--J. R. Laughnan and S. Gabay-Laughnan
Among the many spontaneous nuclear revertants from S-type male-sterile (cms-S) maize that have been identified is a class we refer to as pseudorestorer. A plant heterozygous for one of these pseudorestorer genes is phenotypically male-fertile, having normal anther exsertion and exhibiting the 50% pollen abortion expected of gametophytic restorers of cms-S. However, when these "fertile" plants are crossed as male parents onto cms-S male sterile testers, or onto male-fertile isogenic maintainer plants with normal (N) cytoplasm, there is no seed set. The fact that this class of restorer gene produces nonfunctional pollen has led to the designation Rf-nf (see previous report).
Two alternative hypotheses have been proposed to explain the characteristics of the Rf-nf strains. The Rf-nf gene may code for a product that, while resulting in phenotypically male-fertile plants, does not lead to production of functional pollen. Alternatively, failure of pollen function may result from insertion of a newly arisen Rf element into a gene necessary for male-gametophyte function, thus in effect inducing a male gametophytic lethal mutation. A protocol has been developed that will differentiate between these two hypotheses. It involves crossing a strain heterozygous for an Rf-nf gene in cms-S by a strain carrying an unlinked Rf gene that is male functional. Initial experiments were invalid since we later learned that pollen grains carrying Rf-nf alleles can function in hybrid backgrounds. The crosses are now being repeated in the inbred-line backgrounds Oh51A, B37 and WF9 since it is only in these nuclear backgrounds that we have both functional Rf and Rf-nf genes.
The procedure which, by analogy with the behavior of normal and defective lambda particles in E. coli, we have called the "helper" experiment is as follows: a cms-S Rf-nf rf (50% pollen abortion) plant is crossed as female parent with a pollen parent carrying an unlinked functional Rf gene. Male-fertile plants among the offspring will be of two types: those with 50% aborted pollen, carrying only the functional gene from the male parent, and those with 25% aborted pollen, carrying both the functional Rf and the Rf-nf gene. Those plants with 25% pollen abortion are crossed as female parents by the appropriate nonrestoring maintainer inbred line. Plants with 50% pollen abortion are crossed as controls. The former cross should yield only 1/4 male-sterile plants while the later control cross will produce 1/2 male-sterile offspring. These crosses are made to confirm the pollen records of the female parents. The same plants that are tested as female parents are also crossed as pollen parents onto cms-S male-sterile testers. Testcrosses of the control plants (50% abortion) will produce only male-fertile plants with 50% pollen abortion. Testcrosses of plants carrying both the functional Rf and the Rf-nf (25% abortion) genes will give either of two results depending on which hypothesis is valid. If the Rf-nf gene codes for a defective restorer gene product, the testcrosses will yield both plants with 50% aborted pollen and plants with 25% aborted pollen since, on this hypothesis, pollen grains receiving both the functional Rf (helper gene) and the Rf-nf will be functional. These two types of offspring should occur in a 1:1 ratio if the two kinds of pollen grains, Rf; rf-nf and Rf; Rf-nf, function equally well. If, on the other hand, failure of Rf-nf pollen function results from insertion of a newly arisen Rf into a gene that is indispensable for male-gametophyte function, thus in effect producing a gametophytic lethal, pollen grains carrying an Rf-nf gene, even in the presence of a functional Rf gene, will fail to function and no plants with 25% aborted pollen should be observed in the testcross progeny.
The first tests of the two hypotheses concerning the nature of Rf-nf genes in the Oh51A inbred line background were carried out in the summer of 1988. The progeny of eight different plants exhibiting 25% pollen abortion and testcrossed as pollen parents onto cms-S male-sterile testers were examined. Of 78 plants, 77 were male-fertile and exhibited 50% pollen abortion. No plants with 25% abortion were observed. One plant was too late in maturity to be scored. Four control plants yielded 41 plants, all with 50% pollen abortion. These results indicate, at least for the two Oh51A Rf-nf strains tested, that nonfunctional pollen in pseudorestorer strains can result from disruption of a gene necessary for male-gametophyte function.
As indicated above, only two Rf-nf genes, of separate origin, have been tested by the helper procedure. The finding that in both cases disruption of a gene that is indispensable for male-gametophyte function is involved does not mean that other Rf-nf genes that fail in male-gametophyte function because they are defective versions of restorer genes will not be found. In any case, it now appears that, given numbers of spontaneously revertant Rf-nf genes to analyze by the helper procedure, those that involve lethal mutations, presumably brought about by insertion of Rf elements into genes that are vital for pollen function, provide a unique opportunity to identify, map and characterize them at the molecular level. Although pollen grains that carry Rf-nf in this category are lethal to the male gametophyte, they do not abort. In all but two such strains the pollen grains, upon microscopic examination, are normal in appearance. The two exceptions involve Rf-nf genes that arose spontaneously in cms-S versions of inbred lines M14 and WF9, but even in these strains the pollen grains are partially filled. This means that pollen from these mutant strains and from their isogenic maintainer and Rf-carrying strains with normal pollen function can be analyzed at the DNA, RNA and protein levels. Moreover, if these Rf-nf strains result from transposition and insertion of previously quiescent restorer genes, it is expected that such studies will lead to isolation and characterization of these restorer genes themselves, and following that to a better understanding of the molecular basis for S-type cytoplasmic male sterility. In this connection we note that Rf-nf mutants whose nonfunctional character is based on defective rf genes also afford the opportunity to isolate and characterize both normal and defective versions of the cms-S restorer.
How shall we proceed to obtain the relatively large numbers of Rf-nf
revertants that will be required to establish a reasonably good map of
male-gametophyte genes and an understanding of their functions? First,
we now have in hand a total of seven such revertant strains. Second, we
plan to intensify our search for spontaneous revertants from a variety
of cms-S male-sterile inbred lines. Probably the most fruitful source of
these revertants will come from the use of the Rf-nf strains themselves
as male parents in crosses with cms-S male sterile tester strains. At an
earlier point in this article it was stated that such crosses give no seed
sets. Actually, some such crosses produce seeds at a very low frequency,
most often one or two kernels per ear with most of the ears being barren.
Since both the male-sterile female parent and the Rf-nf male parent
are isogenic (same inbred-line background) contaminants are easily identified
on the basis of the hybrid vigor exhibited by next generation offspring.
Using this crossing procedure we have identified numbers of legitimate
offspring from such crosses. In some cases, even though the male parents
carried an Rf-nf gene, the new derivative carries a functional Rf
gene. In other cases these offspring turn out, like their male parents,
to carry Rf-nf genes. Since these might be rare cases of transmission
through the male gametophyte they can not be regarded as different from
the parental Rf-nf unless procedures such as mapping and allelism
tests prove otherwise. The search for increased numbers of Rf-nf
revertants, and their verification by these methods are underway at the
to the MNL 63 On-Line Index
Return to the Maize Newsletter Index
Return to the Maize Genome Database Page