Development 1995 Talbert 2723 35

8/10/2019 Development 1995 Talbert 2723 35

1/13

INTRODUCTION

Elaboration of the plant body pattern depends primarily on theproper regulation of cell division versus cell differentiation atthe growth sites, called meristems. In seed plants apical growthis carried out by the apical meristems. Although structurallyidentical, shoot apical meristems differ ontogenetically. Aprimary shoot apical meristem originates during embryogene-sis and becomes the apex of the primary shoot. Secondaryshoot apical meristems develop later on the sides of theprimary shoot and form lateral shoots. In many seed plants,radial growth of the shoot is conferred by the cambium, a cylin-drical meristematic layer in the shoot body. Growth of lateralleafy organs (i.e. leaves, petals, etc.) occurs from transientmeristems formed on the ank of the apical meristem. Root

growth occurs from analogous apical and cambial meristems.At present we understand very little of the regulation and inter-action of these different types of meristems.

Morphological and developmental studies suggest thatdifferent meristems interact. An example is provided by thedevelopment of axillary meristems. These meristems form atrepeated positions along the shoot, known as nodes, that areseparated by sections of stem called internodes. A nodetypically bears one or more leaves, each of which contains ameristem in its axil that can form a branch shoot. Axillarymeristems are thought usually to originate directly from theshoot apical meristem as detached meristematic cells in theaxils of the leaf primordia (Garrison, 1955). However, in some

cases axillary meristems appear to arise from the leaf pri-

mordium rather than from separate meristematic cells(Majumdar, 1942; Irish and Sussex, 1992). Pioneering surgicalstudies by Snow and Snow (1942) showed that the formationof an axillary shoot was dependent on the subtending leaf pri-mordium and inhibited by the apical meristem. These resultssuggest that axillary meristems may form in response to pos-itional information set up by the action of opposing morpho-genetic gradients. Genetic or molecular data supporting thismodel are currently lacking.

The owering plant Arabidopsis thaliana has proved useful forthe molecular genetic analysis of developmental problems. Ara-bidopsis has a basal rosette of vegetative leaves and an erectbranched inorescence. The branching pattern of the inor-escence is most easily described in terms of the phytomer concept.

A phytomer (Galinat, 1959) or metamer (White, 1984) is a reit-erated module typically composed of an internode and a nodewith its leaf and axillary branch. Many variations of this basicpattern occur, such as the suppression of leaves or branches, ortheir modication into structures specialized for assorted repro-ductive or vegetative functions. Three types of phytomers havebeen described in the primary shoot of Arabidopsis (Schultz andHaughn, 1991; Fig. 1A). All three are arranged in a spiral phyl-lotactic pattern. Type 1 comprises the basal rosette. Each type 1phytomer has an extremely short internode, a rosette leaf, and anaxillary meristem that may form a branch. Type 2 comprises thebasal portion of the owering stalk. These phytomers have longinternodes, cauline leaves and axillary branches. The rst axillary

2723Development 121, 2723-2735 (1995)Printed in Great Britain The Company of Biologists Limited 1995

The form of seed plants is determined by the growth of anumber of meristems including apical meristems, leaf meristems and cambium layers. We investigated ve

recessive mutant alleles of a gene REVOLUTA that isrequired to promote the growth of apical meristems and tolimit cell division in leaves and stems of Arabidopsis thaliana. REVOLUTA maps to the bottom of the fth chro-mosome. Apical meristems of both paraclades (axillaryshoots) and owers of revoluta mutants frequently fail tocomplete normal development and form incomplete orabortive structures. The primary shoot apical meristemsometimes also arrests development early. Leaves, stemsand oral organs, in contrast, grow abnormally large. We

show that in the leaf epidermis this extra growth is due toextra cell divisions in the leaf basal meristem. The extentof leaf growth is negatively correlated with the develop-

ment of a paraclade in the leaf axil. The thickened stemscontain extra cell layers, arranged in rings, indicating thatthey may result from a cambium-like meristem. Theseresults suggest that the REVOLUTA gene has a role in reg-ulating the relative growth of apical and non-apicalmeristems in Arabidopsis .

Key words: meristem, branching, leaf growth, Arabidopsis, REVOLUTA

SUMMARY

The REVOLUTA gene is necessary for apical meristem development and for

limiting cell divisions in the leaves and stems of Arabidopsis thaliana

Paul B. Talbert, Haskell T. Adler, David W. Parks* and Luca Comai

University of Washington, Department of Botany, Box 355325, Seattle, WA 98195-5325, USA*Present address: Camellia Forest Nursery, 125 Carolina Forest Rd, Chapel Hill, NC, 27516


2/13

2724

branch originates in the uppermost leaf axil and others developbasipetally through the type 2 and 1 phytomers (Alvarez et al.,1992; Hempel and Feldman, 1994). In both of these phytomertypes, a second branch (accessory branch) may later developbetween the axillary branch and its subtending leaf (Fig. 1A).Type 3 phytomers constitute the fertile terminal portion of thestalk, known as the main orescence (Weberling, 1989). Eachtype 3 phytomer has an intermediate-length internode, no leaf anda lateral ower. As observed by Goethe (1790), owers are spe-cialized shoots with oral organs in the place of leaves. Thus theowers of Arabidopsis can be viewed as oral branches whichare serially homologous with the axillary branches. While theoral branches have determinate growth and are composed of special phytomer types (i.e. the whorls of four sepals, four petals,six stamens and two carpels), the indeterminately growingaxillary and accessory branches repeat in part the pattern of phytomers found on the primary shoot and are therefore termedparaclades (Weberling, 1989). The type 2 phytomers of a rst-order paraclade give rise to second-order paraclades, and thesegive rise to higher-order paraclades (Fig. 1A). The type 3phytomers of a paraclade make up a coorescence (Weberling,1989). The primary shoot distal to the rosette and the paracladestogether make up the inorescence.

The study of Snow and Snow (1942) suggests that genesinvolved in axillary meristem development might also have rolesin the primary apical meristem or in the development of leaves.We therefore speculated that mutations in Arabidopsis genesaffecting apical meristems might be identiable by changes in leaf morphology, as is true of the tomato gene lanceolata (Caruso,1968), and screened for mutants accord-ingly. We describe here mutantsdefective in a gene, REVOLUTA ( REV ) ,that is necessary to promote the normalgrowth of apical meristems, including

paraclade meristems, oral meristemsand the primary shoot apical meristem.Simultaneously, REV has an opposingeffect on the meristems of leaves, oralorgans and stems, being necessary tolimit their growth. Thus rev mutationsreveal a novel and fundamental regula-tory feature of plant body pattern elabo-ration.

MATERIALS AND METHODS

Plant growth conditionsPlants were grown in 5.0-cm pots in thegreenhouse in a buffered soil mix of 67%peat moss and 33% pumice. Naturalsunlight was supplemented by 16 hours of illumination with 1000 W high pressuresodium lights. Plants were subirrigatedwith alternating solutions of Peters Peat-Lite Special (20-10-20) or Peters Dark Weather Food (15-0-15), both diluted to100 ppm. Temperature control set pointswere at 18-21C during the day and 9-13Cat night. To assist uniform germination,wetted seeds were kept at 4C for 2 daysprior to planting. Because of the inherentvariability of greenhouse conditions, all

comparisons of wild-type and mutant phenotypes were made oncohorts of plants grown side-by-side under identical conditions.Cohort 1 was grown in May-July of 1994, cohort 2 in July-August1994, cohort 3 in October-November 1994, cohort 4 in March-May1994, and cohort 5 in October-December 1993. To achieve 100% ger-mination and uniform growth, cohort 4 was germinated on growthmedium containing 0.5 Murashige-Skoog salts (Gibco) under con-

tinuous illumination with Philips 40 W Cool White lights (75 E/m2

sec) and transplanted to soil on day 14 after sowing.

Mutagenesis and allelism testsAll ve rev mutations were recovered from the M 2 progeny of M 1seeds mutagenized with ethyl methanesulfonate (EMS). Alleles rev-1, rev-2 and rev-4 were each recovered from an independent batch of M1 seeds of ecotype Nossen-0 (No-0) mutagenized in 10 mM EMSfor 17 hours. This dosage gave siliques with segregating M 2 embryo-lethals on approximately half of the M 1 plants. The rev-3 and rev-5(=spitzen-1 ; Alvarez, 1994) alleles were induced in the Columbiaecotype and generously supplied by Laura Conway and David Smyth,respectively. Allelism of rev-1 and rev-2 was determined by their non-complementation in multiple reciprocal crosses. Further complemen-tation tests were carried out by pollinating a marked er ttg rev-2 stock with a rev-3/REV , rev-4 /rev-4 , or rev-5/rev-5 parent. Individuals thatwere Rev but not Er or Ttg made up half (15/32), all (7/7), and all(23/23), respectively, of the progeny in these three crosses. The rev-1 allele was backcrossed to the wild type three times to eliminate anyeffects of additional mutations on the phenotype. The other alleleshave been backcrossed once.

MappingMapping of rev-1 used the polymerase chain reaction mappingmarkers DFR (Konieczny and Ausubel, 1993) and nga129 (Bell and

P. B. Talbert and others

rosetteparaclade

primaryshoot

rosette leaf

*

**

caulineparaclade

primaryshoot

rosette leaf

caulineleaf

3p

2p2p

3p

REV rev

A B

1

2

3

caulineleaf

accessoryparaclade

*

2p

*

*

caulineparaclade

Fig. 1. Schematic representation of wild-type REV and mutant rev morphologies. (A) Wild type,showing phytomer types (1, 2 and 3) and branching pattern. (B) rev, showing reduced growth of primary shoot, paraclades and owers, and increased growth of leaves and stems . , owers;2p, second-order paraclade; 3p, third-order paraclade; *, axil lacking a paraclade.


3/13

2725Meristem development in Arabidopsis

Ecker, 1994) on chromosome 5. DFR maps to 61.5 mu (AAtDBversion 3-4) and nga129 maps to approximately 86.4 mu (6.6 mubelow m435 ; Bell and Ecker, 1994). A rev-1 mutant (No-0 ecotype)was crossed to wild-type Landsberg erecta and phenotypically RevF2 individuals were scored. We also mapped REV with respect to themorphological marker lfy (Weigel et al., 1992). LFY maps to 89.5 muon the distal tip of chromosome 5 (AAtDB version 3-4). Rev F 2 indi-

viduals from a rev-1 lfy-6/+ cross were progeny tested for segre-gation of lfy rev F3 double mutants (Adler et al., unpublished data).Map distances were corrected for double cross-overs using theKosambi mapping function (Koornneef and Stam, 1992), and the 95%condence interval was computed using the two tails of the binomialdistribution (Woolson, 1987).

Measurements and cell countsMeasurements of plant organ lengths were made with either a Measydial caliper (Max Mgerle, Sen., Switzerland), a calibrated Zeissreticle or a standard metric ruler. Leaf areas in cohort 4 were deter-mined by harvesting leaves and tracing their outlines. The outlinedareas were measured using a Kurta digitizing tablet and SigmaScansoftware (Neff and Van Volkenburgh, 1994). Epidermal peels of theadaxial leaf surfaces were made using Revlon nail strengthener andtransparent tape. The peels were affixed to microscope slides andthree patches of cells from different parts of each leaf were traced ina camera lucida, avoiding the midrib. Cells in each patch were countedand areas of the patches were determined (Neff and Van Volkenburgh,1994). The mean cell sizes of the three patches of each leaf wereaveraged to get the mean cell size for that leaf. Similar results wereobtained when the three patches were grouped as a single patch.Unpaired two-tailed Students t -tests on measurement sets wereconducted using the Statview Student statistical software package.The P value for these tests is the probability that the true means of the two genotypic populations are identical.

Determination of the region of leaf growth was made by markinggrowing leaves from cohort 5 with dots of Sumi black calligraphy ink (Yasutomo & Co.) placed 1-2 mm apart along the midrib andmeasuring subsequent growth relative to the xed dots.

Anatomical sectionsPlant material for sectioning was xed in formalin-aceto-alcohol,dehydrated through an increasing tertiary butyl alcohol series andembedded in Paraplast X-Tra (Oxford Labware) paraffin (Johansen,1940). Serial 10 m sections were cut on a Spencer microtome andmounted on microscope slides with Haupts adhesive (Johansen,1940). Paraffin was dissolved in xylene and the slides were hydratedthrough a decreasing ethanol series before staining 1-2 minutes in0.25% toluidine blue, or in safranin O and fast green FCF essentiallyas described by Johansen (1940) except for the following: picric acidand ammonia were omitted from the washes; 0.25 g of fast green wasdissolved in a mixture of 100 ml each of methyl cellosolve, absoluteethanol and methyl salicylate; and clearing was in a mixture of 50%methyl salicylate, 25% absolute ethanol, and 25% xylene. Sectionswere photographed with a Nikon Microphot FX using Fujicolor 100lm.

RESULTS

rev alleles and mappingWe recovered three independent mutants with a syndrome con-sisting of revolute (downwardly curled) leaves, reducedbranching and many infertile owers from the M 2 self-progenyof the No-0 ecotype plants mutagenized with EMS. The rstmutant was backcrossed to the wild type and its mutantphenotype segregated as a simple recessive mutation in the F 2

generation (160 wild type: 52 mutant; 2 = 0.02; P = 0.9), indi-cating that the mutation is likely to be a loss-of-function. Com-plementation tests showed that the mutations dened a singlelocus that we call REVOLUTA represented by three alleles(rev-1 , rev-2 and rev-4 ). Two EMS-induced alleles ( rev-3 andrev-5 ) were available in the Columbia ecotype.

The rev-1 , rev-2 and rev-4 alleles all had indistinguishablephenotypes. The rev-3 and rev-5 mutants were more and lessseverely affected than these alleles, respectively, but it is notyet clear whether these differences are due to the properties of the latter alleles or to modifying factors from the mutagenesesor ecotype. The phenotype of rev-1 was stably transmittedthrough three backcrosses with only a slight reduction inseverity and this allele was therefore selected for carefuldescription.

REV showed linkage to the fth chromosome markers DFR(11 recombinants/32) and nga129 (13 recombinants/64). Six of the 11 recombinants for DFR were also recombinant fornga129 , suggesting that REV is distal to nga129 . This placedREV approximately 21 mu from nga129 at approximately 108mu (99 mu actual location 124 mu) on the bottom tip of chromosome 5 (Koornneef and Stam, 1992). This map locationwas veried by REV showing tight linkage to LFY (2 recom-binants/33). This second experiment placed REV approxi-mately 6.1 mu from the LFY locus. Because the nga129 datasuggest that REV is distal to LFY , this experiment gives a mapposition for REV of approximately 95.6 mu (90 mu actuallocation 110 mu) on the bottom tip of chromosome 5.

Overview of the Rev vegetative phenotypeThe rev mutations were pleiotropic and strongly affected bothvegetative and reproductive shoot development (Fig. 1B). Incontrast, no differences from wild-type REV controls werediscerned in the roots of rev-1 and rev-2 mutants up to 2-3weeks of age. Older roots have not been examined. We willrst describe the effects of the rev-1 mutation on the vegeta-tive phytomers (types 1 and 2) before describing its effects onthe reproductive phytomers (type 3).

The rev-1 mutation caused overgrowth of both rosette andcauline leaves. The rosette leaves of rev-1 plants were notreadily distinguishable from wild-type No-0 leaves prior tobolting. As bolting began, however, the youngest rosette leavesbecame abnormally large and distorted or uneven in shape asthey matured. This overgrowth was even more dramatic in thecauline leaves. The latter were longer and narrower than wild-type cauline leaves, had rolled-under margins, and curveddownward along their longitudinal axes (Fig. 2A,B). Both theleaves and the primary shoots of rev-1 mutants were oftendarker green than those of wild type. Paraclades frequentlyfailed to develop in both rosette and cauline leaf axils (Fig. 2B).The axils instead appeared empty, contained thin lamentousstructures usually bearing branched trichomes, or bore leaveswith or without a visible supporting stem (Fig. 2C,E-G).

Paraclade formation is reduced in rev-1 mutantsTable 1 shows the mean numbers of vegetative nodes, leavesand paraclades on the primary shoots and rst-order paracladesof REV and re v-1 plants from cohort 1. The rev-1 plants hadan average of about one more rosette leaf and one less caulineleaf on the primary shoot than did the REV plants, but only the


4/13

2726 P. B. Talbert and others

Fig. 2. Morphology of wild-type and rev-1 plants. (A) REV plant. (B) rev-1 plant. (C) REV cauline leaf axil with paraclade. (D) REV orescence apex. (E-G) rev-1 cauline leaf axils: (E) empty axil; (F) axillary lament; (G) axillary leaf. (H) rev-1 orescence apex with owersand tapered laments. (I) rev-1 orescence apex with clustered tapered laments. (J) rev-1 tapered laments subtended by knobs and a lament(arrow). (K) rev-3 arrested orescence apex. f, fertile ower; k, knobs; s; sterile ower: t; tapered lament.


5/13


Fig. 3. Anatomy of cauline leaves and axils. (A-G) Longitudinal axillary sections. (A) Immature REV axil. (B) Near-mature REV axil.(C) Empty REV axil. (D-G) rev-1 axils with: (D) a leaf and terminal knob on a short stem; (E) a lament; (F) club-like projections;(G) trichomes. (H-K) Midlength leaf cross-sections. (H) REV leaf midvein. (I) rev-1 leaf midvein. (J) REV leaf margin. (K) rev-1 leaf margin.Bars, (A-D,F,H,I) 20 m; (E,G,J,K) 100 m. l, leaf; p, paraclade; s, stem.


6/13

2728

leaf on the primary shoot than did the REV plants, but only the

latter difference was signicant at the 95% condence level.On the primary shoots of REV plants, paraclades developedfrom the axils of 100% of the cauline leaves and 72% of therosette leaves. In sharp contrast, the primary shoots of rev-1mutants formed paraclades in 61% of cauline leaf axils and in18% of rosette leaf axils (Table 1). Accessory paraclades werevisible in the axils of about 85% of the REV cauline leaves, butin only about 4% of rosette leaf axils. No accessory paracladeswere formed in rev-1 mutants. Rosette paraclades had morevegetative nodes than cauline paraclades in both genotypes( REV : 3.5 vs. 1.9, P = 0.0001; rev-1 : 2.5 vs. 1.6, P = 0.002).

Second-order paraclades normally developed in the caulineleaf axils of the REV rst-order paraclades, and accessory par-

aclades developed in 41% of these axils. However, some of thevegetative nodes of REV rst-order paraclades (10/151)developed second-order paraclades even though they failed toform a macroscopic leaf. Such second-order paraclades weresubtended by putative vestigial leaf tissue, similar to thatdescribed by Hempel and Feldman (1994) on the primaryshoots of Columbia and Landsberg erecta plants. Rarely (2/151nodes), a second-order paraclade failed to develop in the axil

of a cauline leaf on a rst-order paraclade. Thus about 93% of the vegetative nodes on REV rst-order paraclades bore caulineleaves and about 99% bore second-order paraclades.

On the rev-1 rst-order paraclades, all vegetative nodes hadcauline leaves, but only 4/62 (6%) also bore second-order par-aclades. A leaf, with or without a visible supporting stem, grewfrom 3 of the 80 rev-1 cauline leaf axils on the primary andsecondary shoots.

Data from other cohorts of rev-1 , rev-2 and rev-3 plantsgenerally conrmed the results shown in Table 1, except thatthe number of accessory branches in wild-type plants wasgenerally less and the rev mutants in these cohorts had evenlower rates of paraclade formation, with the fraction of primarycauline leaf axils with paraclades ranging from 0% to 36%(data not shown).

Decapitation of rev-1 plantsWe tested whether the failure of paraclade formation in rev-1plants could be relieved by decapitation of the primary shoot,as might be expected if it resulted from extreme apicaldominance. A strain of rev-1 (from a Landsberg erectaoutcross) was used which typically forms no paraclades.Fourteen owering rev-1 plants were decapitated just above therosette on day 31, and 19 sibs were left undecapitated. 20 dayslater no rosette paraclades had formed on the decapitatedplants, and only one rosette paraclade was present on the unde-capitated plants. An analogous experiment with a normal rev-1 strain also failed to detect a stimulation of paracladeformation by decapitation (data not shown).

Anatomy of rev-1 axilsTo understand better the nature of the defect in rev-1 paracladegrowth, we made serial longitudinal sections through wild-typeand rev-1 cauline leaf axils. In a cauline leaf axil from a 17-day-old wild-type plant from cohort 2 (just prior to visiblebolting), the developing paraclade appeared to arise from thebase of the leaf (Fig. 3A). Procambium was differentiating inthe leaf and primary shoot, and procambial strands wereevident by their darker staining in the developing paraclade. Inthe leaf, vacuolation was more advanced in the abaxial cells


Table 1. Mean numbers of leaves and paraclades in wild-type REV and mutant rev-1 plants

Structures REV rev-1 P value of t -test

Primary shoots*Vegetative nodes 8.2 8.5 0.50Rosette leaves 6.1 7.1 0.06Cauline leaves 2.1 1.4 0.002Rosette paraclades 4.4 1.3 0.0001Cauline paraclades 2.1 0.85 0.0001Accessory rosette paraclades 0.2 0 0.33Accessory cauline paraclades 1.8 0 0.0001

First-order paracladesVegetative nodes 2.7 2.1 0.01Cauline leaves 2.6 2.1 0.06Cauline paraclades 2.7 0.14 0.0001Accessory cauline paraclades 1.1 0 0.0001

*Data are from the primary shoots of 13 REV and 13 rev-1 48-day-oldplants (cohort 1).

Data are from 55 REV and 30 rev-1 secondary shoots of cohort 1,comprised of approximately equal numbers of rosette and cauline paraclades.

Table 2. Mean dimensions of leaves and internodes in REV and rev-1 plants

P valueOrgan Dimension REV rev-1 of t -test

Cotyledon blade* length (cm) 0.27 0.34 0.00011st rosette leaf* length (cm) 1.52 1.91 0.00013rd rosette leaf* length (cm) 2.62 2.93 0.0238Longest rosette leaf length (cm) 4.9 6.8 0.00011st cauline leaf length (cm) 3.7 6.9 0.00012nd cauline leaf length (cm) 2.9 6.3 0.00011st cauline leaf width (cm) 1.3 1.1 0.102nd cauline leaf width (cm) 0.9 0.7 0.00081st cauline leaf area (cm 2) 3.5 5.1 0.00592nd cauline leaf area (cm 2) 2.1 3.4 0.00161st leaf epidermal cell area ( m2) 2100 2200 0.702nd leaf epidermal cell area ( m2) 2000 1600 0.0311st cauline internode length (cm) 7.4 3.8 0.00121st cauline internode diameter (mm) 1.05 1.24 0.0083

*Data from cohort 3, consisting of 17 REV and 20 rev-1 28-day-old plants.Data from cohort 2, consisting of 21 REV and 29 rev-1 37-day-old plants.

Nearly identical mean lengths were made on a rev-3 cohort (data not shown).All cauline leaf measurements and internode length were measured on

cohort 4, consisting of 12 REV plants and 14 rev-1 plants, on materialharvested on day 49. See Materials and Methods and the text for furtherdetails.

Internode diameter was measured just above the rosette on 53-day-oldplants from cohort 1.

Table 3. Growth in the basal portion of leavesMeasurement REV rev-1

Length of rst cauline leaf, day 21 (mm) 12.3 12.3Length of basal region*, day 21 (mm) 2.3 2.7Percentage of total leaf length in basal region, day 21 19 22Increase in total leaf length, day 21-56 (mm) 7.3 20.4Increase in length of basal region, day 21-56 (mm) 6.5 18.1

Percentage of total increase occurring in basal region 89 89

*The basal region is the arbitrary region of the leaf that is proximal to allink dots placed on the expanding 21-day-old leaves (cohort 5) to serve asposition markers.


7/13


than in the adaxial cells, as has been noted in other angiospermspecies (Steeves and Sussex, 1989). On the edges of the leaveswere heavily stained club-like stipules (data not shown). In amore mature (bolting) axil (Fig. 3B), vascular tissue was welldeveloped in the leaf, stem and paraclade. Adaxial leaf cellswere vacuolated; however, differentiation of the palisade layerand spongy mesophyll had not yet taken place.

As mentioned above, some axils on wild-type rst-order par-aclades were found that lacked second-order paraclades. Wesectioned four such mature axils and found all four lacked anyapparent axillary bud or mound of meristematic cells (Fig. 3C).A cryptic meristem of a few cells lacking obvious organizationcould still be present.

In contrast to the all-or-none development of paraclades inwild-type axils, rev axils frequently contained unusual inter-mediate structures. In 29 rev-1 cauline leaf axils, 3% of theaxils bore a leaf (Fig. 2G), 24% bore thin laments (Fig. 2F),and 62% of the axils appeared empty except for small bumpsof tissue present in about half of them. In 44 rev-3 axils, 23%bore leaves, 23% bore laments and 43% were empty or hadsmall bumps. Seven rev-1 axils from bolting plants werechosen for sectioning. One axil contained a paraclade indistin-guishable from a wild-type paraclade (data not shown), whileanother contained an apparent abortive paraclade consisting of a leaf and terminal knob separated from their subtending leaf by a short stem-like region (Fig. 3D). One axil contained ashort lament about ve cell layers in diameter (Fig. 3E). Fourothers contained one or more deeply staining club-shaped pro-

jections, often on a bulge of tissue in the position normallyoccupied by a paraclade (Fig. 3F). These were reminiscent of stipules, but were generally larger, more irregularly shaped,and were not located at the leaf edges, where normal stipuleswere present. These projections appear to correspond to thesmall axillary bumps visible macroscopically. Two of these

axils also supported unbranched trichomes (Fig. 3G), whichwere not observed in wild-type axils. These unusual structuresmay indicate a premature differentiation of the presumptiverev-1 axillary meristem.

The REV gene is required to limit cell division inleavesIn contrast to the abortive development of paraclades, theleaves of rev-1 mutants grew abnormally large. The differencein leaf size between wild-type and rev-1 plants was not obviousin the earlier rosette leaves, but we measured signicant sizedifferences in the cotyledons and rst and third leaves fromcohort 3 (Table 2). Later leaves differed more dramatically: themean length of the longest rosette leaf (ordinarily the youngest

leaf) of rev-1 plants was about 39% longer than wild-typecontrols, and rev-1 cauline leaves became up to twice as longas their wild-type counterparts.

To determine the timing of this excessive leaf growth, wemeasured the lengths of the rst and second cauline leaves of wild-type and rev-1 plants in cohort 4, daily from rst visiblebolting on day 33 until the growth of the cauline leaves of allwild-type and most rev-1 plants had terminated on day 45.(Bolting in this cohort was delayed more than a week relativeto other cohorts because seeds were germinated on plates andtransplanted to soil on day 14.) Over the rst week of mea-surement the average growth rate of the rst cauline leaves wasabout 80% greater for rev-1 plants than for wild-type plants

(5.8 mm/day vs. 3.2 mm/day), and rev-1 second cauline leavesgrew more than twice as fast as their wild-type counterparts(Fig. 4A). Most wild-type leaves quit growing by day 43, butmost rev-1 leaves continued to grow at least 2 days longer.

To determine the respective contributions of cell expansionand increased numbers of cells to the excessive growth of rev-1 cauline leaves, we harvested the rst and second caulineleaves from cohort 4 on day 49 and compared the adaxialsurface areas of REV and rev-1 leaves, as well as the averagesurface areas of the epidermal cells within these leaves (seeMaterials and methods). As shown in Table 2, both the rstand second rev-1 cauline leaves had signicantly larger surfaceareas (about 45% and 63% larger, respectively) than the wild-type controls. Despite these size differences, the averagesurface areas of individual epidermal cells from the rev-1 rstcauline leaves and from wild-type rst and second caulineleaves were not signicantly different. The rev-1 secondcauline leaf cells were smaller, perhaps indicating that not allof them had completed expansion by day 49. Both the rst andsecond rev-1 cauline leaves must therefore contain moreepidermal cells than their wild-type counterparts. We estimatethe average number of epidermal cells on the adaxial surfacesof the rst and second wild-type cauline leaves to be 1.7 105and 1.1 105, respectively. The corresponding estimates forrev-1 leaves are 2.3 105 and 2.2 105 epidermal cells. Thisimplies that one function of the wild-type REV gene must beto limit cell division in the leaves.

Cell division in rev-1 leaves is predominantly in thebasal region of the leafIn an expanding wild-type leaf, cell division is conned to thebasal region (Pyke et al., 1991). To determine if the extragrowth in rev-1 leaves was also conned to the basal regionsor spread over other parts of the leaves, expanding 21-day-old

rst cauline leaves of plants in cohort 5 were marked alongtheir midvein with dots of ink. The positions of dots weredetermined on day 21 and again on day 56. Table 3 comparesa representative REV leaf with a rev-1 leaf with a similar initialsize and placement of dots. Both leaves had about 90% of theirgrowth in the basal fth of the leaf. Similar results with otherleaves (data not shown) conrm that cell divisions in both REV and rev-1 leaves are conned to the basal regions of the leaves.

rev-1 leaf size is diminished when paraclades formTo see whether the overgrowth of leaves in rev-1 mutants isrelated to the failure to form paraclades, we compared thelengths of mature cauline leaves which subtended paracladesand those that lacked paraclades on 25 rev-1 plants from cohort

2. Although rev-1 cauline leaves were enlarged whether or notthey subtended paraclades, the mean length of 11 rst caulineleaves that lacked paraclades was 24% greater than the meanlength of 14 rst cauline leaves that subtended paraclades (6.8cm vs. 5.5 cm, respectively; P = 0.014). Similarly, 15 secondcauline leaves that lacked paraclades were 37% longer onaverage than 6 that had them (5.6 cm vs. 4.1 cm; P = 0.037).This suggests that the growth of a leaf is antagonisticallyrelated to the growth of the meristem in its axil.

rev-1 leaf structureWe cross-sectioned six REV and six rev-1 mature caulineleaves near their midlength. The wild-type leaves were 5 to 9


8/13

2730

cell layers thick between vascular bundles while rev-1 leavesranged from 6 to 12 cell layers. The rev-1 leaves also had morevascular elements in the midvein, with about 30 trachearyelements in the xylem versus about 15 in REV leaves (Fig.3H,I). The leaf margins tapered in wild-type leaves (Fig. 3J)and some rev-1 leaves, but were blunt and thickened in otherrev-1 leaves that had more cell layers and large intercellularspaces (Fig. 3K). The continued division of cells on the adaxialside of the leaf after vacuolation begins on the abaxial side mayaccount for the thickening of leaf margins and the downwardcurling of rev-1 leaves.

rev-1 cauline internodes are shortened andthickenedThe cauline leaves of rev-1 mutants were often lower on theprimary shoot than wild-type cauline leaves. We measured thelength of the rst cauline internode (between the rosette andthe rst cauline leaf) on the bolting plants of cohort 4. Bothwild-type and rev-1 internodes completed growth in about thesame time interval, but rev-1 internodes became only slightlymore than half as long as wild-type internodes on average (Fig.4B). This suggests that the failure of cauline internodeelongation may account for the average of one more rosetteleaf and one less cauline leaf on rev-1 shoots compared to wild-type shoots (Table 1). The rst cauline internodes of rev-1plants were also 18% larger in diameter (Table 2). Despite thisincreased diameter, which was also seen in subsequent intern-odes (data not shown), the shoots of rev-1 mutants were weakerand tended to fall over earlier and more frequently than wild-type shoots.

rev-1 shoots have extra cell layers that arise late indevelopmentTo investigate the cause of thicker stems in rev-1 mutants,

cross-sections of wild-type and rev-1 stems from cohort 2 weremade just above the rosettes of 17-day-old plants (just prior tovisible bolting) and 40-day-old plants (that had completedcauline leaf expansion). Individuals with six to eight procam-bial strands or vascular bundles were observed in bothgenotypes. In four REV and four rev-1 17-day-old plants pro-toxylem elements were visible but other tissues were poorlydifferentiated (Fig. 5A-D). A protodermal layer and about vecortical cell layers were found outside the procambial strands.The most obvious difference between the genotypes at thisstage was that paraclades were present in the axils of wild-typerosette leaves but were frequently missing or reduced tolaments in rev-1 axils.

The six stems of 40-day-old wild-type plants (Fig. 5E,F,I)

differed from the younger stems primarily in cell enlargementand in the presence of well-differentiated tissues. Theepidermal layer surrounded a cortex of about ve layers of chlorenchyma. The vascular bundles contained metaphloemand heavily lignied metaxylem elements. Between the xylemand phloem were a few layers of small cells that may representa fascicular cambium (Fig. 5I). Between and connecting thevascular bundles were a few cell layers of lignied (red-staining) interfascicular cells forming a continuous ring of scleried tissue adjacent to the chlorenchyma.

The stems of 40-day-old rev-1 plants (Fig. 5G,H,J) had astarkly contrasting anatomy. Inside the epidermis were severaladditional cortical layers not present in wild-type or younger

rev-1 stems, which included some very large parenchymatouscells, especially capping the phloem. The additional layers of chloroenchyma probably account for the darker color of rev-1stems. Inside these extra layers the vascular bundles werebroader and appeared to have more xylem and phloem cells.


Days

33 34 35 36 37 38 39 40 41

L e n g t

h ( c m

)

rev-1

REV

Growth of first cauline internodeB

0

1

2

3

4

5

6

7

8

9

A Growth of cauline leaves

33 34 35 36 37 38 39 40 41 42 43 44 45

L e n g t

h ( c m

)

Days

rev-1 1st

REV 1st REV 2nd

rev-1 2nd

0

1

2

3

4

5

6

7

8

Fig. 4. Growth of REV and rev-1 cauline leaves and internodes.(A) First and second cauline leaves. For simplicity, standard errorbars are shown in only one direction. (B) First cauline internode.


9/13


The interfascicular cells did not appear to be as heavilylignied as those in the wild type. This may account for theweaker stems of rev-1 mutants. Between the interfascicularcells and the cortical layers and between the xylem and phloem

was a ring of many small cells in les resembling a cambialzone. We suppose that periclinal divisions in this zone gaverise to the extra layers of cells. Since the extra cell layers werenot present in the 17-day-old rev-1 plants, in which the apical

Fig. 5. Stem cross-sections immediately above the rosette. (A-D) Stems of 17-day-old plants: (A,B) REV , (C,D) rev-1 . Some leaves weredamaged by dissection. (E-J) Stems of 40-day-old plants: (E,F,I) REV , (G,H,J) rev-1 . F and H are backlit. Bars, (A,C,E-H) 100 m; (B,D,I,J)20 m. Arrowheads, protoxylem elements; arrows, cambium-like zone; c, cortex; e, epidermis; if, interfascicular region; l, leaf; p, paraclade;pc, procambial strand; pd: protoderm; ph: phloem; v, vascular bundle; x, xylem; *, axillary laments.


10/13

2732

meristems had already initiated 10-15 owers, these layersmust have arisen from cells that retained a capacity to divideat a considerable distance behind the apical meristem.

Floral branch types in rev-1 orescencesIn addition to the striking effects rev-1 had on vegetativegrowth, it also dramatically affected the reproductive struc-tures. In describing these rev-1 mutants, we will use the termoral branches to refer both to the owers and to thoseabnormal structures lacking oral organs that occurred in theposition of owers. The phyllotaxy of the oral branches alongrev-1 orescence axes was commonly irregular, and internodelength was often reduced. This reduction, together with theshorter vegetative internodes, gave plants an overall shorterstature: the mean length of the primary shoot of 30 rev-1 plantsin cohort 3 was 24.1 cm, whereas 22 REV plants had a meanlength of 46.0 cm ( P = 0.0001). The rev-1 oral branchesexhibited a range of abnormalities which could be grouped intothree branch types (Fig. 2D,H): fertile owers that were largerthan normal, sterile and usually misshapen owers that lackedpistils, and tapered lamentous structures.

Frequency of oral branch typesThe relative frequency of these three types of oral branchesin an inorescence varied widely even among siblings in thesame cohort. Most rev-1 orescences bore some fertile owers,especially on the rst few oral nodes. In extreme cases,however, rev-1 orescences could consist almost entirely of closely spaced tapered laments and have a brush-like appear-ance (Fig. 2I). Table 4 gives the relative frequencies of thedifferent oral branch types for cohort 1. The number of fertileowers per rev-1 main orescence ranged from 0 to 34 in thiscohort. The overall fraction of fertile rev-1 owers among theoral branches was about 32% for the main orescences and

about 24% for the coorescences, compared to over 90% forall orescences in the wild type. Although sterile owers wereobserved in wild-type orescences, these were always well-formed, in contrast to the sterile owers of rev-1 plants whichlacked pistils. None of the tapered laments seen in rev-1 o-rescences were observed in wild-type plants.

Fertile owers and seeds are enlarged in rev-1plantsWe investigated in more detail the structures of the three oralbranch types. Fertile rev-1 owers generally had a completeset of oral organs and a normal shape, but the oral organswere consistently larger than those of wild-type owers (Table5). Sepals, petals and stamens from rev-1 owers actively

shedding pollen (stage 15 of Smyth et al., 1990) were found tobe 37%, 58% and 28% longer, respectively, than wild-typecontrols. Since pedicels and pistils are actively expanding atstage 15, we compared mature REV and rev-1 pedicels andsiliques just prior to dehiscence. Pedicels were about 28%larger in rev-1 compared to wild type. No difference was seenin REV and rev-1 silique length, but rev-1 siliques were 56%larger in diameter. These measurements suggest that the rev-1oral organs experience an overgrowth analogous to that seenfor leaves. We also measured the size of REV and rev-1 seedsfrom cohort 3. The rev-1 seeds were 15% longer than REV seeds (Table 5). Although we did not measure seed diameter,this appeared to be increased proportionally in rev-1 seeds.

Sterile rev-1 owers are missing oral organsThe oral organs of the rev-1 sterile owers ranged in sizefrom as large as the fertile owers to extremely dwarfed formswith incomplete development. The sepals in these sterileowers usually were of varying sizes, giving an asymmetricalappearance to the owers. The sepals sometimes bore branchedtrichomes, which are normally found only on leaves in wild-type plants. In both sterile and fertile rev-1 owers, rare indi-vidual organs of either the rst or second whorl were foundwith a hybrid identity that was part sepal and part petal. Dis-section of 50 sterile owers revealed that they were missing avariable number of oral organs in an acropetally increasingmanner: the mean numbers of organs found were 2.1 sepals,0.88 petals, 0.08 stamens and no carpels. Filamentous struc-tures were found in 26% of the owers. Two adjacent organsin a whorl were fused proximally in 6/50 sepal whorls, in 2/27petal whorls, and 1/2 stamen whorls. One of the two organsinvolved in a fusion was usually lamentous or reduced in size.A similar dissection of 17 rev-4 owers gave means of 3.0sepals, 1.9 petals, 0.24 stamens and no carpels, with lamen-tous structures in 29% of the owers. The acropetally increas-ing loss or reduction of oral organs suggests that the rev-1and rev-4 oral meristems become exhausted prematurely.

Tapered laments and the rev-1 orescence apexThe tendency to form incomplete owers and incomplete oral


Table 5. Mean dimensions of oral organs and seeds in REV and rev-1 fertile owers

Dimension (mm) REV rev-1 P value of t -test

Length of sepal* 1.9 2.6 0.0001Length of petal* 2.6 4.1 0.0001Length of long stamen* 2.5 3.2 0.0001Length of mature silique 14.8 14.7 0.85Width of mature silique 0.9 1.4 0.0001Length of long pedicel 13.0 16.7 0.0001Length of seed 0.46 0.53 0.0001

*Lengths of one sepal, one petal and one of the four long stamens weremeasured from a stage 15 ower from each of 20 REV and 22 rev-1 plantsfrom cohort 2 on day 38.

Measurements on mature siliques and pedicels were made on three of thelargest siliques and pedicels (ordinarily from the most basal oral branches onthe main orescence except when these were sterile) from each of 12 REV and 12 rev-1 plants from cohort 1 on day 48.

Five seeds were measured from basal siliques of each of 9 REV and 9 rev-1 plants from cohort 3 on day 80.

Table 4. Occurrence of oral branch types in REV and rev-1 plants

P valueFlorescence Floral branch type REV rev-1 of t -test

Main orescencesFertile owers 26.5 14 0.0049Sterile owers 2.9 22 0.0001Tapered laments 0 8 0.0022

CoorescencesFertile owers 22.1 8.0 0.0001Sterile owers 0.7 13.6 0.0001Tapered laments 0 12.2 0.0001

The mean numbers of fertile owers, sterile owers and tapered lamentson 13 REV and 12 rev-1 main orescences and on 52 REV and 29 rev-1coorescences from 48-day-old plants in cohort 1 are shown.


11/13


organs appeared to reach its extreme in the third type of oralbranch, the tapered laments. These resembled pedicels thatterminated development before forming any normal oralorgans. Both tapered laments and pedicels could be subtendedby knobs somewhat resembling leaf bases (Fig. 2H-J) that areabsent from wild-type pedicels. In addition to the knobs, thetapered laments could also be subtended or anked by addi-tional laments (Fig. 2J). The positions of these additionallaments are suggestive of homology to leaves and stipules.Some of the tapered laments bore branched trichomes. Thismay indicate that the laments had partial leaf identity;however, we have infrequently observed branched trichomeson wild-type pedicels. The phyllotaxy of the oral branchestended to be more irregular for the tapered laments than forrev-1 pedicels (Fig. 2H), with the internode distance oftenseverely reduced, as in the brush-like orescence in Fig. 2I.The excessive production of tapered laments was followed bythe premature termination of growth of the rev-1 orescenceapex.

Premature termination also occurred in other contexts that

are not well dened. Sporadic rev-1 orescences terminatedabruptly in a lament. In a signicant fraction of rev-3 plants,apical growth terminated only a few nodes above the rosette ina cluster of primordia (Fig. 2K). A similar phenotype wasobserved in preliminary experiments on rev-1 plants grown at29C (unpublished data). These observations indicate that the

REV gene has a role in the maintenance of the primary shootapical meristem.

To investigate the structures of the tapered laments and therev-1 primary apical meristem, we made sections through theapices of 17-day-old and 40-day-old REV and rev-1 plantsfrom cohort 2. No difference in size or organization of theprimary shoot apical meristem could be found between REV and rev-1 plants at day 17 (Fig. 6A,B). The densely cytoplas-

mic cells of the apical meristem stained deeply in bothgenotypes, as did the procambial strands. In contrast, the apicesof the 40-day-old REV and rev-1 plants (Fig. 6C,D) hadmarkedly different staining. The REV plant had a domed,deeply staining meristem surrounded by new oral primordiaand developing owers (Fig. 6C). The rev-1 plant had a moreattened apex with severely reduced staining and dramaticallylarger cells (Fig. 6D). The apex was surrounded by arrestedprimordia. We suggest that the rev-1 apical meristem cells hadstopped dividing and were far along in cell enlargementcompared to the dividing cells in the wild-type apex. We alsosectioned some other arrested rev-1 apices. Arrested primordiaand tapered laments with their subtending knobs exhibitedlight staining (Fig. 6E), indicating a lack of meristematic andvascular tissues and suggesting that the development of therev-1 oral primordia arrested quite early.

DISCUSSION

REV performs similar roles in vegetative andreproductive phytomersWe have recovered mutations in a gene, REVOLUTA , whichis necessary to regulate basic cell division patterns in Ara-bidopsis . The rev mutations are pleiotropic and affect all aerialparts of the plant. Although the effects of rev-1 on vegetative

and reproductive structures appear to be quite different, theyshare some notable similarities when they are examined in lightof the phytomer concept.

Both the vegetative and reproductive phytomers of rev-1mutants undergo excessive leaf or leaf-homologous growthand simultaneously have reduced branch structures. In the rev-1 vegetative phytomers, leaves grow too large and formationof many paraclades is drastically reduced to club-shapedobjects and laments. In the type 3 phytomers rev-1 mutantsgrow knobs, laments and stipule-like objects beneath theiroral branches in the positions homologous to those of leavesand stipules in vegetative phytomers. The oral branches areoften reduced to incomplete owers or tapered laments.Within the ower itself, oral organs are regarded as homolo-gous to leaves, and rev-1 oral organs are enlarged like rev-1leaves. Thus the same growth-altering processes appear to acton all phytomers. This contrasts with other known mutationsaffecting axillary meristem development such as the tomatomutants lateral suppressor (Malayer and Guard, 1964) andtorosa-2 (Mapelli and Kinet, 1992), which affect branchformation in only a specic subset of vegetative phytomers andhave no reciprocal effect on leafy organ growth.

REV limits the growth of non-apical meristemsAt least two disparate aspects of the Rev phenotype, the over-growth of leaves and the thickening of the stems, are due tothe presence of extra cells. This is compelling evidence that

REV is involved in limiting cell division. Extra cell divisionsin rev-1 leaves are largely conned to the base of the leaf asin the wild type. The extra cell layers in rev-1 stems have nocounterpart in greenhouse-grown wild-type plants, but are notdue to a generalized deregulation of cell division. The apparentderivation of the extra cells from a cambium-like layer mayreveal a potential for secondary growth in Arabidopsis stems.

Secondary growth has been observed in the roots of wild-typeplants (Dolan et al., 1993). A structure resembling a fascicularcambium is visible in wild-type stems. A cryptic interfascicu-lar cambium may also be present in wild-type stems and bothcambia may be deregulated in rev-1 mutants. If this is true,both leaf and stem growth in rev-1 plants may arise from theenhancement of normal meristematic activities.

REV is required to maintain apical meristem growthAlthough rev-1 mutants have enhanced cell division in leavesand stems, they have a defect in apical (including oral)meristem activity that results in the premature termination of the rev-1 shoot apex and in the formation of abnormal orincomplete structures in place of paraclades and owers. These

incomplete structures imply the defect is in apical meristemorganization or maintenance, in contrast to the failure toinitiate meristematic growth that is seen in empty wild-typeaxils. There is a hierarchy to the probability of apical meristemfailure: the primary apical meristem is least likely to fail, whilerst-order paraclades and owers fail more frequently and thesecond-order paraclades fail even more frequently (Table 1).This suggests that the partitioning of some resource mayinuence the probability of meristem failure.

Models of REV functionWhat is the role of REV in maintaining apical meristems? It isunlikely that REV is required for general cell viability or


12/13

2734

division, since rev-1 leaves and stems grow excessively andrev-1 plants are vigorous. Although the effects of rev-1mutations on growth resemble the pleiotropic actions of phy-tohormones, the particular suite of defects in rev-1 mutants arenot easily interpretable in terms of the known actions of any

one hormone. REV appears to regulate cell proliferation withopposing effects on apical versus non-apical meristems.Explaining these opposing effects is crucial to understanding

REV function.One possibility is that REV controls the partitioning of

nutrients or growth factors between apical and non-apicalmeristems, either by direct regulation of nutrient allocation orby indirect re-allocation to non-apical meristems caused bylack of competition for nutrients from the failed apicalmeristems. This model requires the unusual assumption thatthe formation of de novo stipules and knobs below oralbranches is normally limited by nutritional partitioning.Another possibility is that REV directs meristematic cells to be

incorporated into the axillary (or oral) meristem instead of contributing to the formation of leaves (or stipules, knobs andlaments). This possibility is particulary interesting becauseaxillary meristems are clonally related to their subtendingleaves in Arabidopsis (Furner and Pumfrey, 1992; Irish and

Sussex, 1992). However, it is difficult to explain the over-growth of cotyledons and oral organs by this model. A thirdpossibility is that REV could promote or inhibit cell divisionaccording to the presence of cofactors that differ betweenapical and non-apical meristems. Finally, REV could controlthe production of or response to a morphogen necessary tomaintain apical growth and to inhibit non-apical growth.Regardless of which model is correct, REV clearly hasprofound regulatory effects on the development of meristems.Together with other mutations affecting meristem growth, therev mutations offer the exciting possibility of dissecting theprocesses involved in the regulation and interactions of meristems that result in plant morphogenesis.


Fig. 6. Longitudinal sections of primary orescence apices. (A,B) Apical meristems of 17-day-old plants: (A) REV , (B) rev-1 . (C,D) Apices of 40-day-old plants: (C) REV , (D) rev-1 . (E) Backlit rev-1 arrested apex. Bars, (A-D) 20 m; (E) 100 m. f, ower or oral primordium; kt,knobs and tapered laments; m, apical meristem.


13/13


We thank Laura Conway, Elliot Meyerowitz, David Smyth and RyMeeks-Wagner for supplying seeds. We thank Kevin Lease, AnnePaul, Arp Schnittger, Gene Tanimoto, Doug Ewing and the green-house staff for technical assistance. This work was supported by Uni-versity of Washington Royalty Research Fund 629 to L. C. andNational Institute of Health Postdoctoral Fellowship GM14355-02 toH. T. A.

REFERENCES

Alvarez, J. (1994). The SPITZEN gene. In Arabidopsis: An Atlas of Morphology and Development (ed. J. Bowman), pp. 188-189. New York:Springer-Verlag.

Alvarez, J., Guli, C. L., Yu, X.-H. and Smith, D. R. (1992). terminal ower: agene affecting inorescence development in Arabidopsis thaliana . Plant J. 2,103-116.

Bell, C. J. and Ecker, J. R. (1994). Assignment of 30 microsatellite loci to thelinkage map of Arabidopsis. Genomics 19 , 137-1444.

Caruso, J. L. (1968). Morphogenetic aspects of a leaess mutant in tomato. I.General Patterns in development. Amer. J. Bot. 55 , 1169-1176.

Dolan, L., Janmaat, K., Willemsen, V., Linstead, P., Poethig, S. andRoberts, K. (1993). Cellular organisation of the Arabidopsis thaliana root.

Development 119 , 71-84.Furner, I. J. and Pumfrey, J. E. (1992). Cell fate in the shoot apical meristem

of Arabidopsis thaliana . Development 115 , 755-764.Galinat, W. C. (1959). The phytomer in relation to the oral homologies in the

American Maydea. Bot. Mus. Lea., Harv. Univ. 19 , 1-32.Garrison, R. (1955). Studies in the development of axillary buds. Am. J. Bot.

42 , 257-266.Goethe, J. W. von (1790). Versuch die Metamorphose der Panzen zu

Erklren . Gotha: C. W. Ettinger.Hempel, F. D. and Feldman, L. J. (1994). Bi-directional inorescence

development in Arabidopsis thaliana : acropetal initiation of owers andbasipetal initiation of paraclades. Planta 192 , 276-286.

Irish, V. F. and Sussex, I. M. (1992). A fate map of the Arabidopsis embryonicshoot apical meristem. Development 115 , 745-753.

Johansen, D. A. (1940). Plant Microtechnique . New York and London:McGraw-Hill Book Company, Inc.

Konieczny, A. and Ausubel, F. M. (1993). A procedure for mappingArabidopsis mutations using co-dominant ecotype-specic PCR-basedmarkers. Plant J. 4, 403-410.

Koornneef, M. and Stam, P. (1992). Genetic analysis. In Methods in Arabidopsis Research (ed. C. Koncz, N.-H. Chua and J. Schell), pp. 83-99.Singapore: World Scientic.

Majumdar, G. P. (1942). The organization of the shoot in Heracleum in thelight of development. Ann. Bot. 6, 49-81.

Malayer, J. C. and Guard, A. T. (1964). A comparative developmental studyof the mutant sideshootless and normal tomato plants. Am. J. Bot. 51 , 140-143.

Mapelli, S. and Kinet, J. M. (1992). Plant growth regulator and graft control of axillary bud formation and development in the TO-2 mutant tomato. Pl.Growth Reg. 11 , 385-390.

Neff, M. M. and Van Volkenburgh, E. (1994). Light-stimulated cotyledonexpansion in arabidopsis seedlings. Plant Physiol. 104 , 1027-1032.

Pyke, K. A., Marrison, J. L. and Leech, R. M. (1991). Temporal and spatialdevelopment of the cells of the expanding rst leaf of Arabidopsis thaliana L.Heynh. J. Exp. Bot. 42 , 1407-1416.

Schultz, E. A. and Haughn, G. W. (1991). LEAFY , a homeotic gene thatregulates inorescence development in Arabidopsis. Plant Cell 3, 771-781.

Smyth, D. R., Bowman, J. L. and Meyerowitz, E. M. (1990). Early owerdevelopment in Arabidopsis. Plant Cell 2, 755-767.

Snow, M. and Snow, R. (1942). The determination of axillary buds. NewPhytol. 41 , 13-22.

Steeves, T. A. and Sussex, I. M. (1989). Patterns In Plant Development .Cambridge: Cambridge University Press.

Weberling, F. (1989). Morphology of Flowers and Inorescences . Cambridge:Cambridge University Press.

Weigel, D., Alvarez, J., Smyth, D. R., Yanofsky, M. F. and Meyerowitz, E.M. (1992). LEAFY controls oral meristem identity in Arabidopsis. Cell 69 ,843-859.

White, J. (1984). Plant metamerism. In Perspectives In Plant Population Ecology (ed. R. Dirzo and J. Sarukhan), pp. 15-47. Sunderland, MA: SinauerAssociates Inc.

Woolson, R. F. (1987). Statistical Methods for the Analysis of Biomedical Data . New York: John Wiley and Sons.

(Accepted 24 June 1995)

Date post:	02-Jun-2018
Category:	Documents
Upload:	atika-ayu-kusumaningtyas
View:	218 times
Download:	0 times

Development 1995 Talbert 2723 35

Documents