CC BY
44
UDC 633.12:581.167:581.145
GENETIC ANALYZES OF SOME MORPHOLOGICAL CHANGES ASSOCIATED WITH THE DIVERGENCE BETWEEN WILD SELFER FAGOPYRUM HOMOTROPICUM OHNISHI AND CULTIVATED OUTCROSSER F. ESCULENTUM MOENCH
Fesenko I.N., Fesenko A.N., Researchers All-Russia Research Institute of Legumes and Groats Crops, Orel City, Russia
E-mail: ivanfesenko@rambler.ru
ABSTRACT
Genetic analysis of interspecific differences between outcrosser Fagopyrum esculentum Moench and selfer F. homotropicum Ohnishi was conducted in both flower size and number of partial inflorescences. Both the characters are under polygenic control. Plus-alleles are dominant in loci influencing flower size, but are recessive in loci affecting the number of partial inflorescences: the «wild type» of floral display in buckwheat is large showy flowers combined with a small number of partial inflorescence. Evolution of self-pollinated species F. homotropicum was accompanied by reducing flower size only. On the contrary, variety population of cross-pollinator F. esculentum maintains genetic system providing large flower (primitive feature) and large inflorescence (derived feature). The maintaining of the complex of genes that provide a small number of metamers (partial inflorescences) in the inflorescence in the wild forms was likely to optimize pollen regime when population density was low, and seed forming regime in barren conditions of mountains.
KEY WORDS
Breeding system evolution; Floral display; Domestication; Fagopyrum.
Plant species with varied types of breeding system, besides presence or absence of self-incompatibility system, are differing, as a rule, in several traits which associated with mode of pollen transfer on compatible stigma. Self-pollinators may produce a little number of pollen grains to pollinate stigma of the same flower. Cross-pollinating species, in addition to incompatibility system, must maintain some properties for successful transfer of pollen between plants. These are additional pollen production and (for insect-pollinated species) some attributes for insect attraction (as a rule, large floral display) [1]. In other words, reproductive system is associated with several costs, and cross-pollination requires additional costs in comparison with self-pollination [2]. Evolution of mating system toward self-pollination may be accompanied by reduction of the costs for reproductive system. Morphologically, it is mainly reduction of floral display and pollen/ovule ratio [3].
Genus Fagopyrum belonging to Polygonaceae family comprises about 15 species, including both selfers and insect-pollinated outcrossers [4]. Cross-pollination is associated here with dimorphic heterostyly; self-pollinators, as a rule, produce more small and less showy homostylous flowers. Since destruction of the self-incompatibility and other mechanisms providing cross-pollination is more simple process in comparison with those of evolutionary synthesis, it seemed that evolution of breeding system is possible from cross- to self-fertilization only [5, 6]. Therefore, it may be concluded that homostylous self-pollinating Fagopyrum species arise from cross-pollinating ancestors.
F. esculentum Moench. and F. homotropicum Ohnishi are closely related species: in crosses F. esculentum * F. homotropicum fertile interspecific hybrids have been obtained [710]. F. esculentum is a cultivated outcrosser with relatively large flowers which combined in large showy inflorescences. Wild species F. homotropicum is a highly polymorphic selfer: lineages with both large (almost like wild F. esculentum) and small (approximately two times smaller) flowers are known.
The objective of this article is to clarify genetic control of the interspecific differences in flower and inflorescence size between F. esculentum and F. homotropicum and to reveal the associations of interspecific variation of the traits with evolution of breeding system.
MATERIAL AND METHODS
Plant material. F. esculentum: a variety Skorospelaya 86 (Sk. 86), which characterized by minimal variation of metameric structure* (number of nodes) of vegetative zone of main stem, and by usual size of infloresences (bred without selection on the trait); F. homotropicum: an accession C9139 from collection of Kyoto University, which is a almost pure line with small homostyle flowers; F1 and F2 hybrids of Sk. 86 * C9139 (* metamer -repeated unit of plant structure (for example, vegetative or generative node, partial inflorescence etc. [11-13]).
Methods. Types of genetic interactions and approximate number of the genes participating in the segregation, were estimated in according with algorithm described by N.A. Sobolev [14] who combined approaches for analyses of genetic interactions elaborated by K. Mather [15] with Castle’s and Wright’s scheme for estimating number of genes [16,17].
The biometrical concept of K. Mather [15, 18] used a genetic algebra to measure the different genetic interactions. For simplest genetic situation including two parents (P1 and P2) and two hybrid generations (F1 and F2), the mean component of dominance in F2 is equal a half (1/2) from component of dominance in F1, and the mean component of epistasis (interloci, or non-allelic genetic interaction) in F2 is equal a quarter (1/4) from component of epistasis in F1. At present time, this approach is standard in quantitative genetics. It was used by N.A. Sobolev [14] for the measurement of genetic interactions. S. Wright [17] shown that genetic variance a2g in a segregate population directly depends on the strength of individual genes involved in the segregation. Genetic variance is maximal when all of the intrapopulational differences are due to segregation on single gene with strong effect. Participation in the segregation of many genes with weak effect reduces the genetic variance in proportion to the number of these genes.
For isolating of the genetic component from phenotypic variances of the F2 generation and non-homozygous parent the principle of additivity of the variances were using, a2g= a2F2 -a2e, where a2e= a2F1 or (a2P1+ a2P2)/2 [19]. Phenotypical variance is directly correlated with the mean value of the trait (X). Therefore, when differences between mean values of lines and hybrids evaluated are large, the variances may be replaced by the coefficient of variation (Cv= a/X). Cv2g=Cv2ph - Cv2e, and a2g=X2 * Cv2g [18].
The method of estimating of the number of additive genes which are segregating in F2, designed by S. Wright [17], has been improved by N.A. Serebrovsky [20], who introduced in the formula the index of dominance.
N.A. Sobolev [14] tried to perfect this method. Also, it is needed to note the works by R. Lande [21] and C.C. Cockerham [22], which generalized the formulations to accommodate two heterogeneous parent populations and their crosses. Today, however, the ideal biometric method for estimating of genes number does not exist [18, 23].
Hybrids and parents were grown in a greenhouse. Flower size (corolla radius) was measured under microscope with an aid of micrometer. Five flowers of each plant were measured, and the mean values were calculated and used for statistics. Inflorescence size (second inflorescence from the bottom on the main stem was analyzed) was evaluated as the number of metamers, which are the partial inflorescences.
Mean values of parents (P1 and P2), hybrids (F1 and F2) and its phenotypical variances (a2ph) were evaluated in the experiment. Dominance and epistasis measures and, also, approximate number of genes participating in the segregation, were calculated by the formulas of Sobolev [14]:
Xa - arithmetic mean of parents, Xa =0.5(P1+P2);
C1 - sum component of genetic interactions in F1, C1=F1-Xa;
C2 - sum component of genetic interactions in F2, C2=F2-Xa; d1 - dominance component in F1, d1=4C2-C1; d2 - dominance component in F2, d2=0.5d1; f1 - epistasis component in F1, f1=2C1-4C2; f2 - epistasis component in F2, f2=0.25f1;
a/ - rough estimation of additive component, a/ =0.5(P1-P2);
D - degree of dominance, D=d1/a/;
E - degree of epistasis, E=f1/a/;
N - approximate number of segregated genes,
N = a|(a| + |d2| + |f2|)/2 a2g,
where a2g - genetic variance in F2: a2g= a2phF2 - a2e; a2e is equal to a2ph of pure line or F1 hybrids.
RESULTS OF RESEARCH
Inheritance of flower size in the cross F.esculentum x F.homotropicum. All investigated plants of F.esculentum produced larger flowers than those of F.homotropicum. A flower (corolla) radius variable in cross-breeder population (Sk.86) considerably wider (in the range of 3.46...4.71 mm, see Table 1) in comparison with F. homotropicum. Relative variability of F. esculentum also was wider (Table 1).
Table 1 - Genetic analysis of differences in flower size between F. homotropicum (C9139)
and F. esculentum (Sk.86)
Parents, hybrids Individuals analyzed Corolla radius, mm a2ph, mm2 °2g2 mm sO 3 Dominance degree Epistasis degree Estimated number of genes
X±m Range
P1, Sk.86 23 4.14±0.06 3.46.4.71 0.0934 0.0715 7.3 - - -
P2, C9139 22 2.19±0.01 2.07.2.31 0.0055 - 3.4 - - -
F1(P1*P2) 5 3.30±0.04 3.24.3.41 0.0063 - 2.4 2.02 -1.83 12
F2(P1*P2) 93 3.69±0.03 2.80.4.68 0.1079 0.1000 8.9 - - -
Additional genetic variance of cross-breeder (P1) in comparison with the selfer (P2) can be estimated as a2gP1 = X2P1CV2g = X2P1(CV2P1 - CV2P2) = 4.142(0.0732 - 0.0342) = 0.07 1 5. Hereditability of this trait in the cultivar population was H2 = a2g/a2ph = 0.0715/0.0934=0.766.
The hybrid population F1 (Sk. 86 * C9139) was small (n=5) and relatively homogeneous (Cv=0.024). Inheritance of the trait was intermediate (range: 3.24...3.41). The mean value of the trait in hybrid population F2 (Sk. 86 * C9139) was significantly different from the one of F1 (Table 1; t=7.8): it suggests that non-additive interactions influence variation of the trait. Variability range in F2 (2.80...4.68 mm) was significantly less than one between parents (2.07.4.71 mm) because of absence of the most small-flowered individuals in hybrid population. Thereby, it is clear that many genes participate in the segregation [24].
Estimation of types and values of genetic interactions was made following to Sobolev’s scheme [14].
P1 = 4.14; P2 = 2.19; F1 = 3.30; F2 = 3.69.
Xa = (4.14 + 2.19)/2 = 3.165; a/ =0.5(P1-P2) = 0.5(4.14 - 2.19) =0.975;
C1=F1-Xa = 3.30 - 3.165 = 0.135; C2=F2 - Xa = 3.69 - 3.165 = 0.525;
d1=4C2-C1 = 2.1 - 0.135 = 1.965; d2=0.5d1 = 0.9825;
f1=2Cr4C2 = 0.27 - 2.1 = -1.83; f2=0.25f1 = -0.4575;
D=d1/a/ = 1.965/0.975 = 2.02;
E=f1/a/ = -1.83/0.975 = -1.88;
N = a/(a/ + |d2| + |f2|) / 2 a2g = 11.58.
Thus, intermediate value (in comparison with parents) of flower size of the F1 hybrids was conditioned by interaction of positive (promote large value of the trait) over-dominance (D = 2.02) with negative (promote small value of the trait) over-epistasis (E = -1.83). The
assumption of the polygenic control of the interspecific difference has proved true, but it is necessary to note that in the situation the exact evaluation of genes number is impossible.
Inheritance of inflorescence size. Mean number of inflorescence metamers in the Sk. 86 population was almost twice as large as the line of F. homotropicum. However, intra-populational variation in the cultivar of F. esculentum was larger too, and the ranges of both species were overlapping (Table 2).
Table 2 - Genetic analysis of differences in number of partial inflorescence between F. homotropicum (C9139) and F. esculentum (Sk.86)
Parents, hybrids Individuals analyzed Number of partial inflorescence „2 a ph a2g vO > O Dominance degree Epistasis degree Estimated number of genes
X±m Range
P1, Sk.86 23 15.3±0.7 9.22 10.23 6.02 20.9 - - -
P2, C9139 22 8.1 ±0.2 7.10 1.18 - 13.4 - - -
F1P1XP2) 5 12.6±0.7 11.15 2.29 - 12.0 - 2.25 2.5 25
*F1(P1M*P2) - - - - - - -1.8 2.8 8
F2(P1*P2) 93 9.9±0.2 5.16 2.48 0.72 15.9 - - -
*Analysis of hypothetical data (assumption, that Pi= Pim= Fi), see explanations in the text.
Additional genetic variance of the cultivar (P1) in comparison with the wild line (P2) can
be estimated as a2gP1 = X2P1(CV2P1 - CV2P2) = 15.32(0.2092 - 0.1342) = 6.02 1 9. So, heritability
of the trait in Sk. 86 population was H2=6.0219/10.2253=0.589.
The hybrid population F1 (Sk.86 * C9139) was sufficiently homogeneous, with almost intermediate (in comparison with parental forms) average expression of the trait. Population means of F2 (Sk.86 * C9139) and F1 (Sk.86 * C9139) were significantly differed (Table 2; t=3.71), that forces to refuse the assumption about additive mode of interaction of the genes, causing intermediate inheritance of the trait by F1 hybrids. Range of variability among F2 hybrids (5 ... 16) was much less than the range between parental forms (7 ... 22): in the hybrid population the plants with the largest inflorescences were absent, that testifies to participation of large number of genes in the segregation [24]. Genotypic variance of F2 (Sk.86 * C9139) also was sufficiently small: a2gF2 = X2F2(Cv2F2 - CV2P2) = 9.92(0.1592 - 0.1342)
= 0.72. Following genetic analysis was conducted in accordance with N.A.Sobolev's scheme:
Xa = (15.3 + 8.1)/2 = 11.7; a/ = 0.5(15.3 - 8.1)/2 = 3.6;
C1=F1-Xa = 12.6 - 11.7 = 0.9; C2=F2-Xa = 9.9 - 11.7 = - 1.8;
d1=4C2-C1 = 4(-1.8) - 0.9 = - 8.1; d2=0.5d1 = - 4.05; f1=2C1-4C2 = 1.8 - (- 1.8) 4 = 9.0; f2=0.25f1 = 2.25;
D=d1/a/ = (- 8.1)/3.6 = - 2.25; E=f1/a/ = 9.0/3.6 = 2.5;
N = a/(a/ + |d2| + |f2|) / 2 a2g = 3.6(3.6 + 4.05 + 2.25) / 1.44 = 24.8.
Thus, intermediate value (in comparison with parents) of the inflorescence size of the interspecific F1 hybrids is conditioned by overlapping of negative over-dominance (D =-2.25) with positive over-epistasis (E = 2.50). The assumption about polygenic control of the interspecific difference for this trait has proved true, but in the situation the exact estimation of the genes number is impossible.
Specificity of the genetic situation is a significant excess of genetic variance of the population P1 (Sk. 86) in comparison with that of population F2 (Sk. 86 * C9139) (Table 2).
Besides, in population F2 (Sk. 86 * C9139) the transgression toward the least value of the
trait was observed, but there were no plants with large inflorescences. Possibly, the interspecific hybrids have been obtained on the basis of a genotype, which was homozygous for minus-alleles at the strongest loci determining variability of the cultivar’s population of F. esculentum. For correction the results of the genetic analysis, model of this situation was made. Unknown value of this model is P1M, a hypothetical mean of Skorospelaya 86. Probably, this value should be closer to F1, than the experimental mean of P1. In other words, the modeling value P1M should be equated to the experimental mean of F1. The analysis will be as:
P1M= F1=12.6; P2=8.1; F1=12.6; F2=9.9; Xa=0.5(12.6+8.1)=10.35; av=0.5(12.6-8.1)=2.25;
C1=12.6-10.35=2.25; C2=9.9-10.35=-0.45; d1=-1.8-2.25=4.05; d2=-2.03; f1=6.3; f2=1.57; D=-
1.8; E=2.8; N=7.6.
Modeling (replacement of P1 by P1M) has not changed the type of genetic interactions revealed: over-dominance of small value of the trait (D =-1.8) together with positive over-epistasis (E=2.8), which together create the illusion of dominance of large value of the trait (positive dominance) in F1. However, the estimated number of genes looks here more plausibly.
DISCUSSION
By genetic analyses it is revealed that the different components of floral display demonstrate the different directions of evolutionary changes, and, therefore, its alterations affected by different selective forces. Since the plus-alleles are dominant in the loci influencing flower size, but are recessive in the loci affecting the number of partial inflorescences (metamers), the «wild type» of floral display in buckwheat is large showy flowers combined in small number of partial inflorescence. Evolution of self-pollinated species F. homotropicum was accompanied by reducing of flower size only. On the contrary, variety population of cross-breeder F. esculentum maintains genetic system providing large flower (primitive feature) and large inflorescence (derived feature).
Large inflorescence is obviously needed for cultivated varieties. Selection for yield has led to increasing of inflorescence metamers number. In this work a cultivar of F. esculentum was used which did not pass a special selection for the inflorescence size, and it was shown that intravarietal polymorphism on this trait proved quite high. The variation of inflorescence metameres number in the cultivar Skorospelaya 86 is conditioned by loci with more powerful effect than polygenes determining the interspecific differences. High genetic heterogenity for inflorescence size, possibly, is a component of genetic system of intra-population heterosis, which is required for maintenance of the cultivar productivity. Successful selection is carried out for further increasing of the inflorescence size [25].
Large flowers are needed for an entomophilous species as its should be visibable to pollinators. On the other hand, a priori, a large inflorescence would be to perform the same functions. However, as shown in this work, a large inflorescence is due to selection of recessive genes, and, apparently, is a result of secondary evolution, probably already after domestication. It remains an open question as to why cross-pollinated ancestor of the F. homotropicum maintained genetic system that ensures the formation of small inflorescence?
It was established experimentally [26] that the simultaneous opening of many flowers in the inflorescence promotes geitenogamy (pollination by pollen grains from adjacent flowers). Thus, the maintaining of the complex of genes that provide a small number of metamers in the inflorescence in the wild forms was likely to optimize pollen regime when population density was low. In cultivated populations, plant density is sufficient for the normal functioning of cross-pollination, and possibility of geitenogamy is minimized. Maintenance of the small number of partial inflorescences in F. homotropicum may be considered also in frame of the concept of reproduction costs.
Wild buckwheat grows on stony hillsides in Southern China [27]. In such conditions possibility of above-ground biomass growth is sharply limited and should not exceed capabilities of root system in the stony ground: high competitiveness is not required. The morphotype of the line C9139 as a whole satisfies to these requirements. This plant of the alpine type with slowed down growth of sprouts that minimizes loading on root system (windage, water stress etc.) and promotes reliable rooting of a young plant. Generative period takes place slowly and stretched in time due to the consecutive formation and flowering of numerous small inflorescences.
Common buckwheat cultivars have much more intensive ontogenesis rhythm. After sowing in the warmed soil such buckwheat competes successfully with annual weeds, then blooms and sets the seeds intensively. Most likely, evolution of F. esculentum from the alpine
type to the competitive one has occurred in the course of «descent» of this species from mountains in valleys. Apparently, this «descent from the mountains» was possible due to cross-pollination, since selfer F. homotropicum could not overcome this barrier.
ACKNOWLEDGEM ENTS
Authors are grateful to Dr. Ohmi Ohnishi, Plant Germ-Plasm Institute, Kyoto University, for kindly supplying seeds of the accession C9139. This work was supported by the State contract 14.512.11.0063.
REFFERENCES
1. Grant, V. Plant speciation / V. Grant. - New York: Columbia Univ. Press, 1981.
2. Solbrig, O.T. A cost-benefit analysis of recombination in plants / O.T. Solbrig // In: O. T. Solbrig, S. Jain, G. B. Johnson & P. H. Raven (eds.). Topics in plant population biology. -New York: Columbia University Press, 1979. - P. 114-130.
3. Cruden, R.W. Pollen-ovule ratios: a conservative indicator of breeding systems in flowering plants / R.W. Cruden // Evolution. - 1977. -V.24. - P. 32- 46.
4. Yasui, Y. 1998. Evolutionary processes of Fagopyrum inferred from the molecular phylogenetic analyses / Y.Yasui, T. Ohsako, O. Ohnishi // Proc. 7th Intl. Symp. Buckwheat at Winnipeg. - 1998. - Part 6. - P. 50-60.
5. Jain, S.K. The evolution of inbreeding in plants / S.K. Jain // Annual. Rev. Ecol. Syst. -
1976. - V. 7. - P. 469-495.
6. Takebayashi, N. Is self-fertilization an evolutionary dead end? Revisiting an old hypothesis with genetic theories and a macroevolutionary approach / N. Takebayashi, P.L. Morrell // Amer. J. Botany. - 2001. - V. 88. - P. 1143-1150.
7. Campbell, C. Interspecific hybridization in the genus Fagopyrum / C. Campbell // Proc. 6th Intl. Symp. Buckwheat at Ina. - 1995. - V. 1. - P. 255-263.
8. Fesenko, N.N. Some genetic peculiarities of reproductive system of wild relatives of
common buckwheat Fagopyrum esculentum / N.N. Fesenko, A.N. Fesenko, O. Ohnishi //
Proc. 7th Intl. Symp. Buckwheat at Winnipeg. - 1998. - Part 6. - P. 32-35.
9. Fesenko, I.N. Compatibility and congruity of interspecific crosses in Fagopyrum / I.N. Fesenko, N.N. Fesenko, O. Ohnishi // Proc. 8th Intl. Symp. Buckwheat at Chunchon. -2001. - V. 1. - P. 404-410.
10. Fesenko, N.N. Functional fragments of a relictual gametophytic self-incompatibility system are associated with the loci determining flower type of the heterostylous outcrosser Fagopyrum esculentum Moench. and the homostylous selfer F. homotropicum Ohnishi / N.N. Fesenko, I.N. Fesenko // Russian J. Genetics. - 2011. - V. 47. - P. 41-48.
11. Meyen, S.V. Plant morphology in its nomothetical aspects / S.V. Meyen // Bot. Rev. -1973. - V.39. - P. 205-260.
12. Harper, J.L. Population biology of plants / J.L. Harper. - London: Academic press, 1977.
13. White, J. The plant as a metapopulation / J. White // Ann. Rev. Ecol. Syst. - 1979. - V.10. - P. 109-145.
14. Sobolev, N.A. Genetic analyses of quantitative traits / N.A. Sobolev // Cytology and Genetics. - 1976. - V. 10. - P. 424-436 (in Russian).
15. Mather, K. Biometrical Genetics / K. Mather. - London: Methuen, 1949.
16. Castle, W.E. An improved method of estimating the number of genetic factors concerned in cases of blending inheritance / W.E. Castle // Science. - 1921. - V. 54. - P. 223.
17. Wright, S. Evolution and the Genetics of Populations, Vol. 1. Genetic and biometric foundations / S. Wright. - Chicago: University of Chicago Press, 1968.
18. Mather, K. Biometrical genetics / K. Mather, J.L. Jinks. - London, New York: Chapman and Hall, 1982.
19. Wright, S. The results of crosses between inbred strains of guinea pigs, differing in number of digits / S. Wright // Genetics. - 1934. - V. 19. - P. 537-551.
20. Serebrovsky, A.S. An analysis of the inheritance of quantitative transgressive characters / A.S. Serebrovsky // ZIAV. - 1928. - V. 48. - P. 229-243.
21. Lande, R. The minimum number of genes contributing to quantitative variation between and within population / R. Lande // Genetics. - 1981. - V. 99. - P. 541-553.
22. Cockerham, C.C. Modifications in estimating the number of genes for a quantitative character / C.C. Cockerham // Genetics. - 1986. - V. 114. - P. 659-664.
23. Petrov, A.P. A mode of testing of some methods of statistical genetics / A. P. Petrov, B.V. Bondarev // Cytology and Genetics. - 1982. - V. 16 . - P. 53-59 (in Russian).
24. Powers, L. Gene analysis by the partitioning method when interactions of genes are involved / L. Powers // The Botanical Gazette. - 1951. - V. 113. - P. 1-23.
25. Martynenko, G.E. Inheritance of a large inflorescence at determinant buckwheat / G.E. Martynenko // Bull. VNII ZBK. - 1996. - V. 42. - P. 95-98 (in Russian).
26. Harder, L.D. Mating cost of large floral displays in hermaphrodite plants / L.D. Harder,
S.C.H. Barrett // Nature. - V. 373. - P. 512-515.
27. Ohnishi, O. Discovery of new Fagopyrum species and its implication for the studies of evolution of Fagopyrum and of the origin of cultivated buckwheat / O. Ohnishi // Proc. 6th Intl. Symp. Buckwheat at Ina. - 1995. - P.175-190.
