Lessons from Two Decades of Field Trials with Genetically Modified Trees in the USA: Biology and Regulatory Compliance

Abstract

We summarize the many field trials that we have conducted in the USA beginning in 1995 and continuing to this day. Under USDA APHIS federal regulatory notifications and permits, we have planted nearly 20,000 trees derived from approximately 100 different constructs in more than two dozen field experiments. The large majority of the trials were in Populus and included hybrid white poplars (P. tremula × alba INRA 717-1B4 and P. tremula × tremuloides INRA 353-53), but also included diverse hybrid cottonwoods such as P. trichocarpa × deltoides and P. deltoides × nigra. One field trial used transgenic sweetgum ( Liquidambar ). Most trials were conducted on Oregon State University (OSU) land, but several were also conducted on the land of industry collaborators in Oregon, Washington, and other states. The main traits we have studied are floral sterility and flowering time modification; size and growth rate modification by gibberellin perturbation; activation-based gene tagging; stability of reporter gene expression and RNAi suppression; herbicide and pest resistance gene impacts on plantation productivity; lignin modification and its impacts on physiological processes; and effects of isoprene reduction on growth and stress tolerance. The most significant lessons from these years of trials are: (1) Visual abnormalities in form or growth rate due to the transformation and in vitro regeneration (somaclonal variants) have been observed in several experiments, but are extremely rare (below 1 % of events produced). (2) Gene expression and RNAi-induced gene suppression have been highly stable—with a virtual absence of transgene silencing —over many years for virtually all transgenic trees whether assayed by a visual phenotype (reporter gene, flowering time, sexual sterility, herbicide or pest tolerance), or by molecular measures of transgene expression (e.g., quantitative RT-PCR). (3) The regulatory process has largely been efficient and workable, though it imposes significant biological constraints, costs, and risks that are very difficult for an academic laboratory to bear when trials span several years. It is most difficult where flowering is needed. (4) Field environments invoke complex and largely unpredictable changes to expression and associated phenotypes when studying physiology-modifying transgenes, including those affecting wood properties, suggesting the need to study several field sites, genetic backgrounds, and gene insertion events over many years, similar to common practices of conventional breeding. However, regulatory requirements make this very difficult to do for transgenic trees. (5) Collaborative field trials with industry have shown that common transgenic traits, such as herbicide and insect resistance, can have large productivity benefits in near-operational plantation conditions (e.g., two-year volume growth improvements of ~20 %)—suggesting that it could be highly beneficial to incorporate transgenic traits into production programs. Regulatory reforms to focus on product benefits as well as risks, and that do not assume harm from the use of recombinant DNA methods, are needed if transgenic technology is to provide significant benefits in forestry.