Tải bản đầy đủ - 0trang
CHAPTER 4. THE GENETIC CONTROL OF FLOWERING AND GROWTH IN SORGHUM
J. R. QUINBY
D. Temperate and Tropical Pairs of Varieties . . . . . . . . . . . . . . . . . . .
E. Height Genotypes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
XI. Summary and Discussion of Genetic Control of Growth in Sorghum .
XII. Implications to Plant Breeding .................................
A. Sorghum Breeding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
B. Plant Breeding in General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
XIII. References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Plant hormones have been known since 1928, and the literature concerning the influence of hormones on plant growth and development has
become voluminous (Leopold, 1964). Nevertheless, no mechanism of
genetic control over growth that involves hormones has been recognized
and reported. A dogma in the physiology of flowering is that the floral
stimulus is hormonal and that a vegetative growing point differentiates into
a floral bud whenever the floral hormone, that is produced in the leaves,
reaches the proper level at the growing point. Because the different maturity genotypes of sorghum initiate floral buds at different times, there is
reason to assume that the maturity genes, in some way, control hormone
Fifty years ago, Kiesselbach (1922) measured pith cells of parents and
hybrids of Zea mays L. and concluded that hybrid vigor resulted largely
from more rapid cell division in the apical meristem. Other data from corn
cited by Patanothai and Atkins (1971) and recent data from sorghum
seem to indicate the validity of that conclusion (Quinby, 1970; Patanothai
and Atkins, 197 1 ) . If a greater rate of cell division is the basis of hybrid
vigor, an understanding of the genetics and physiology of hybrid vigor
should contribute to an understanding of the genetic control of plant
The inheritance of time of floral initiation and duration of growth in
sorghum, Sorghum bicolor L. Moench, has been known since 1945
(Quinby and Karper, 1945; Quinby, 1966). More recently, Lane (1963)
studied the influence of light on flowering in sorghum and Miller et al.
(1968a) observed sorghum in monthly plantings throughout the year in
Puerto Rico. Based on the information from these studies and on some
knowledge of the genetic control of flowering in sorghum, it is now possible
to formulate some hypotheses concerning the physiology of flowering in
sorghum. Until now, genetics has contributed little to an understanding
of the physiology of flowering and of growth. It is the purpose of this chapter to present a working hypothesis to explain the genetic control of flowering and growth which can be subjected to experimentation. It is hoped
GENETIC CONTROL OF FLOWERING AND GROWTH IN SORGHUM
that the presentation of this model will stimulate an interest in the genetics
and physiology of growth and result in work that will prove or disprove
Genetics of Flowering
Four gene loci that control time of floral initiation and duration of
growth have been recognized in sorghum (Quinby and Karper, 1945;
Quinby, 1966). The four gene loci segregate independently. Although no
additional gene loci have been found, a number of recessive alleles at locus
1 (mal) and locus 3 (ma,)have been recognized. The assignment of dominance to late maturity was made because F, plants of certain MILO crosses
were late to flower, but the lateness was due to gene interaction or epistasis
rather than to dominance or recessiveness at one or more loci (Table I )
The maturity genes control leaf number because no leaves are laid down
in the meristem after a floral bud is initiated. Maturity genes also control
plant size, irrespective of growth rate, because the longer a plant grows
the larger it becomes. In the past, it seemed reasonable to attribute only
the initiation of the floral bud to the maturity genes (Quinby, 1967). It
has become apparent that the maturity genes influence rate of growth as
well as time of floral initiation because comparable leaves of plants of
different durations of growth differ in size (Quinby, 1972). Additional
functions must now be assigned to the maturity genes.
The continuous variation of maturity in sorghum is thought to exist because of alleles, both dominant and recessive, at the four maturity loci
and not to modifiers at other loci. The allelic situation has been discussed
previously (Quinby, 1967, 1 9 7 2 ~ ) .Most tropical varieties are dominant
at all four maturity loci and one reason for this conclusion is that a recessive at any locus causes temperate zone adaptation. Several single-gene
recessive, temperate varieties have been identified (Quinby, 1967). Most
of the 3000 temperate varieties in the world are recessive at locus 1 and
are dominant at the other three. Because mutations to temperate zone
adaptation must have occurred at many times in the past, recessive alleles
at locus 1 must be numerous. Only one recessive allele is known at locus
2 and only one at locus 4, and recessive allelic series are not known at
those loci. Five recessive alleles are known at locus 3, and three of the
five occurred in the United States since sorghum was introduced about
a century ago.
More than 100 tropical varieties have now been converted to temperate
zone adaptation by substituting recessive ma, for dominant Ma,. Because
J. R. QUINBY
Assumed Hormone Level and Duration from Time of Planting
t o Floral Initiation and Flowering of Maturity Genotypes
Days t o
genotype and hormone
Data from a June 1, 1972, planting at Plainview, Texas.
“A” indicates high, and “a” low, auxin level.
“G” indicates Iiigh, and “g” low, gibberellin level.
most tropical varieties are dominant at all four maturity loci, the converted
varieties must be of the genotype malMazMa,Ma, with the recessive at
the first locus the same in each. (Here, and henceforth in this chapter,
a single gene symbol per locus will be used to indicate homozygosity) .
Nevertheless, the converted varieties vary in flowering from 60 to 85 days
(Quinby, 1 9 7 2 ~ ) Apparently,
early-maturing tropical varieties are converted into early-maturing temperate varieties and late-maturing tropical
varieties into late-maturing temperate varieties. Because of the diversity
in time of flowering among the converted varieties, the dominants at maturity loci 2, 3, and 4 must differ among varieties. Dominant Mal from
lOOM M I L O causes later flowering in segregating populations than dominant Mal from Hegari. Furthermore, according to Haupt (1969), Wellensick (1965) found evidence in Pisum sutivum L. to indicate allelic series
GENETIC CONTROL OF FLOWERING AND GROWTH IN SORGHUM
in late-, intermediate-, and early-flowering varieties of that species. Because
dominant and recessive allelic series exist, it is not only the dominant or
recessive condition at each locus that is important in determining time of
floral initiation in sorghum; but, also, which allele is present at each locus.
The Floral Stimulus
The identity of the floral stimulus is the subject of much discussion in
the literature. The literature has been reviewed many times and recently
by Salisbury (1963), Lang (1965), and Evans (1969). Hamner (1940),
using data from soybeans (Glacine max (L) Merrill), concluded that
flower induction is governed by processes in the light, processes in the
dark, and processes from an interaction between the two. Chailakhian
(1961) suggested that the floral stimulus consists of two components and
that the long-day component is gibberellin. The short-day component has
not been identified but is probably auxin. Auxin is known to be involved
in the flowering process; is used to initiate floral buds uniformly in the
production of pineapples (Clark and Kerns, 1942) ; and interactions between auxin and gibberellin are well known (Leopold, 1964). The interaction of auxin and gibberellin in differentiation and growth of tissues has
been discussed by Evans (1969). Some of the reasons for the failure to
recognize auxin as a part of the floral stimulus are discussed in Section
XI. In the rest of this chapter, auxin will be assumed to be the short-day
component even though proof is lacking.
There is reason to think that a vegetative bud changes into a floral bud
whenever the floral stimulus reaches the proper level at the growing point.
Assuming that a combination of auxin and gibberellin is the flowering
stimulus, differences in time of floral initiation between maturity genotypes
shown in Table I could be interpreted to mean that the floral stimulus
accumulates at the growing point at different rates in different genotypes.
This should indicate that auxin and gibberellin are being synthesized in
the leaves at different rates in different genotypes and that there must be
some genetic mechanism to control the rate of synthesis of the two hormones. Dominant and recessive alleles at the maturity gene loci and gene
interaction appear to exercise this control.
Physiology of Flowering
Phytochrome appears to be a regulatory agent in the growth of all seed
plants (Hendricks and Borthwick, 1963) and is both a pigment and an
enzyme. Phytochrome exists in a far-red absorbing form (Pis,,) and a red
J. R. QUINBY
absorbing form (Paso) Because the sunlight that reaches the earth through
the atmosphere contains both red and far-red wavelengths, plants growing
in sunlight contain both forms of phytochrome in about equal amounts.
During the night, and after conversion of P730 to P660, most of the
phytochrome is in the P 6 6 0 form. P,,, was considered by Hendricks and
Borthwick (1963) to inhibit flowering in short-day plants but to promote
flowering in long-day plants. Phytochrome is generally considered to be
involved, in some way, in controlling the reactions that lead to flowering.
Some understanding of the genetics of flowering in sorghum has prompted
some surmises regarding the role phytochrome plays in controlling time
of floral initiation.
Borthwick and Cathey (1962) have pointed out three possibilities to
explain the long time that phytochrome must be in the P 7 3 0 form to
prevent flowering of some varieties of Chrysanthemum. They concluded
that different varieties might ( 1 ) contain different absolute amounts of
phytochrome; (2) have different rates of conversion of PTs0to P660; or (3)
have different levels of PTs0that would inhibit flowering. Differences in
sensitivity to inhibition by phytochrome is a fourth possibility and is akin
to the third possibility that different levels of P,30 might inhibit flowering.
The MILO maturity genotypes of sorghum have contributed greatly to
an understanding of the genetics of flowering in sorghum. Their origin
appears in Section X, A. Lane (1963) determined that the MILO maturity
genotypes do not differ in phytochrome content or rate of conversion of
P,,, ,to Po60 in darkness. It seems logical, therefore, to assume that
dominant and recessive maturity alleles cause differences in sensitivity
to inhibition by phytochrome.
Assuming that the floral stimulus consists of auxin and gibberellin,
P 7 3 0 appears to inhibit the synthesis of auxin during daylight and P660 to
inhibit synthesis of gibberellin during the night. Because of these inhibitions,
auxin is synthesized largely during the night and gibberellin largely during
daylight. P,,, converts to P 6 6 0 in darkness and, after the conversion to
P660, synthesis of auxin is not inhibited for the rest of the night. P 6 6 0
appears to inhibit the synthesis of gibberellin during the night; and, if
this is true, P660 is not inactive as argued by Hendricks (1960); but is
inhibitory to flowering as suggested by Takimoto (1969).
Alleles at loci 1 and 2 appear to cause differences in sensitivity to
inhibition by PTs0,and alleles at loci 3 and 4 to cause differences in
sensitivity in inhibition by P,,,. It is likely that recessive alleles at loci 1
and 2 allow the synthesis of some auxin during the day to supplement
that produced during the night despite the presence of PTs0during the day.
Likewise recessive alleles at loci 3 and 4 probably allow the synthesis of
more gibberellin during the day than dominant alleles.
GENETIC CONTROL OF FLOWERING AND GROWTH IN SORGHUM
The assumed effects of dominant and recessive alleles on duration from
planting to floral initiation and to flowering and on hormone levels are
shown in Table I. Dominant alleles result in low levels of auxin and gibberellin and recessive alleles in higher levels of the two hormones. 100M,
that is dominant at all four loci, is assumed to synthesize small amounts
of both hormones because of sensitivity to inhibition by Ps30and by P G G o .
80M is recessive at ma2, is assumed to synthesize more auxin than 100M,
and flowered earlier. SM80 (rnalma2)is assumed to synthesize more auxin
than either 80M or 100M and flowered earlier than either. However,
SMlOO (malMaz) is assumed to synthesize less auxin than SM80 but
flowered earlier indicating that too much auxin, as well as too little delays
floral initiation. This is not an isolated case because Early Kalo (malMa,)
flowers earlier than Kalo (malmar) (Quinby, 1967) and KMH-1
(Malma2)flowers earlier than KMH-2 (rnalrna,) (Quinby, 1 9 7 2 ~ )SM90
(ma1Ma2)flowered earlier than 60M (Malmaz) and this is assumed to
indicate that recessive ma, and recessive ma, result in different amounts
All ten maturity genotypes of MILO are dominant at Ma,. Recessive
mu3 is assumed to result in more gibberellin than dominant Ma, and all
recessive ma3 genotypes are earlier to flower than the dominant Mas counterparts. 38M and 44M differ from the other eight maturity genotypes of
Milo in being recessive ma,R. Presumably, these two genotypes synthesize
more gibberellin than the others. Both 38M and 44M in addition to being
early to flower, have long narrow leaves, slender stems, and long internodes. Even the first four embryonic leaves are long and narrow and these
two genotypes can be recognized in the seeding stage.
The early flowering of 44M and 38M appears to be caused by a high
rate of synthesis of gibberellin during the day that results from a low level
of sensitivity to inhibition by P,;co.44M and 38M initiate floral buds before SM60, and the presence of recessive ma3 appears to result in too little
gibberellin for earliest floral initiation when the level of auxin is at the
level resulting from either Ma,ma, or ma,mu,. But when recessive masR
is present, the higher level of gibberellin that is synthesized matches the
high level of auxin; and early floral initiation results. Unlike auxin, high
levels of gibberellin appear not to delay flowering.
It is apparent that the maturity genes, by controlling hormone levels,
control plant growth until a floral bud initiates and begins rapid growth.
It is not apparent when the maturity genes cease to function, but it is obvious that a second genetic control of growth begins to function as soon
as the developing inflorescence becomes large enough to synthesize appreciable amounts of auxin. The genes that begin to function then have
been recognized as height genes.
J. R. QUINBY
Genetics and Physiology of Cell Elongation
Dwarfing genes have been useful in sorghum because they favor me-.
chanical harvesting and reduce lodging. Four, nonlinked, brachytic
dwarfing genes are known (Quinby and Karper, 1954). Tallness is partially dominant to shortness. Five height classes have been recognized. The
tall or zero-dwarf class may grow to be as tall as 3 or 4 m and 4-dwarf
plants may be as short as 1 m. Plants of the same height genotype may
differ in height by as much as 10 to 50 cm, and these differences are
thought to be due to allelic series at the four height loci (Quinby and
Schertz, 1970). A dwarfing gene may reduce height by as much as 50
cm or more but the reduction is less if there are dwarfing genes at other
height loci. Differences between the height of 3-dwarf and 4-dwarf genotypes may be as little as 10 or 15 cm.
The first maturity gene ma, is linked to the second height locus dw,
but crossovers occur (Quinby and Karper, 1945). All height genotypes
occur in all maturity genotypes and it is assumed that the height genes
and maturity genes are at different loci. However, each dwarfing gene
might be linked to one of the maturity genes, but the possible linkage between them has not yet been established.
OF CELL ELONGATION
The usual sites of hormone production in vascular plants are meristems
and enlarging tissues (Leopold, 1964). It is reasonable to assume, therefore,
that the developing inflorescence and expanding leaves synthesize auxin
and that the exposed leaves also synthesize gibberellin. The presence of
a floral, rather than a vegetative, bud has profound effects. Soon after a
floral bud is initiated, the grand period of growth begins, internode elongation accelerates, tillering is suppressed, and the inflorescence begins rapid
growth. It seems likely that the maturity genes continue to control synthesis
of auxin and gibberellin in the leaves as long as they are growing. A leaf
blade is almost fully expanded by the time the ligule can be seen due to
the elongation of the internode below the node of leaf sheath attachment.
After all the leaves are fully grown and exposed a few days before the
head is in the boot, the leaves appear to no longer be the major site of
hormone synthesis. The expanding inflorescence then becomes the major
site of hormone synthesis and the height genes appear to be in control
of hormone synthesis.
The reduction of cell elongation in the internodes of short-statured
GENETIC CONTROL OF FLOWERING AND GROWTH IN SORGHUM
plants probably results from inhibition to cell elongation due to an excess
of auxin and too little gibberellin. During the period from floral initiation
until emergence from the boot, the developing panicle is protected from
sunlight and the phytochrome in the panicle should be in the Pfjo0form.
If this is true, the developing panicle, in the absence of Pi,,, should synthesize auxin during the day as well as during the night, the amount depending on the height genotype. However, developing panicles of all genotypes should synthesize little or no gibberellin because of inhibition by
P,,,. The upper, expanding leaves, however, should synthesize both auxin
The height genes are assumcd to control levels of auxin and gibberellin,
and it is reasonable to assume that inhibition by PTs0and P,,, is the
mechanism involved. The hormone levels that are assumed to result from
dominant and recessive alleles at the four height loci are shown in Table
11. Alleles at height gene loci 1 and 2 are assumed to control levels of
auxin and alleles at loci 3 and 4 to control levels of gibberellin. Four dominant genes for height result in a low level of auxin and a high level of
gibberellin. For this reason, even though both maturity and height genes
control hormone levels, different functions must be assigned to the height
genes at loci 3 and 4 as compared to the maturity genes at loci 3 and
4. Apparently, dominants Mu, and Ma, result in a low level of gibberellin
and dominants Dw3 and Dw, in a high level of gibberellin.
Schertz et al. (1971) concluded that the dw, height locus in sorghum
may regulate the expressed level of peroxidase activity. There is no agreement about the function of peroxidase in plants at present but peroxidase
is generally thought to be involved in interactions with gibberellin or auxin.
How peroxidase may be involved in the control of plant growth is not
Several brachytic dwarfing genes exist in Zea mays L. and applications
of gibberellin in microgram amounts to plants of five of them result in
plants of tall height (Phinney, 1961). The probable explanation of this
response to gibberellin application in corn is that short plants contain too
much auxin for normal cell elongation unless some gibberellin is added.
This is the same response recognized in the early floral initiation in
sorghum that results from the presence of recessive
and greater synthesis of gibberellin. Presumably, any sorghum genotype that is high in
auxin content or is low in gibberellin content would grow taller if gibberellin were applied at the proper time and at appropriate concentrations.
C. HEIGHTGENESAND LEAFSIZE
Schertz (1973) has compared leaves of a doubled haploid of 4-dwarf
SA403 with leaves of a 3-dwarf obtained from a tall mutation of the same