Flowers exist in a variety of arrangements on plants but basically they can be considered as occurring singly or clustered in inflorescences. When they occur singly the flower usually arises by the conversion of a vegetative SAM into a flower meristem. In the case of inflorescences, the vegetative SAM is converted to an inflorescence meristem and then flower meristems can arise directly on the inflorescence with no preceding vegetative period. Thus these meristems go through no change to become flower meristems.
Cytological changes and Morphological changes
Despite the developmental differences between single flowers and inflorescences, the changes in the SAM from vegetative to floral are remarkably similar. One of the earliest events in the transition from vegetative to floral is an increase in mitotic activity, particularly in the central zone. The distinction between the central zone and peripheral zone is lost as mitotic activity becomes uniformly high. Associated with the increased mitotic activity is an increase in meristem size and usually a change in shape.
The phyllotaxy changes as the meristem becomes floral. Commonly the phyllotaxy becomes whorled (ie. >2 primordia are initiated at the same node). Sometimes the first set of organs (the sepals) initiates sequentially, but in subsequent organs (petals and stamens) organs of a whorl initiate simultaneously. In maize the meristem shifts from alternate to spiral phyllotaxy as it assumes inflorescence identity.
Internode elongation is minimal in flowers resulting in the compact arrangement of organs.
In flowers, the meristem changes from an indeterminate to determinate mode of growth. In other words the meristem produces a set number of organs and then stops. Inflorescence meristems can be determinate or indeterminate depending on the species.
The general sequence of events in a flower meristem is that the meristem initiates sepal primordia, either sequentially or simultaneously. Then petal primordia are initiated, usually simultaneously, in positions that alternate with the sepals. Stamen primordia arise shortly after or simultaneous with the petals, and in alternate positions to them. Finally the carpels initiate in the center of the meristem to give rise to the pistil. As the different organs initiate the meristem gets progressively smaller. After the carpels initiate from the center, the meristem is "used up".
SAM cell layer contributions to floral organs
Periclinal chimera fate mapping experiments have shown that the cell layers of the floral meristem have the same basic relationships with floral organs as with leaves. That is L1 gives rise to epidermis, L3 to core tissues and L2 to intermediate tissues of sepals, petals etc. The key point about flower lineage studies is that in majority of cases, germ cells are derived from the L2. Thus sexually produced progeny from a G-W-G type chimera (with an albino L2 layer) will all carry the albino marker. (Note: because a gamete is a single cell, it is not possible to propagate the chimeral arrangement of cells sexually) This has implications for genetic engineering because for a gene to be heritably introduced into a plant it must be incorporated into the L2.
Genetic control of flower development
A collection of mutants has been generated that provides considerable information about the genetic control of flower morphogenesis. The 2 plants that have been most informative are arabidopsis and Antirrhinum (snapdragon). Essentially 2 classes of genes have been identified: meristem identity and organ identity genes. The same genes are involved in both plants. In addition, snapdragon has genes regulating the bilateral symmetry of the flowers (arabidopsis flowers are radially symmetrical). For simplicity we will only discuss arabidopsis except for the bilateral symmetry.
There are three types of meristems involved in flower development: vegetative, inflorescence and flower. As such there are 2 different transitions involved (veg to infl., infl. to flower). Genes have been identified that control both these transitions.
The phenotype of a mutant called embryonic flower (emf) suggest that a floral meristem is a default pathway. emf mutants form an inflorescence meristem during embryogenesis. Upon germination they form an inflorescence with no vegetative stage. Because this is a recessive loss of function mutation, it suggests the normal gene function is to suppress floral identity and allow vegetative meristem identity. This gene has not been cloned yet. Presumably the transition of a normal vegetative meristem to an inflorescence involves the suppression of the normal EMF gene.
The identity of floral meristems is controlled by the opposing action of 2 sets of genes. A mutation called terminal flower (tfl) results in the conversion of the inflorescence meristem to a flower. Thus TFL is required to promote inflorescence or suppress flower meristem identity in the inflorescence meristem. Two mutations called leafy (lfy) and apetala1 (ap1) result in the production of inflorescences in place of flowers. These genes are normally required to promote flower or suppress inflorescence identity. The balance between these 2 sets of genes controls meristem identity.
TFL is expressed in the corpus of inflorescence meristems. LFY is normally expressed throughout flower meristems. In the tfl mutant LFY is expressed throughout the apical meristem (which should be an inflorescence meristem but has been converted to a flower meristem). Therefore TFL in the corpus normally represses LFY expression in the tunica and corpus.
LEAFY is required for the expression of downstream flower organ identity genes.
Mutations in genes controlling the identity of flower organs results in homeotic conversions of one organ to another. There are 3 classes of homeotic genes involved in flower organ identity and each class affects the identity of 2 adjacent whorls of organs. The classes of genes have been designated A, B and C. The A class gene is called apetala2 and ap2 mutations cause the sepals and petals to form carpels and stamens respectively. 2 genes, apetala3 and pistillata belong to class B and mutations in either ap3 or pi cause petals and anthers to form sepals and carpels respectively. The class C gene is called agamous and ag mutations cause the stamens and carpels to form petals and sepals respectively. In addition, ag mutant flowers are indeterminate and reiterate this pattern of floral organs.
A model was proposed where the overlapping action of these genes would specify organ identity. A function would be present in the first 2 whorls, B function in whorls 2 and 3, and C function in 3 and 4. A alone would specify sepals, A+B=petals, B+C=stamens and C alone specifies carpels. In addition, A and C would negatively regulate each other (be mutually exclusive) so that in the absence of A function, C function would be expressed in whorls 1 and 2, and in C mutants, A function would be expressed in whorls 3 and 4. This model accounts for all the basic mutant phenotypes observed. Recent models also include a D function which is involved in ovule development.
Most of these genes (except TFL, AP2 and LFY) contain a conserved DNA-binding motif called the MADS box. This motif appears to have evolved a similar function as the homeobox which is important in specifying organ identity in animals.
SAM cell layer interactions in flower development
Periclinal chimeras in which different cell layers contain genotypes that confer different floral characteristics have shown that cell layers interact and cooperate in the development of flowers. In a classical work the flowers of a periclinal chimera between 2 species of camellia were examined. The L1 consisted of a species with normal flowers ( fertile and containing all the normal organs) while the L2 and L3 consisted of a species with sterile flowers containing no stamens or carpels. The chimera formed functional pollen even though the pollen is derived from the L2 of the genotype that normally does not form reproductive structures. Thus the fertile L1 induced reproductive development in tissues that lacked the genetic capacity to induce their own reproductive development.
Another example is in tomato where periclinal chimeras were made with a mutant called fasciated. This mutant makes too many carpels. All chimeras with fas in the L3 produced extra carpels but the presence of fas in the L1 or L2 had no effect. Thus L3 specified carpel number.
A third example involves the floricaula mutation of snapdragon (which corresponds to the leafy mutation of arabidopsis). Wild type Flo+ is required for expression of a downstream organ identity gene called deficiens (def). In chimeras where Flo+ was only present in one of the apical cell layers, def mRNA was expressed throughout the meristem indicating that somehow Flo+ expression in one cell layer was inducing def expression in other cell layers.
Finally there is the example of TFL that we already discussed.
Protein trafficking is likely to be involved in some of these interactions. Two snapdragon genes called GLO and DEF (homologs of PI and AG3) function cell autonomously in the L1 of periclinal chimeras but non-cell autonomously in the L2. That is that wild type L2 can rescue mutant L1 but NOT vice versa. These transcription factors can apparently be transported between cells through the plasmodesmata (similar to gap junctions of animal cells). GLO can traffic from L2 to L1 but cannot from the L1 to L2. Thus trafficking is polar.
Flowers that contain both male and female reproductive organs are called perfect (bisexual). Flowers that lack one or the other are called imperfect (unisexual). The developmental decision of which sexual organs to form (sex determination) can occur at several points. Some plants contain all male or all female flowers and therefore the sex determination event occurs at the whole plant level. Other plants such as maize have male and female flowers located on separate inflorescences. In the majority of cases, unisexuality arises by the selective abortion of one type of reproductive organ. For example, flowers on the maize ear initiate both stamens and a pistil but at an early stage the stamens abort. Conversely on the tassel the pistil aborts. There are several mutations called tasselseed that result in female flowers developing on the tassel. ts2 has been cloned and encodes a short chain alcohol dehydrogenase. Normal expression of ts2 in the pistil primordium of male flowers induces apoptosis (programmed cell death) leading to pistil abortion.
No mutants are known that convert the ear to completely male flowers. GA deficient mutants form bisexual flowers on the ear indicating that GA is important for suppressing stamen development on the female inflorescence.
Many flowers have fused floral organs. It is common for petals or sepals to be fused into a tubular structure and the pistils of many flowers consist of fused carpels. The fusion events can be either congenital (occurring before or during organ initiation) or postgenital (occurring after organ initiation).
Postgenital carpel fusion has been studied in a plant called Catharanthus roseus (Matagascar periwinkle). The process involves the dedifferentiation of the carpel epidermis and is mediated my a diffusible substance. Plant epidermis is usually resistant to the formation of organ fusions so this represents a highly specialized process. Initially the separate carpels are completely covered with an epidermis. As the growing organs come in contact, the epidermis in the region of contact dedifferentiates. Insertion of a permeable barrier between the organs allows the dedifferentiation event to proceed but a non-permeable barrier will block the event. Therefore a diffusible signal mediated the dedifferentiation. If an agar impregnated porous barrier is inserted between the carpels, the agar can be removed and applied to another portion of the carpel epidermis not involved in the contact. Therefore the whole carpel epidermis is competent to respond to the dedifferentiation signal, but only the area of contact receives or produces the signal. Application of the agar to other organs such as anthers does not cause epidermal dedifferentiation so competence to respond to this signal is specific to carpels. The molecular basis of this process is not known.