Stanford University

"Pink" kernels in bz2 testers--a connection to ABA levels?

--Virginia Walbot, Juli Nash and Maria-Ines Benito

The bz2 tester lines we have received (in W23 or K55 backgrounds from Ed Coe, EH, Jr.; bz2 in combination with diverse sources of novel Adh1 alleles from Mike Freeling, M) or constructed (various exotic backgrounds) exhibit a sporadic expression of medium to dark pink kernel color. Typically just a few kernels per ear are affected, but in some ears most kernels are pink; some ears in the same family will have no pink color. The hue of the darkest pink kernels is annoyingly similar to dilute (poor) expression of Bz2. We do not have this problem with testers for any other gene of the pathway. Crosses with plants derived from bz2 kernels selected as pink, dark bronze or the light, standard bronze color did not resolve the nature of the pink phenotype other than indicate that some lines were more prone to pink coloration. At the suggestion of J. D. Smith, JD who paints ABA onto developing kernels to overcome vivipary, we treated bz2 ears in summer 1991 with this hormone at a developmental stage when ABA is known to penetrate into the developing seed. Longitudinal quadrants representing four treatment conditions were created on each ear: exposure to light, light + ABA, dark + ABA, and the dark only control. We cut through the husks 12-14 dap of ears in the hybrid W23/K55 background to expose kernels on three-fourths of each ear. We put a strip of folded paper towel wetted with 10-4M ABA on half of the ear; the husks were then smoothed into place over one-half of the towel and secured in place by a rubber band. The brown paper pollinating bag was replaced and nothing further done. At harvest about half (9/16) of the well-filled, treated ears had a clear sector of dark pink kernels in the ABA + light quadrant and an enhanced number of pink kernels in the dark + ABA quadrant. The other treated ears lacked such a clear demarcation of the treatment areas but had an enhanced number of pink kernels in one or more patches. Parallel treatment of a1 ears in a similar background yielded only the expected colorless kernels. Tests are in progress in Hawaii with r-g, c1, c2, a2, bz1, and bz2 testers to confirm our findings and determine whether the ABA effect is specific to bz2.

Our observations suggest that high levels of ABA can induce the pink phenotype in a bz2 tester. A plausible explanation for the sporadic appearance of pink kernels would be that a variable number of kernels per ear experience sufficient stress, probably water deficit, to increase endogenous ABA to a level sufficient to cause the pink phenotype. If this model is true, then many lines of maize contain an ABA-inducible gene that can carry out BZ2 function, albeit at a very low level as pink rather than purple pigmentation results. Based on genetic criteria, Bz2 encodes the terminal step of anthocyanin synthesis in maize. This step could be a malonyl transferase or other enzyme that decorates the glucose moiety to make a more stable molecule; the enzyme giving the pink phenotype would then be postulated to normally carry out a similar reaction on a different substrate but when induced to high levels by ABA and in the presence of substantial anthocyanin precursor, the "pink" enzyme catalyzes the BZ2 reaction. Alternatively, Bz2 could encode a protein that guides anthocyanin into the vacuole from its site of synthesis in the cytoplasm, and the "pink" product performs a similar task for a related molecule. To date, Bz2 shows the highest similarity to a stress-induced gene of soybean, Gmhsp26 (Nash, J and Walbot, V, this issue), but the biochemical function of this gene is also unknown.

Return to the MNL 66 On-Line Index
Return to the Maize Newsletter Index
Return to the MaizeGDB Homepage