susceptible pests, thereby diluting the frequency of resistant genes in the field. However, the actual effectiveness of this strategy has been a subject of ongoing debate.

The "refuge" strategy has, to some extent, become a barrier to market entry. Large multinational corporations have established technological moats through patent technology and complex approval processes, consolidating their monopoly in the market. This makes it difficult for small R&D institutions or emerging companies to bear the high patent licensing fees and lengthy approval cycles, excluding them from the mainstream market. This structural barrier not only hinders the rapid and diversified development of agricultural technology but also deepens farmers' dependence on a few companies. To achieve true technological innovation and fair competition, it is urgent to reconstruct an open and equitable ecosystem, break patent blockades and approval monopolies, and provide space for innovators to thrive.

Furthermore, the correlation between target pest resistance and the "refuge" strategy may not be as strong as anticipated. In the United States, for instance, despite good implementation of the refuge strategy for the European corn borer, this pest did not develop resistance, and its population density became very low. However, research suggests that this was less related to the refuge and more to the presence of two different receptors for the Cry1Ab protein in the corn borer's gut, making it difficult for it to develop single-gene resistance. Conversely, under the same refuge strategy, corn rootworms rapidly developed significant field resistance to various Cry proteins, with current effectiveness largely relying on the later-marketed and more active Cry34/Cry35 dual-protein synergistic action. This indicates that high-dose and multiple mechanisms are key to delaying resistance development, while the role of refuges might be quite limited.

Similarly, the rapid development of resistance to Cry1F-containing GM corn by fall armyworm in South America also appears to have little to do with the presence of refuges. The actual reason lies in the significantly lower expression level of Cry1F in corn plants compared to prokaryotic bacteria, failing to reach the "high-dose" standard required to effectively kill pests. When the insecticidal protein expressed by the crop is insufficient, some "robust individuals" survive and gradually adapt, thereby selecting for resistant populations.

Moreover, pest resistance evolution is a complex process driven by multiple factors, including crop layout, pest population dynamics, and gene expression environment. Relying solely on the "refuge" strategy is insufficient to address the complexities of field realities. Especially in China, fragmented land and diversified production, intercropping of various crops, cultivation of vegetables and fruit trees, as well as windbreak forests, green belts, and trees and bushes around villages, inherently provide extensive natural "refuges." Large-scale continuous planting of corn in China is limited, mainly concentrated in the Heilongjiang reclamation area, which is located in a "cold region" with limited insect pest incidence. The primary pest there is the corn borer, which can be effectively controlled by the Cry1Ab gene, preventing both damage and resistance development.

Scientific and Technological Innovation: The Fundamental Solution for Resistance Management

Fortunately, continuous scientific and technological innovation offers new hope for resistance management. With the rapid advancement of artificial intelligence (AI) technology, discovering new insecticidal proteins and directed evolution of existing ones has become easier and more efficient, effectively preventing the problem of lacking effective traits when target pests develop resistance. The discovery and functional optimization of novel insecticidal proteins are accelerating, with greatly improved efficiency. This benefits from AlphaFold 3's accurate prediction of protein structures from known amino acid sequences, and the recreation or remodeling of amino acid sequences of known functional and structural proteins to generate a large number of structurally similar candidate sequences. After cluster analysis and virtual screening, highly probable sequences can directly proceed to experimental validation.

The aggregation of multiple insecticidal genes also significantly reduces the probability of target pests developing resistance. By combining insecticidal proteins with different mechanisms of action, pests would need to simultaneously acquire multiple independent resistance mutations to survive, exponentially decreasing the probability. By simulating natural selection pressure, designing chimeric proteins or temporal expression systems that are less prone to inducing resistance can significantly extend the effective lifespan of traits. Currently, companies such as Qingyuan, Dabenong, Ruifeng, and Longping Biological have successfully stacked multiple insecticidal proteins with different mechanisms, including Cry1Ab, 2Ab, 1F, and A105, resulting in very low resistance risk. At the same time, high-throughput screening platforms accelerate the translation of new proteins from laboratory to field, forming a reserve梯队 of traits. This not only mitigates the risk of resistance development but also provides rapid response capabilities to sudden pest succession.

In practical production, the presence of other types of pests such as aphids and spider mites still necessitates the use of insecticides, which itself can slow down the development of resistance in target pests. The synergistic use of chemical and biological control methods can reduce selection pressure and delay the increase in resistance gene frequency. Combined with precision pesticide application technologies, interventions against non-target pests are carried out only when necessary, avoiding damage to natural enemy populations by broad-spectrum insecticides. Simultaneously, linking insecticide use to resistance monitoring results establishes dynamic decision-making models, optimizing control thresholds and application timing. This integrated strategy ensures the long-term effectiveness of Bt traits and enhances the stability of agricultural ecosystems, providing a viable path for future green pest control.

Abandoning the Single "Refuge" Approach: Building a Chinese Solution

The isolated "refuge" strategy not only increases the burden on farmers and seed companies and reduces actual returns, but its effectiveness is also greatly diminished in the context of multi-gene stacking, and it conflicts with the purity requirements of the current Seed Law. If a 10% planting area standard for refuges were required, the affected corn area would reach 60 million mu, with an annual yield loss of 4 million tons. Additionally, seed companies would incur extra costs for backcrossing and breeding, safety assessments, variety approval, base isolation, parent seed amplification, seed production, mixed packaging, logistics management, and distribution. The potential losses and comprehensive social costs incurred annually would be approximately tens of billions. The widespread implementation of a uniform refuge model lacks flexibility, making it difficult to adapt to regional differences in planting structures and pest occurrence patterns. Furthermore, due to the difficulty of supervision and low farmer compliance, actual implementation often becomes a mere formality.

A more realistic path is to leverage the high activity and high expression of insecticidal proteins in plants, the discovery and stacking of more new genes, and comprehensive management strategies combined with monitoring the sensitivity of target pests to insecticidal proteins. This approach not only meets the demand for efficient ecological pest control but also reduces production and management costs, offering a Chinese solution to complex production and biological risks.