Spatial and temporal assessment of pollen- and seed-mediated gene flow from genetically engineered plum Prunus domestica

Abstract

Pollen flow from a 0.46 ha plot of genetically engineered (GE) Prunus domestica located in West Virginia, USA was evaluated from 2000-2010. Sentinel plum trees were planted at distances ranging from 132 to 854 m from the center of the GE orchard. Plots of mixed plum varieties and seedlings were located at 384, 484 and 998 m from the GE plot. Bee hives (Apis mellifera) were dispersed between the GE plum plot and the pollen flow monitoring sites. Pollen-mediated gene flow from out of the GE plum plot to non-GE plums under the study conditions was low, only occurring at all in 4 of 11 years and then in only 0.31% of the 12,116 seeds analyzed. When it occurred, gene flow, calculated as the number of GUS positive embryos/total embryos sampled, ranged from 0.215% at 132 m from the center of the GE plum plot (28 m from the nearest GE plum tree) to 0.033-0.017% at longer distances (384-998 m). Based on the percentage of GUS positive seeds per individual sampled tree the range was 0.4% to 12%. Within the GE field plot, gene flow ranged from 4.9 to 39%. Gene flow was related to distance and environmental conditions. A single year sample from a sentinel plot 132 m from the center of the GE plot accounted for 65% of the total 11-year gene flow. Spatial modeling indicated that gene flow dramatically decreased at distances over 400 m from the GE plot. Air temperature and rainfall were, respectively, positively and negatively correlated with gene flow, reflecting the effects of weather conditions on insect pollinator activity. Seed-mediated gene flow was not detected. These results support the feasibility of coexistence of GE and non-GE plum orchards.

Figure 1. Plot T0 source block for gene flow from genetically engineered (GE) plum trees.

Figure 2. Experimental plantings of genetically engineered plums T0 and gene flow sampling plots (S1–S5; M1–M3).

A total of 12,116 seeds were sampled for GUS assays over 11 years. Distances of sample plots from plot T0 and % GUS positive seeds/total seeds (white numbers) and % GUS positive seeds/total seeds/plot (black numbers ) are indicated. Borderline indicates USDA property line.

Figure 3. Estimates of the proportion of positive seeds per sampled tree for the 52 sampled tree locations outside of the GE plot (Plot T0) from 2000–2010, based on the fit of the exponential dispersal model to the data.

Predictions and 95% prediction intervals are shown for the three tree locations surveyed in plot T0 (distances of 67, 95, and 97 m).