Apomixis and endosperm development
--Tyrnov, VS

The problem of apomixis is of great economic importance. At present efforts have been on experimental production of apomictic forms with traditional genetical-breeding methods as well as methods of molecular biology and genetic engineering. As a rule, insufficient attention is given to endosperm and often to the problem of double fertilization as a whole.

On the practical side, the ideal form of apomixis is an autonomous one, when both embryo and endosperm develop without fertilization, and pollen is not necessary at all. Autonomy not only contributes to fixation of heterosis and other unique characters, but it also decreases the dependence of seed productivity on influence of unfavourable temperature, drought, rain and other abiotic factors during the process of pollination. In addition, it decreases dependence on insect-pollinators. It can not be excluded, that, in perspective, maize genes can be transferred by methods of molecular biology to other cultures, including insect-pollinating ones.

Unfortunately, in maize reliable examples of absolutely autonomous formation of embryo and endosperm and, as a consequence, seeds are unknown. At the same time in maize, in all probability, pseudogamy, where the embryo is apomictic, but the endosperm can develop only after fertilization, is distributed widely (Tyrnov and Selivanov, pp. 124-134 in Apomixis and its use in breeding, Moscow, "Kolos", 1976. In Russian).

By using the method of genetical marking we found repeatedly diploid plants of the maternal type, grown from kernels, the endosperm of which had all marker genes of the paternal form. We examined more than 10 million genetically marked kernels and seedlings. The frequency of occurrence of maternal type diploids in different materials was in the range 0.01-0.001%.

Pseudogamy is also a characteristic of haploids (reduced apomicts) produced experimentally. We analyzed more than 20,000 maize haploids, arising in marker crosses. They all had hybrid endosperm, possessing characters (genes) of a parental form. Our investigations showed that it was a result of single fertilization of the central cell (Enaleeva et al., Dokl. Biol. Sci. 353:225-226, 1997). Among a low number of non-marked kernels haploids were not discovered. In all probablity, the origin of such kernels is from pollen contamination.

We also described lines of maize (Tyrnov, MNL, 71:73-74, 1997; Enaleeva and Tyrnov, MNL, 71:74-75, 1997) having haploid parthenogenesis with a frequency up to 100%. At initial developmental stages, according to embryological data, autonomous apomixis takes place, i.e. the development of embryo and endosperm proceed without pollination. However, endospermogenesis does not go to the end. Haploid plants develop only in that case, if endosperm is formed by fertilization. And so, haploid parthenogenetic embryos can develop normally without fertilization, but endosperm can not.

Pseudogamy is also characteristic of the apomictic maize relative Tripsacum, which is regarded by many scientists as a donor of apomixis genes.

If anyone takes a look at the problem of pseudogamy from evolutionary positions, he can conclude that the evolutionary process leads first to apomictization of embryo, but not endosperm. Since endospermal function is limited, specific and short-timed, anyone would think that the contrary phenomenon must be observed - in the first turn apomictization of endosperm, but not embryo. It is probable there is a profound sense in such "unlogical evolution". We suppose that one of the important functions of endosperm connects with keeping the stability of sexual reproduction.

The analysis of evolutionary preconditions, experimental data and logical constructions allows some conclusions (Tyrnov, pp. 112-114 in Gametic and zygotic selection, Kishinev, 1987. In Russian).

1. Embryo development without fertilization is a rather frequent phenomenon. It is, probably, a constant component of the plant reproduction system. This puts under a threat reproduction by sexual means, that, in absolute expression, is not always worth while or useful for onto- and phylogenesis. At the same time it is the precondition for successful artificial selection of forms with increased tendency to apomixis.

2. A high level of spontaneous apomictization has led to creation of a binary system (embryo - endosperm), securing conservation of stability for a sexual reproduction system. Apomictic endosperm, as a rule, has genome content that does not allow it to develop normally. Undeveloped endosperm leads to the ruin of the embryo and, as a consequence, to the elimination of forms with an apomictic tendency.

What is the genomic content of endosperm? In diploids, it has, as a rule, 3n ploidy. Even in polyploids (4n, 6n and etc.) the ploidy is divisible by 3. When the endosperm develops without fertilization, it must have "n" or "2n" ploidy (2n or 4n in tetraploids) in connection with fusion of polar nuclei.

The possibility of endosperm development with another ploidy than 3n was discovered in crosses 2n x 4n and 4n x 2n, producing endosperm with 4n and 5n ploidy. Our data and those of other investigators show, that tetraploid endosperm, as a rule, is not able to develop, and pentaploid endosperm has various degrees of completeness.

This situation can be extrapolated on apomictic forms. In embryo sacs, potentially able to undergo haploid parthenogenesis, fusion of polar nuclei gives 2n ploidy. In case of non-reduction each of the polar nuclei has 2n ploidy and their fusion gives 4n. Thus, in all cases, the nuclei are formed with ploidy leading to degeneration of endosperm. Fertilization gives origin to development of endosperm with 3n and 5n ploidy. The last can be rather developed, and an apomictic embryo has a chance to survive. Maybe that is just what contributes to pseudogamy and makes the passage to autonomous apomixis difficult.

Can the mentioned limitation be overcome anyhow? Probably, yes. Autonomous apomicts occur, although rarely, in nature in different species. Almost all apomicts are polyploids, including triploids. It is possible that increased endosperm ploidy can contribute satisfactory development. It can not be excluded, that the work on experimental production of apomicts must be carried out on a triploid level, since endosperm will have optimal ploidy - 3n or 6n. We examine this hypothesis now. In addition, species are known, in which only one of two polar nuclei is fertilized, and their endosperm is, possibly, diploid. The regularities of endospermogenesis in such species merit special investigation.

We also draw attention to one circumstance. By unreduced pseudogamous apomixis the endosperm will have, most probably, 5n (in diploids). At the same time very puny kernels are formed. And so, instead of heterosis, significant reduction of seed productivity can take place. We observed this phenomenon in model experiments with tetraploids of apomictic maize line AT-1. Is any progress possible in this field? It can not be excluded. We pollinated the common tetraploids (non-apomictic) by pollen of the different lines and hybrids. Theoretically, expected ploidy of the endosperm must be 5n. We observed various degrees of endosperm development: from forms like glumes to half of normal. Single kernels look like the ordinary small kernels located usually at the top of the ear. It points to the possibility of increasing, by selection, the tendency to formation of full-formed endosperm with atypical ploidy.

The given data speak about necessity to examine the regularities, concerning the role of endosperm in the reproduction system, since it is difficult to reckon on success in production of apomictic forms, even if anyone will be able to isolate and to clone apomixis genes.

Return to the MNL 72 On-Line Index
Return to the Maize Newsletter Index
Return to the MaizeGDB Homepage