SummaryThe sterol biosynthesis pathway of Arabidopsis produces a large set of structurally related phytosterols including sitosterol and campesterol, the latter being the precursor of the brassinosteroids (BRs). While BRs are implicated as phytohormones in post-embryonic growth, the functions of other types of steroid molecules are not clear. Characterization of the fackel (fk) mutants provided the ®rst hint that sterols play a role in plant embryogenesis. FK encodes a sterol C-14 reductase that acts upstream of all known enzymatic steps corresponding to BR biosynthesis mutants. Here we report that genetic screens for fklike seedling and embryonic phenotypes have identi®ed two additional genes coding for sterol biosynthesis enzymes: CEPHALOPOD (CPH), a C-24 sterol methyl transferase, and HYDRA1 (HYD1), a sterol C-8,7 isomerase. We describe genetic interactions between cph, hyd1 and fk, and studies with 15-azasterol, an inhibitor of sterol C-14 reductase. Our experiments reveal that FK and HYD1 act sequentially, whereas CPH acts independently of these genes to produce essential sterols. Similar experiments indicate that the BR biosynthesis gene DWF1 acts independently of FK, whereas BR receptor gene BRI1 acts downstream of FK to promote post-embryonic growth. We found embryonic patterning defects in cph mutants and describe a GC±MS analysis of cph tissues which suggests that steroid molecules in addition to BRs play critical roles during plant embryogenesis. Taken together, our results imply that the sterol biosynthesis pathway is not a simple linear pathway but a complex network of enzymes that produce essential steroid molecules for plant growth and development.

To determine if standard breeding methodology is applicable to transgenes, phenotypic recurrent selection was used to select for increased transgene expression in white clover, Trifolium repens L. Plants were transformed with nptII and gusA, and selected on 100 mg L−1 of kanamycin. Independently transformed plants were intercrossed, and the progeny was germinated on 200, 300, or 400 mg L−1 of kanamycin. Those seedlings surviving on 400 mg−1 were in turn intercrossed, and the progeny was selected on 300, 400, or 500 mg L−1 of kanamycin. NPTII levels were measured in each selected population, and Southern blots were made from individuals in each population. The highest‐expressing individual in the T2 had levels of NPTII that were more than four times higher than those in the highest parent. With selection on increasing levels of kanamycin, average expression across each generation went from 0.033 ng μg−1 NPTII in the parents to 0.095 ng μg−1 in the selected T1 plants to 0.539 ng μg−1 in the selected T2 plants. Southern hybridization suggested that plants displaying a heightened level of nptII expression in the T1 and T2 fell into two categories. The first contained one particular transgenic event, implicating the importance of other genomic factors in modulating gene expression. Alternatively, the plants had an accumulation of various nptII loci, suggesting an association between multiple transgene copies and high expression levels. On the basis of these results, selection for transgene expression appears to be a viable option for plant breeding programs.