| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Development and Morphogenesis |
Department of Biology, University of Missouri—St. Louis, 8001 Natural Bridge Road, St. Louis, Missouri 63121 USA
Received for publication December 13, 2001. Accepted for publication April 9, 2002.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Key Words: Cenchrus • development • inflorescence • morphology • Pennisetum • phylogeny • Poaceae • Setaria
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
). There are approximately 10 000 species and 600–700 genera (Clayton and Renvoize, 1986
; Watson and Dallwitz, 1992
), and taxa have been morphologically distinguished one from another primarily by differences in inflorescence and spikelet morphology, leaf and culm anatomy, and photosynthetic pathway (Clayton and Renvoize, 1986
). Of these, the most obvious morphological distinctions among many genera and species are differences in inflorescence architecture, ranging from spikes to many-branched panicles (Clayton and Renvoize, 1986
). These differences can be characteristic of certain tribes (Clayton and Renvoize, 1986
), but analysis of the phylogenetic distribution of inflorescence variation in the grasses suggests that different forms have arisen independently many times (Kellogg, 2000a
). Homoplasy in the distribution of inflorescence forms across the grass phylogeny obscures the mechanisms by which inflorescence diversity has arisen. The inflorescence is the outcome of a temporal and spatial pattern of gene expression, which drives the process of morphological development. Named inflorescence architectures, such as panicle or spike, have no particular meaning in and of themselves, but are merely convenient descriptors for the outcomes of these developmental processes; the mature inflorescence is an historical summary of development, rather than a fixed or ideal type. Therefore, it may be preferable to look for homologies at the level of changes in development rather than in morphology of mature inflorescences.
A study of inflorescence development across the entire grass family would be impracticable, but analysis of a smaller group of taxa may lead to results that are applicable to the rest of the family. We have chosen to concentrate on the so-called "bristle clade" (Paniceae), which has 25 genera and approximately 310 species, 110 of which are in the genus Setaria. The bristle clade has been found to be monophyletic in morphological (Zuloaga, Morrone, and Giussani, 2000
) and molecular studies (Gómez-Martinez and Culham, 2000
; Giussani et al., 2001; A. N. Doust and K. Giljum, unpublished data). It includes all panicoid species in which some inflorescence branch meristems are converted to setae or bristles. These are novel structures that have some aspects of spikelet identity and some of branch identity (see below). Homology of the inflorescence bristles was not recognized by previous workers, who placed members of the clade in different subtribes (Clayton and Renvoize, 1986
). Members of the bristle clade have inflorescences that may be elongate or condensed and may have few or many orders of branching (Figs. 1–6). Setaria, Pennisetum, and Cenchrus are the three most species-rich genera within the bristle clade. By concentrating on these genera, we hope to discern changes in developmental processes that can lead to diversity in mature inflorescence architecture.
|
; Schuman, 1890
; Sohns, 1954
, 1955
; Pensiero, 1995
, 1999
), modified branches (Schuster, 1910
; Arber, 1931
, 1934
; Butzin, 1977
), or as being derived from both these sources (Vegetti and Pensiero, 1990
). The extent to which bristles and spikelets share common developmental pathways will have to be answered on the level of gene expression and interaction, but analysis of inflorescence development provides the context and observations that will allow this question to be properly formulated and answered.
This paper presents a molecular phylogenetic analysis of the bristle clade, which is then used to orient investigation of the developmental variation leading to inflorescence diversity. Characters developed from comparisons of the developmental sequences are optimized onto the molecular phylogeny to trace patterns of character evolution. Thus, the phylogeny is based only on molecular data sets, as opposed to a total evidence approach (Deleporte, 1993
; Grandcolas et al., 2001
), so that the molecular phylogeny is used as an a priori standard against which to assess developmental evolution. The chloroplast DNA markers used are the trnL intergenic spacer region (Taberlet et al., 1991
) and the NADH dehydrogenase gene (ndhF) (Olmstead and Sweere, 1994
; Clark, Zhang, and Wendel, 1995
); these have previously been used to construct phylogenies of the grass tribes Andropogoneae and Paniceae (Spangler et al., 1999
; Gómez-Martinez and Culham, 2000
; Giussani et al., 2001
). Integration of a robust molecular phylogeny, developmental analysis, and optimization of developmental characters onto the molecular phylogeny will enable a more detailed understanding of patterns of inflorescence evolution in the bristle grass clade and help in the larger task of elucidating the patterns of inflorescence evolution in grasses.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
). A larger set of outgroup taxa was originally used (all those in the x = 9 Paniceae clade described by Giussani et al., 2001
) but these were pruned down to five as choice of outgroups had little effect on the topology within the bristle clade. Ingroup taxa were chosen to show all of the different inflorescence forms known in the three genera (except for the extremes of spinescence found in some species of Cenchrus). All taxa received as seed were grown to flowering, vouchered, and their identity checked using keys in Rominger (1962)
, DeLilse (1963)
, Correll and Johnston (1970)
, and Cope (1982)
. DNA was extracted using the protocol described in Giussani et al. (2001)
. The trnL intron and the trnL-trnF spacer were amplified via the polymerase chain reaction (PCR) using primer pairs C–F, designed by Taberlet et al. (1991)
, and primers C, D, E, and F were used for sequencing; ndhF was amplified using PCR in two overlapping fragments: 5F/1318R and 972F/2110R. Ten sequencing primers were used, nine designed by Olmstead and Sweere (1994)
and one (1821R) designed by Clark, Zhang, and Wendel (1995)
. The PCR products were cleaned with the QIAquick PCR purification kit (Qiagen, Valencia, California, USA), quantified by comparison with DNA of a known concentration (pGEM 10 and 25 ng; Applied Biosystems, Foster City, California, USA) and fluorescence-labelled using the "Big Dye" (Applied Biosystems) cycle sequencing protocol. Both forward and reverse strands were sequenced on an ABI 377 automated sequencer. Contig assembly and editing of sequences used Sequencher, version 3.1 (Gene Codes Corporation, Ann Arbor, Michigan, USA). Sequences were a minimum of 85% double stranded. All sequences were aligned in Clustal W (Thompson, Gibson, and Higgins, 1994
) followed by manual adjustments in Se-Al version 1 (Rambaut [1996], available via file transfer protocol (FTP) from ftp://evolve.zo.ox.ac.uk/packages/Se-Al/). Genbank accession numbers are archived on the BSA web site at http://ajbsupp.botany.org/v89.
Phylogenetic analyses were performed with a maximum parsimony algorithm, treating character states as unordered. Analyses were conducted using PAUP* version 4.0b4a (Swofford, 1999
), with heuristic searches, tree bisection-reconnection (TBR) branch swapping, 100 random addition sequence replicates, and gaps treated as missing data. Full heuristic bootstrap analyses were conducted using 1000 replicates, with remaining parameters identical to those used in the parsimony analysis (Felsenstein, 1985
).
Congruence between the two chloroplast data sets was tested using the incongruence length difference (ILD) test with a significance level of P < 0.01 (Farris et al., 1994
; Cunningham, 1997
), implemented in PAUP*. An ILD test was also used to investigate congruence between the molecular and developmental data sets. Templeton tests (Templeton, 1983
) were used to investigate differences between molecular and developmental data sets by testing the length of a molecular tree constrained by the topology sugggested by the developmental data against an unconstrained molecular tree.
Developmental characters, defined by comparison of developmental sequences (see below), were optimized on each of the most parsimonious trees using MacClade 4.0 (Maddison and Maddison, 2000
). In cases where character states were equivocal at a node, all equally parsimonious reconstructions under both ACCTRAN and DELTRAN options were examined.
Morphological and developmental analysis
Developmental data were obtained for 15 of the 23 species sequenced (see supplementary information at http://ajbsupp.botany.org/v89). Most species of Setaria, Pennisetum, and Cenchrus used in this study were obtained as seed from the United States Department of Agriculture (USDA) and grown in the glasshouse at the University of Missouri—St. Louis. One species obtained from the USDA was labelled as Setaria paniculifera but was reidentified by us as S. poiretiana. It is included in this analysis as S. poiretiana A, while a second but somewhat different accession from USDA labelled as Setaria poiretiana was included as S. poiretiana B. Inflorescences of Setaria palmifolia were obtained from plants in the Missouri Botanical Garden. Inflorescences were harvested at all stages of development, dissected while fresh, and fixed either in PFA/glutaraldehyde (phosphate buffered 4% paraformaldehyde followed by phosphate buffered 4% glutaraldehyde) or FAA (formalin-acetic acid-70% ethanol, 10 : 5 : 85 volume/volume). Some of this material was dehydrated using an ethanol series, critical point dried in an SPI Jumbo critical point drier (Structure Probe, West Chester, Pennsylvania, USA) and sputter-coated with gold in a Polaron E5000 sputter coater (Quorum Technologies, Hailsham, UK). Specimens prepared in this manner showed significant charging of bristles in the electron beam and therefore the rest of the specimens were rehydrated and post-fixed with osmium tetroxide (OsO4) using the OTOTO method (Murphy, 1978
). This treatment consisted of overnight fixation in phosphate buffered 1% OsO4, six deionized water washes, 30 min fixation in freshly made and filtered 1% thiocarbohydrazide (TCH), six deionized water washes, 1 h in 1% OsO4, six washes in deionized water, 30 min fixation in freshly made and filtered 1% TCH, six deionized water washes, 1 h fixation in 1% OsO4, followed by a final six washes in deionized water. This treatment ensured that all tissue of each specimen was fully electron conductive. The OTOTO specimens then underwent dehydration and critical point drying, but were not sputter-coated with gold. Both types of specimens were imaged in an Hitachi S450 scanning electron microscope (SEM) (Tokyo, Japan) at 20 kV.
The SEM images were used to construct developmental sequences for each taxon and to compare inflorescence development between taxa. The start of the inflorescence development pathway was taken as the transformation of the hemispherical vegetative meristem into an elongated inflorescence meristem (Evans and Grover, 1940
), and the end as the point at which all inflorescence axes had ceased meristematic activity and had differentiated. Comparison of developmental sequences allowed identification of points in the developmental continuum where differences between taxa exist. Character states were then defined as those variants that could be consistently distinguished from one another.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
|
), three of which are native to the Old World. Of these, S. italica is the cultivated grain crop foxtail millet, while S. verticillata and S. viridis are weeds of cultivated crops. The fourth species, S. grisebachii, is New World in distribution. Setaria clade 2 consists of the rest of the species of Setaria. Setaria barbata, S. poiretiana A, S. palmifolia, and S. poiretiana B are all from section Ptychophyllum (Rominger, 1962
), while S. geniculata and S. parviflora are from section Setaria. The two accessions of S. poiretiana have pronounced sequence differences and fall into two different subclades. The four representatives of Ptychophyllum were checked by re-extracting DNA and re-sequencing the ndhF gene; an identical result was found. The third clade consists of all species of Pennisetum and Cenchrus, with Cenchrus forming a subclade within a paraphyletic Pennisetum.
Development
The transition to reproductive growth, the initiation of the main branch axes, and the later development of the spikelets were similar in all species of Setaria, Pennisetum, and Cenchrus investigated. Other aspects of development vary widely, and therefore the development of the three clades will be discussed separately.
For the purposes of this discussion we define the main inflorescence axis as the primary (1°) axis, the major branches as secondary (2°) axes, branches on these as tertiary (3°) axes, and so on. The 1° axis of the inflorescence is that created by the activity of the apical meristem as it initiates 2° axis branch primordia. The structure of the 2° axis and the primordia it produces is more or less repeated throughout the inflorescence and is referred to as the secondary axis complex (2°AC).
Setaria clades 1 and 2: early development
Early inflorescence development in all species of Setaria is similar. The beginning of inflorescence development can be identified by the elongation of the shoot meristem as it undergoes the transition from vegetative (Fig. 8) to reproductive growth (Fig. 9). The shoot meristem initiates 2° axis primordia that become the main branches from the central axis. These primordia are arranged polystichously, in whorls or spirally (Figs. 10, 11). Species differ in the number of 2° axes produced; some varieties of Setaria viridis have only approximately 25 2° axes, while others can have up to 200 such axes. The number of 2° axes in Setaria italica is high but falls within the range of that of S. viridis. Although numbers of 2° axes are affected by growing conditions, differences between accessions are maintained when plants are grown under the same levels of temperature, moisture, fertilizer, and light.
|
|
The number of orders of branching in each 2°AC varies within the inflorescence, with basal axes having the most orders of branching. The 2°ACs near the apex of the inflorescence may have only one or two orders of branching, and the very uppermost branches may themselves differentiate into spikelets and thus produce no higher orders of branching at all (Fig. 17). Differentiation of branch axes begins at the apex of the inflorescence, so that spikelets and bristles are differentiating in 2°ACs near the apex at the same time as 3° and 4° axis primordia are produced in basal 2°ACs (Fig. 18). The species illustrated (Fig. 18) is actually from Setaria clade 2, but all species of Setaria share this basipetal pattern of inflorescence development.
|
|
, 1814
; Clayton and Renvoize, 1986
) (Figs. 23, 24). At maturity the spikelets disarticulate above the level of insertion of the bristles and fall individually, leaving the bristles attached to the inflorescence. In S. italica (foxtail millet) spikelets do not fall from the inflorescence, a trait associated with domestication. Axes that develop into bristles are variously ornamented, with prickle hairs in antrorse or retrorse orientation. The meristems at the tips of the bristles may degenerate or may form abortive spikelets (Figs. 25, 26). Rarely, these bristle-borne spikelets may produce viable seed (Ayyangar, Narayanan, and Rao, 1933
; A. N. Doust, greenhouse observations on Setaria italica x S. viridis cross).
|
|
|
|
|
|
|
In both Pennisetum and Cenchrus, higher order axes do not elongate as much as lower order axes, leading to two distinct size classes of bristles at maturity (Figs. 47, 48). This contrasts with Setaria where all bristles elongate equally.
In all species examined, elongation of the primary axis occurred late in development, after glume and lemma formation were nearly complete, and more or less simultaneous with the elongation of the culm below the inflorescence. The extent of elongation was variable among species (Figs. 1–6), with the greatest elongation occurring in S. palmifolia, S. poiretiana B, S. poiretiana A, and S. barbata.
Developmental differences among taxa
Comparison of developmental sequences yielded 12 characters whose states varied between taxa (Table 1). In most cases both meristic and quantitative characters fall into discrete character states. Some potential characters, such as number of 2° axes, were not coded or optimized onto the phylogeny, as numbers were too variable within and among species. Uncertain relationships among the outgroups make character polarization ambiguous, and therefore it is not possible to assign ancestral states to the characters.
|
The bristle clade as a whole is defined by three developmental characters: greater than five orders of axes (no. 1), primordia that develop into bristles (no. 5), and relative lack of elongation of the 1° axis (no. 8) (Fig. 49), although two of these (nos. 1 and 8) reverse in a subclade of Setaria clade 2. The tree shown places Setaria clade 1 as sister to all other taxa in the bristle clade, with Setaria clade 2 and Pennisetum/Cenchrus forming a monophyletic group. No morphological character is shown as supporting this group, although one optimization of character no. 6 places the change from state 0 (initiating equal numbers of spikelets and bristles) to state 1 (more bristles than spikelets initiating) at this point. That optimization also requires a reversal in this character in the subclade containing S. palmifolia and S. barbata. In the optimization shown (Fig. 49), state 1 of character no. 6 appears twice, rather than once as an origin and once as a reversal. Setaria clade 1 is supported by the character of suppression of the spikelets on higher order axes late in development (no. 7). Setaria clade 2 is not supported by any developmental character, although its subclades have support. The subclade of S. geniculata and S. parviflora is supported by the change from equal numbers of bristles to more bristles being initiated than spikelets in early development (no. 6). The subclade of S. palmifolia and its allies is supported by reversals of two of the characters that define the bristle clade, having four orders of axes (no. 1), and an elongated primary axis (no. 8). Two gains define this subclade: an increase in the number of 3° axes (no. 2) and the occasional presence of polystichously arranged 3° axes on basal 2° axes (no. 3). The Pennisetum/Cenchrus clade is defined by reduction of the internode on the secondary axis (no. 9), the reduction of internodes on other axes (no. 4) with the resultant appearance of a central axis surrounded by concentric rings of bristles, and the differential elongation of the bristles at maturity (no. 10), as well by a change from equal numbers of bristles to more bristles than spikelets being initiated in early development (no. 6). Cenchrus is distinguished from Pennisetum by bristles that are partially fused (no. 12). Within Cenchrus, species can have bristles that are flattened or more or less terete (no. 11).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
). The two Setaria clades do not fit previous assessments of relationships within the genus; clade 1 contains species from section Setaria only, whereas clade 2 contains species from both section Setaria and section Ptychophyllum (Rominger, 1962
). Rominger placed a few species in other sections and considered them morphologically similar to section Ptychophyllum, but these were not included in this study.
In Setaria clade 1, the sister species to S. italica is S. verticillata, although with only moderate bootstrap support (62%). These two plus S. viridis make up a strongly supported subclade within clade 1 (100% bootstrap). All three species are Old World in origin, with both S. viridis and S. verticillata being troublesome weeds worldwide (Rominger, 1962
). There are reports of hybrids between S. viridis and S. verticillata (Cope, 1982
), and it is likely that these two plus S. italica form part of a hybrid complex. Setaria italica is commonly regarded as a domesticated variant of S. viridis (Wang et al., 1998
), and both share similar genome structure (Benabdelmouna, Abirached-Darmency, and Darmency, 2001
; Benabdelmouna et al., 2001
). Setaria verticillata has been shown to be an allotetraploid, with one of the parental genomes being similar to that of S. italica and S. viridis (Benabdelmouna, Abirached-Darmency, and Darmency, 2001
; Benabdelmouna et al., 2001
). The chloroplast DNA copies recovered in this study are most probably from that genome. Sequences of the nuclear gene waxy (A. N. Doust and K. Giljum, unpublished data) also point to the presence of two dissimilar genomes within S. verticillata, with only one copy closely related to the genome of S. italica and S. viridis. Setaria grisebachii is the other species in Setaria clade 1, although its placement there is not strongly supported. It is an annual weedy New World species, and its inclusion in clade 1 will need to be verified with other data.
Setaria clade 2 has strong bootstrap support (98%) and contains a number of well-supported subclades. The subclade of S. geniculata and S. parviflora consists of plants with condensed inflorescences and strap-shaped leaves. In these respects they are morphologically similar to the Setaria species of section Setaria in clade 1, and Rominger (1962)
included them in this section also. More recent treatments (Pensiero, 1999
) treat S. parviflora as being synonomous with S. geniculata, but we have included both accessions in this analysis as the plants are distinct both in terms of sequence variation and in their morphology (S. parviflora has much longer and more golden-colored bristles than S. geniculata). The S. geniculata/S. parviflora clade is sister to another containing two subclades; S. barbata plus S. poiretiana A and S. palmifolia plus S. poiretiana B. These three taxa were assigned to section Ptychophyllum by Rominger (1962)
and characterized by a longer, more lax inflorescence and plicate leaves. The two accessions of S. poiretiana included in this analysis are morphologically very similar in both vegetative and inflorescence characters but show pronounced sequence differences. Such findings emphasize the need for further taxonomic work on this group of Setaria species.
Clade 3, which comprises all of the sampled species of Pennisetum and Cenchrus, has strong bootstrap support (91%). Pennisetum is paraphyletic with respect to the embedded monophyletic subclade of the Cenchrus species. It is not surprising that Pennisetum and Cenchrus should be closely related; various species have been described in both genera at different times, and several authors have indicated that the two genera may need to be merged (DeLisle, 1963
; Correll and Johnston, 1970
).
Development
The overall picture of development is one of early similarity followed by diversification at later stages. Developmental characters were optimized onto one of the most parsimonious trees and were mostly congruent with the molecular phylogeny. However, an ILD test found a significant lack of congruence between the two data sets. By visual inspection we found that the major difference between the topologies was the change in position of the subclade containing S. palmifolia and its allies (section Ptychophyllum). When the subclade was removed and the ILD test rerun, the data sets proved to be congruent. A Templeton test showed that there was a significant difference between an unconstrained molecular phylogeny and one constrained by having the Setaria palmifolia group basal in the ingroup (as suggested by the developmental data). The results of these two tests suggest that the differing position of the subclade of Setaria palmifolia and its allies is responsible for the lack of congruence between the molecular and developmental data sets. Setaria as a whole and Setaria clade 2 are not defined by any developmental character, but Setaria clade 1 is defined by failure of higher order spikelets to develop. Pennisetum and Cenchrus together are defined by four inflorescence characters, and Cenchrus by one, but no character defines Pennisetum itself. This is consistent with the paraphyly of Pennisetum in the molecular tree. The obvious key character separating Pennisetum from Cenchrus is the fusion of the bristles, although this is variable within the genus. Most species of Cenchrus also have at least some flattened bristles; the most parsimonious optimization of this character suggests that there has been more than one evolution of flattened bristles.
Relative numbers of initiated bristles and spikelets (no. 6) is the only developmental character that is ambiguous in its optimization. The change from initiation of equal numbers of bristles and spikelets to initiation of more bristles than spikelets can be optimized in two different ways. In one optimization (not shown) the character links Setaria clade 2 and Pennisetum/Cenchrus, with a reversal within the S. palmifolia subclade of Setaria clade 2. In the other optimization (shown) this character originates twice, once at the base of Pennisetum/Cenchrus, and once defining the subclade of S. geniculata and S. parviflora in Setaria clade 2. The ambiguous optimization suggests that this character state in Setaria may not be homologous with the similar condition in Pennisetum/Cenchrus.
Variation appears at different times throughout development. Changes involving inflorescence ramification and primordial differentiation occur very early, as the result of activity of the apical meristems of the branch axes. Changes involving elongation of the axes, including the differential elongation of the bristles seen in Pennisetum and Cenchrus, occur much later. This suggests that the various developmental characters may be under different genetic control.
In most cases, temporal variation in developmental characters would not have been detected by investigating mature morphology. For example, the four species of Setaria with lax inflorescences—S. palmifolia, S. barbata, and the two accessions of S. poiretiana—differ from other species of Setaria by having more 3° axes and fewer orders of axes, as well as lacking a condensed inflorescence (Figs. 1–6). Production of 3° axis primordia occurs early in development, but further ramification of the axes does not occur, resulting in a low number of orders of axes at maturity. In contrast, elongation of inflorescence internodes leading to the characteristic appearance of the inflorescence of these species occurs late in development, at a later stage than any of the electron micrographs shown here.
The greater detail afforded by developmental analysis can also provide a better explanation of differences among species than mature morphology alone. In Setaria the number of bristles that subtend the spikelets at maturity has been used as a diagnostic taxonomic character (Rominger, 1962
); in species such as S. palmifolia bristles and spikelets appear to be paired, with spikelets lateral to bristles, while in others, such as Setaria viridis, multiple bristles appear to subtend each spikelet in the mature secondary axis. In early development, however, most species have similar numbers of spikelets and bristles. Differences appear at maturity because the spikelets on higher order branch meristems fail to develop in some species while in others all initiated spikelets develop. In all species all initiated bristles grow on to maturity. Therefore the "character" of number of bristles per spikelet is a composite of at least three developmental characters, that is, number of orders of branching, number of primordia per order of branching, and number of spikelets that fail to develop. In contrast, in Setaria species such as S. geniculata and S. parviflora, and also in Pennisetum and Cenchrus, the high number of bristles relative to spikelets is determined in the early stages of development and thus is truly nonhomologous to the situation in Setaria clade 1.
Differences among species described here result from a combination of changes from different stages of the developmental process. This is a more complex view of developmental evolution than can be explained by simple heterochrony or heterotopy. It also suggests that developmental changes can be combined in a variety of ways, leading to diverse morphologies at maturity. However, not all combinations of developmental processes are found within the bristle clade, suggesting that there may be constraints on developmental evolution, whether they be phylogenetic or physically intrinsic to the processes of growth and differentiation of meristems.
Pensiero and Vegetti (2001)
have recently published a typological analysis of the inflorescences of Setaria, in which they apply the terminology of Troll (1969)
to some of the same species as we describe here. Their observations of mature morphology are largely congruent with our developmental data, although they note the occurrence of sterile (presumably suppressed) spikelets in S. palmifolia and S. poiretiana, whereas we have not observed this.
Mabee (1993)
and Hufford (1995)
have suggested that development could be described as a linear series of transformations that could be mapped onto a phylogeny. From this it would be possible to determine the relative frequency of terminal deletions or additions or of novel or reciprocal substitutions. One example of such a linear transformation series is the number of orders of branching, where change in the number of orders of branches can be seen as terminal additions or deletions. However, other variation does not readily fit into a linear series; for example, a nonterminal addition to the developmental pathway has apparently occurred in Setaria palmifolia, where its basal secondary branches may be polystichous rather than distichous as in all other species examined. In this case the apical meristem of the secondary axis is appreciably larger than meristems that produce higher order primordia distichously. Such a change to the inflorescence development pathway does not obviously involve a slowing down or speeding up of the ancestral developmental pathway and is therefore not heterochronic (Guerrant, 1988
; Zelditch and Fink, 1996
; Klingenberg, 1998
; Zelditch, Sheets, and Fink, 2000
). Another example includes the change in position of the 3° and higher order axes from ramified in Setaria to surrounding the 2° axis in Pennisetum and Cenchrus. This may be better described as a heterotopic change, although it is also in part dependent on heterochronic change in 2° axis elongation (the 2° axis elongates very little in Pennisetum and Cenchrus). In general, inflorescence variation is difficult to fit into a linear framework and may be better described by combinatorial factors (Kellogg, 2000b
).
Our ultimate aim is to identify genes involved in morphological diversity and to elucidate how these genes and the pathways to which they belong have changed over evolutionary time. This paper presents the basic phylogenetic and developmental background against which such effort must be viewed. In the grass family the synteny of genomes allows information on gene location from one species to be transfered to other species (Gale and Devos, 1998
; Devos and Gale, 2000
; Kellogg, 2001
). The great number of morphological mutants found in maize also allows the use of a "candidate gene" approach, where mutant phenotypes can be compared to the morphology of other grass species. For example, the tasselseed 4 mutant of maize produces an extensively ramified inflorescence strikingly similar to the inflorescence of Setaria italica (Neuffer, Coe, and Wessler, 1997
). Where phenotypes are similar, we hypothesize that genes controlling the mutant phenotype may also control similar morphology in other grasses.
Comparative developmental morphology is the crucial link between genetic studies in model systems and evolution of morphological diversity in large clades, because developmental morphology is the direct outcome of spatial and temporal patterns of gene expression. Diversification of developmental processes in related taxa leads to morphological diversification. Analysis of changes in developmental morphology with molecular phylogenies adds a temporal component to the study of morphological evolution that is not available when only mature structures are analyzed.
|
FOOTNOTES |
|---|
2 Author for reprint requests (adoust{at}umsl.edu
)
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Arber A. 1934 The Gramineae, 186–194. MacMillan, New York, New York, USA
Ayyangar G. N. R. T. R. Narayanan T. N. Rao 1933 The inheritance of characters in Setaria italica (Beauv.), the Italian millet. Part IV. Spikelet-tipped bristles. Indian Journal of Agricultural Science 3: 552-558
Benabdelmouna A. M. Abirached-Darmency H. Darmency 2001 Phylogenetic and genomic relationships in Setaria italica and its close relatives based on the molecular diversity and chromosomal organization of 5S and 18S-5.8S-25S rDNA genes. Theoretical and Applied Genetics 103: 668-677[CrossRef][ISI]
Benabdelmouna A. Y. Shi M. Abirached-Darmency H. Darmency 2001 Genomic in situ hybridization (GISH) discriminates between the A and the B genomes in diploid and tetraploid Setaria species. Genome 44: 685-690[Medline]
Brown R. 1810 Prodromus Florae Novae Hollandiae et Insulae Van-Diemen, vol. 1. J. Johnson, London, UK
Brown R. 1814 General remarks, geographical and systematical, on the botany of Terra Australis. In M. Flinders [ed.], A voyage to Terra Australia, undertaken for the purpose of completing the discovery of that vast country, and prosecuted in the years 1801, 1802, and 1803, vol. 2, 533–613. W. Bulmer, London, UK
Butzin F. 1977 Evolution der infloreszenzen in der Bortstenhirsen-verwandtschaft. Willdenowia 8: 67-79
Clark L. G. W. Zhang J. F. Wendel 1995 A phylogeny of the grass family (Poaceae) based on ndhF sequence data. Systematic Botany 20: 436-460
Clayton W. D. S. A. Renvoize 1986 Genera Graminum. Her Majesty's Stationery Office, London, UK
Cope T. A. 1982 Poaceae. In E. Nasir and S. I. Ali [eds.], Flora of Pakistan 143: 177–246
Correll D. S. M. C. Johnston 1970 Manual of the vascular plants of Texas. Texas Research Foundation, Renner, Texas, USA
Cunningham C. W. 1997 Can 3 incongruence tests predict when data should be combined?. Molecular Biology and Evolution 14: 733-740[Abstract]
Deleporte P. 1993 Characters, attributes, and tests of evolutionary scenarios. Cladistics 12: 427-432
DeLisle D. G. 1963 Taxonomy and distribution of the genus Cenchrus. Iowa State Journal of Science 37: 259-351
Devos K. M. M. D. Gale 2000 Genome relationships: the grass model in current research. Plant Cell 12: 637-646
Evans M. W. F. O. Grover 1940 Developmental morphology of the growing point of the shoot and the inflorescence in grasses. Journal of Agricultural Research 61: 481-520
Farris J. S. M. Källersjö A. G. Kluge C. Bult 1994 Testing the significance of incongruence. Cladistics 10: 315-319[CrossRef][ISI]
Felsenstein J. 1985 Confidence limits of phylogenies: an approach using the bootstrap. Evolution 39: 783-791[CrossRef][ISI]
Gale M. D. K. M. Devos 1998 Plant comparative genetics after 10 years. Science 282: 656-659
Giussani L. M. J. H. Cota-Sanchez F. O. Zuloaga E. A. Kellogg 2001 A molecular phylogeny of the grass subfamily Panicoideae (Poaceae) shows multiple origins of C4 photosynthesis. American Journal of Botany 88: 1993-2012
Goebel K. 1884 Beiträge zur Entwickelungsgeschichte einiger inflorescenzen. Jahrbucher fur wissenschaftlichen Botanik 14: 1-42
Gómez-Martinez R. A. Culham 2000 Phylogeny of the subfamily Panicoideae with emphasis on the tribe Paniceae: evidence from the trnL-F cpDNA region. In S. W. L. Jacobs and J. Everett [eds.], Grasses: systematics and evolution, 136–140. CSIRO, Melbourne, Australia
Grandcolas P. P. Deleporte L. Desutter-Grandcolas C. Daugeron 2001 Phylogenetics and ecology: as many characters as possible should be included in the cladistic analysis. Cladistics 17: 104-110[ISI]
Guerrant E. O. 1988 Heterochrony in plants: the intersection of evolution, ecology, and ontogeny. In M. L. McKinney [ed.], Heterochrony in evolution: a multidisciplinary approach, 111–133. Plenum, New York, New York, USA
Hufford L. 1995 Patterns of ontogenetic evolution in perianth diversification of Besseya (Scrophulariaceae). American Journal of Botany 82: 655-680[CrossRef][ISI]
Kellogg E. A. 2000a Molecular and morphological evolution in the Andropogoneae. In S. W. L. Jacobs and J. Everett [eds.], Grasses: systematics and evolution, 149–158. CSIRO, Melbourne, Australia
Kellogg E. A. 2000b A model of inflorescence development. In K. L. Wilson and D. A. Morrison [eds.], Monocots: systematics and evolution, 84–88. CSIRO, Melbourne, Australia
Kellogg E. A. 2001 Evolutionary history of the grasses. Plant Physiology 125: 1198-1205
Klingenberg C. P. 1998 Heterochrony and allometry—the analysis of evolutionary change in ontogeny. Biological Reviews of the Cambridge Philosophical Society 73: 79-123[Medline]
Mabee P. M. 1993 Phylogenetic interpretation of ontogenetic change: sorting out the actual and artefactual in an empirical case study of centrarchid fishes. Zoological Journal of the Linnaean Society 107: 175-291
Maddison D. R. W. P. Maddison 2000 MacClade 4: analysis of phylogeny and character evolution. Version 4.0. Sinauer, Sunderland, Massachusetts, USA
Murphy J. A. 1978 Non-coating techniques to render biological specimens conductive. Scanning E
