Abstract
The narrow sheath mutant of maize displays a leaf and plant stature phenotype controlled by the duplicate factor mutations narrow sheath1 and narrow sheath2. Mutant leaves fail to develop a lateral domain that includes the leaf margins. Genetic data are presented to show that the narrow sheath mutations map to duplicated chromosomal regions, reflecting an ancestral duplication of the maize genome. Genetic and cytogenetic evidence indicates that the original mutation at narrow sheath2 is associated with a chromosomal inversion on the long arm of chromosome 4. Meristematic sectors of dual aneuploidy were generated, producing plants genetically mosaic for NARROW SHEATH function. These mosaic plants exhibited characteristic half-plant phenotypes, in which leaves from one side of the plant were of nonmutant morphology and leaves from the opposite side were of narrow sheath mutant phenotype. The data suggest that the narrow sheath duplicate genes may perform ancestrally conserved, redundant functions in the development of a lateral domain in the maize leaf.
THE maize leaf is a particularly tractable organ for studies of developmental genetics. Comprised of the basal sheath, which is separated from the distal blade by the ligule-auricle boundary, the maize leaf is the target of numerous, domain-specific mutations. These include mutations affecting the development of the following: lateral veins (Hake and Freeling 1986; Sinha and Hake 1990); the ligule-auricle region (Becraftet al. 1990; Sylvesteret al. 1990; Becraft and Freeling 1994; Schneeberger et al. 1995, 1998; Harper and Freeling 1996; Morenoet al. 1996; Walshet al. 1997); the midrib (Fowler and Freeling 1995; Schichness and Freeling 1997); the margins (Bertrand-Garcia and Freeling 1991; Scanlonet al. 1996; Scanlon and Freeling 1997); and the dorsal/ventral and proximal/distal axes (Muelhbaueret al. 1997; Schneebergeret al. 1998; Timmermans et al. 1998, 1999; Tsiantiset al. 1999). These analyses have provided a model for maize leaf development whereby different domains of the leaf are independently assigned discrete developmental programs (Freeling 1992). In this view, the leaf is a mosaic of different domains that are coordinately assembled along three axes into a composite organ.
The leaf is one component of the repeating vegetative segment or phytomer, comprised of the leaf, node, internode (stem), and axillary bud (Sharman 1942; Galinat 1959). Successive phytomers are initiated from lateral regions of the shoot apex in a distichous phyllotaxy, 180° apart and in two ranks. Clonal analyses have revealed that the maize phytomer develops from ~250 founder cells, recruited from the shoulders of the shoot apical meristem (Poethig 1984). Founder cell recruitment begins on one flank of the meristem (the eventual midrib) and proceeds laterally around the apex until the margins of the incipient leaf overlap the opposite flank of the meristem (Figure 1B; Steffenson 1968; Poethig 1984; Scanlon and Freeling 1997). This pattern of founder cell initialization is mirrored by the downregulation of KNOX (KNotted1-like homeobOX) proteins in maize meristems (Smithet al. 1992; Jacksonet al. 1994). The border of KNOX protein accumulation/downregulation is thought to demarcate the boundary between leaf/meristem identity in the incipient maize phytomer. A clonal sector induced in the founder cell domain of the shoot apical meristem (SAM) will mark a single maize phytomer (Poethig and Syzmkowiak 1995). However, owing to the distichous phyllotaxy and overlapping margins of founder cell-staged leaves, a sector located above the founder cell region of the SAM will necessarily intersect at least two successive maize phytomers. These factors render maize especially suitable to clonal analyses of leaf development.
Cell lineage maps of the maize meristem have shown that meristematic cells have predictable destinies. A strong correlation is documented between the lateral position of a shoot meristematic cell and the particular leaf domain into which that cell will divide (Steffenson 1968; Poethig 1984; Scanlon and Freeling 1997). Precise fate mapping of maize phytomer development has revealed that the individual components of the phytomer (i.e., the leaf, the subtending stem, and the axillary bud) do not become clonally separate compartments until primordial stages, when the phytomer is 4–6 mm in length (Sylvesteret al. 1990; Poethig and Syzmkowiak 1995). Differentiation of the incipient phytomer proceeds basipetally, from tip to base. Therefore, the tip of the leaf blade develops first, whereas the opposing flank of the stem develops last (Poethig 1984; Poethig and Syzmkowiak 1995).
The narrow sheath mutant phenotype is a deletion of a lateral domain that includes the margins of the lower leaf. (A) Seedling leaf 4 from a narrow sheath mutant (left) and a nonmutant sibling (right) plant. (B–D) A model for NARROW SHEATH function during recruitment of leaf founder cells in a lateral domain shown in yellow. (B) Model of leaf founder cell recruitment from the SAM, beginning in the midrib domain and proceeding to the margin domain on the opposite flank of the SAM. Previous analyses of KNOX downregulation in narrow sheath mutant founder cells and fate mapping of mutant meristems indicate that a lateral domain of founder cells (shown in yellow) is not recruited during development of narrow sheath mutant leaves. The resulting mutant primordium (C) is missing a large lateral domain that includes the margins of nonmutant sibling leaves (D). See text for further explanation. s, sheath; b, blade; L, ligule-auricle; md, midrib; mr, margin.
The narrow sheath mutant was first described as a pattern deletion phenotype affecting the lower leaf and stem (Figure 1; Scanlonet al. 1996). Structures normally found at the nonmutant leaf edge are absent from the affected region of the mutant leaf and leaf homologous organs (Scanlon and Freeling 1998). The narrow sheath mutant phenotype is inherited as a duplicate factor trait, conferred by homozygosity for each of the unlinked mutations narrow sheath1 and narrow sheath2 (Scanlonet al. 1996). Immunohistochemical analyses of the narrow sheath mutant SAM have shown that the downregulation of KNOX proteins, meristematic markers of the maize leaf/nonleaf boundary (Smithet al. 1992; Jacksonet al. 1994), is disrupted in a lateral domain of the mutant founder cells that forms the leaf margin in nonmutant plants. Fate mapping of the mutant meristem revealed that founder cells in this same meristematic region are not recruited to form the ns mutant leaf (Scanlon and Freeling 1997). These data have contributed to a model for NARROW SHEATH function during founder cell recruitment in a lateral domain that forms the margins of the lower leaf blade, the sheath, and the stem (as shown in Figure 1).
In this article we show that the narrow sheath mutations map to ancestrally duplicated regions of the maize genome and provide genetic and cytogenetic evidence that the ns2 mutation is linked to a chromosomal inversion. Furthermore, we exploit the unique properties of the maize B-A translocations in order to generate meristematic sectors of narrow sheath mutant tissue in otherwise nonmutant plants. These mosaic plants provide evidence that both of the narrow sheath mutations are null alleles. Moreover, despite the advance of over 11 million years of evolutionary time since the duplication of the maize genome (Gaut and Doebley 1997) the narrow sheath genes perform redundant functions during maize leaf development. NARROW SHEATH function is not required for the development of central domains of the phytomer. These data further support a model for compartmentalized development in plant shoots (Freeling 1992; Scanlonet al. 1996).
Placement of the narrow sheath mutations using B-A translocations
MATERIALS AND METHODS
Genetic stocks and B-A translocation mapping: The narrow sheath1-O (ns1-O) and narrow sheath2-O (ns2-O; Scanlonet al. 1996) mutations were kindly donated by E. Elsing and M. Albertson (Pioneer Hi-Bred International, Johnston, IA). The narrow sheath mutant phenotype displays single factor inheritance in the genetic background in which the original ns mutations were identified (i.e., the ns1:1 line; Scanlonet al. 1996), indicating that one of the duplicate pair ns-O mutations is homozygous in this specific line. In all other genetic backgrounds investigated, the narrow sheath mutant phenotype is inherited as a duplicate factor trait (Scanlonet al. 1996).
A series of endosperm-marked (Beckett 1994) B-A translocations (Table 1) was provided by D. S. Robertson and P. S. Stinard (Iowa State University). The B-A translocations of maize are efficient tools used to map recessive mutations to a specific chromosome arm (Roman and Ulstrup 1951; Beckett 1978). The particularly useful feature of the supernumerary B chromosomes of maize is the high rate of nondisjunction of the B centromere during the second mitotic division of the microspore following meiosis (Randolph 1941; Roman 1947). Therefore, plants containing a single B chromosome produce pollen grains with nonconcordant sperm nuclei; one sperm nucleus may contain two B chromosomes (i.e., hyperploid) whereas the other nucleus will contain no B chromosomes (hypoploid). When a “normal” A chromosome and a supernumerary B chromosome undergo reciprocal translocation, nonconcordant sperm nuclei are generated that contain either two copies (hyperploid) or no copies (hypoploid) of the genes translocated to the B centromere. Because such hypoploid, F1 progeny can permit phenotypic expression of recessive mutations distal to the translocation, B-A translocation stocks are especially effective tools for mapping mutations to a particular chromosome arm. To identify the chromosome arm locations of both narrow sheath loci, a modification of the strategy described in Beckett (1978) was employed.
Plants hyperploid for particular B-A translocations were crossed as male onto plants containing ns1-O and ns2-O. The parental, hyperploid plants were heterozygous for the B-A translocation and a specific endosperm marker gene in repulsion (Beckett 1994). Therefore, the general genotype of the hyperploid male plants used in the mapping strategy was A-B, B-AM, B-AM/Am, where M and m represent dominant and recessive alleles of chromosome-arm-specific endosperm markers, respectively.
Putative hypoploid plants were selected among the F1 progeny. Typically, hypoploid plants exhibit ~50% pollen sterility due to the abortion of microspores deficient for genes contained on the B-AM chromosome (described in Beckett 1978). Moreover, plants heterozygous for the ns2-O allele produce ~25% aborted pollen owing to the presence of a linked paracentric inversion (see results). Further complications to the use of pollen abortion to select F1 hypoploid plants are introduced by the use of the compound translocation TB-1Sb-2L4464 (Table 1, Rahka and Robertson 1969), which conditions ~50% pollen abortion in plants heterozygous for the B-A chromosome (Beckett 1994). However, plants hypoploid for this compound translocation are relatively small in stature, because they are hemizygous for genes contained on two different chromosome arms (M. J. Scanlon, unpublished observation; Beckett 1978, 1994). In this study, therefore, putative hypoploid plants were identified among F1 progeny by selecting individuals that segregated for exceptionally high levels of pollen abortion as compared to sibling plants (assayed in the field using a pocket microscope) and/or plant phenotypes associated with arm-specific hypoploidy (described in Beckett 1978). Putative hypoploid plants were backcrossed to narrow sheath mutant plants, and the resultant progeny were screened for segregation of the narrow sheath mutant phenotype. Hypoploid F1 plants heterozygous for ns1 and ns2 are expected to segregate 25% mutant plants when backcrossed to narrow sheath mutant plants if neither ns mutation is linked to the hypoploid chromosome arm. However, if one of the ns loci is linked to the hypoploid chromosome arm, backcrossed progeny are expected to segregate ~50% narrow sheath mutant plants because hypoploid megaspores are not transmitted (Table 1).
RFLP recombination data for narrow sheath1
The map positions of the ns loci were confirmed and mapped to higher resolution by conventional restriction fragment length polymorphism (RFLP) analyses of F2 narrow sheath mutant plants generated following outcrosses to the standard inbred lines B73, Mo17, and W23 (Tables 2 and 3). All RFLP probes utilized in these analyses were obtained from T. Muskett and E. Coe (University of Missouri).
Transposon tagging an independent allele of ns2: A second allele of ns2 (designated ns2*Mu77) was obtained in a modified, directed transposon-tagging experiment using Robertson's Mutator (Robertson 1978). Mutant plants were crossed onto a line containing Mutator activity (Robertson 1978), and nonmutant F2 plants were selected from the progeny. F2 plants that segregated no mutant progeny following a second round of self-pollination were used in RFLP analyses of markers linked to ns1 and ns2 (see results, Tables 2 and 3) in order to identify an F2 plant of the inferred genotype ns1-O/ ns1-O; Ns2-Mu/Ns2-Mu. Once identified, Mu-active progeny of this F2 plant (designated G11-2) were crossed as female by pollen from narrow sheath mutant plants (genotype ns1-O/ns1-O; ns2-O/ns2-O). Among >5500 F1 progeny of this directed tagging experiment, a single mutant plant was identified of the presumed genotype ns1-O/ns1-O; ns2-O/ns2*Mu77. RFLP analyses using markers tightly linked to ns2 were used to identify plants harboring the new mutant allele, ns2*Mu77, in subsequent crosses to the inbred lines B73 and Mo17. Mutant plants homozygous for the new ns2 mutation (genotype ns1-O/ns1-O; ns2*Mu77/ns2*Mu77) are identical in phenotype to ns-O mutant plants (genotype ns1-O/ns1-O; ns2-O/ns2-O).
RFLP recombination data for narrow sheath2
Identification of the homozygous ns mutation in nonmutant siblings of ns-O mutant plants: To determine which ns-O locus is homozygous in the ns1:1 line (described above and in Scanlonet al. 1996), true-breeding, nonmutant plants were generated by self-pollinating nonmutant individuals in the ns1:1 line. These plants were inferred to be homozygous for one ns mutation and homozygous nonmutant for the other ns locus. Because the ns1:1 line is so highly introgressed, it is not practical to use mutant and nonmutant plants produced within this line in RFLP analyses to distinguish which ns mutation is homozygous in the true-breeding nonmutant individuals. Therefore, RFLP markers linked to ns1 and ns2 were employed to identify an F2 plant of the genotype ns1-O/Ns1-Mu; ns2-O/ns2-O among the F2 progeny obtained by self-pollination of F1 double heterozygous plants of the genotype ns1/Ns1-Mu; ns2/Ns2-Mu (see above). True-breeding, nonmutant plants generated from the ns1:1 line were crossed to plants of the genotype ns1-O/Ns1-Mu; ns2-O/ns2-O (i.e., heterozygous for ns1-O, homozygous for the ns2-O mutation). Subsequently, the F1 progeny of this cross were self-pollinated, and the F2 seedlings were screened for segregation of the ns mutant phenotype.
DNA hybridization analyses: DNA was isolated from maize tissue samples using a urea-based protocol described by Chen and Dellaporta (1994). DNA-DNA gel-blot analyses were performed according to the procedures described by Southern (1975), with changes as described in Scanlon and Myers (1998). DNA fragments used as probes were isolated in 0.8% low-melt agarose (GIBCO BRL, Gaithersburg, MD) and radio-labeled using 32P (ICN) and random oligonucleotide primers (Pharmacia, Piscataway, NJ).
Cytology and microscopy: For cytological investigations, tassels containing meiocytes undergoing meiosis were dissected from ns/w23 heterozygous plants and fixed in ethanol/glacial acetic acid solution (3:1) according to the methods described in Dempsy (1994). Meiocytes fixed during anaphase I and pachytene were stained in 2% acetocarmine (Dempsy 1994) and examined with an Axioplan II microscope (Zeiss, Thornwood, NY) using a 100× objective lens.
For analyses of pollen abortion, pollen obtained from fresh, unextruded anthers was examined under low power with a Zeiss Stemi 2000 dissecting microscope. Pollen was examined from at least six F1 plants of each of the following 10 genotypes: (1) W23, (2) Mo17, (3) B73, (4) ns1-O/ns1-O; ns2-O/ns2-O × W23, (5) ns1-O/ns1-O; ns2-O/ns2-O × Mo17, (6) ns1-O/Ns1; ns2-O/ns2-O × W23, (7) ns1-O/Ns1; ns2-O/ns2-O × Mo17, (8) Ns1/Ns1; ns2-O/ns2-O × W23, (9) Ns1/Ns1; ns2-O/ns2-O × Mo17, and (10) ns1-O/ns1-O; ns2-O/ns2*Mu77 × B73.
X-irradiation of seed and analyses of sectored plants: 1030 F1 seed obtained from the cross of TB-4Lf hyperploid plants onto narrow sheath mutant females were imbibed overnight and subjected to 1017 rad of X-irradiation over 3.2 min, through a 0.35-mm Cu filter with a Philips RT250 X-ray machine run at 225 kV. Irradiated seeds were hand-planted and grown to maturity in Santa Clara, CA. Sectored leaves were excised from the plants and photographed; tissue samples were taken for use in DNA gel-blot analyses (described above).
RESULTS
Genetic mapping of the narrow sheath duplicate genes: Expression of the narrow sheath mutant phenotype requires homozygosity at two unlinked, recessive mutations (Scanlonet al. 1996). Therefore, a modified approach to traditional B-A translocation mapping (described in materials and methods) was employed in order to place both of the ns mutations to a particular chromosome arm (Table 1). The mapping data indicate that ns1 is located on the long arm of chromosome 2, whereas ns2 is located on the long arm of chromosome 4.
RFLP analyses were conducted on F2 narrow sheath mutant progeny obtained following outcrossing to the nonmutant inbred lines W23, B73, and Mo17. Recombination analyses of ns1 and RFLP markers found on chromosome arm 2L confirmed the map position of ns1 inferred from B-A translocation mapping (Table 2) and located ns1 approximately midway between the markers umc5a and csu270 (Figure 2). Likewise, RFLP mapping of ns2 confirmed the placement of this second mutation on chromosome arm 4L, although no F2 progeny were recovered that had undergone recombination between the ns2 mutation and an ~12.3-cM expanse of chromosome 4 encompassing the region between markers bnl10.05 and umc52 (Table 3, Figure 2).
To determine whether or not the lack of recovery of viable, recombinant progeny in this region of chromosome 4 is caused by mutation at the ns2 locus, a second independently induced mutant allele of ns2, designated as ns2*Mu77 (see materials and methods for a description of the generation of this allele) was analyzed for recombination in portions of this 12.3-cM region. Plants homozygous for ns2*Mu77 display a phenotype that is identical to plants homozygous for ns2-O. The genetic data (Table 3, Figure 2) reveal that F2 progeny are recovered that exhibit recombination between the ns2*Mu77 allele and two RFLP loci within the 12.3-cM region affected in the ns2-O allele. Therefore, viable gametes exhibiting recombination between ns2-O and linked marker loci in a 12.3-cM region of chromosome 4L are not recovered, although viable gametes are recovered that have undergone recombination between ns2*Mu77 and the same linked loci. These results indicate that mutation at the ns2 locus does not condition the mortality of recombinant gametes described above. Consequently, we suspected that some cytogenetic anomaly, such as a chromosomal inversion (described below), is the probable cause of the aberrant RFLP recombination data obtained for the ns2-O mutation.
RFLP recombination maps of the narrow sheath loci. The duplicate factor genes ns1 and ns2 map to duplicated regions on the long arms of chromosomes 2 and 4 and in the vicinity of a second duplicate factor pair, whp1 and c2. The recombination data for two independently isolated alleles of ns2 yield disparate recombination data in an interval of ~13 cM (shown bracketed by double arrows). The blue numbers to the left of the chromosomes indicate the recorded recombination interval distances between successive core RFLP loci (in boxes), as reported on the 1999 maize RFLP working map (www.agron.missouri.edu/map.html). The red numbers to the right of the RFLP markers indicate the recombination data for that particular RFLP locus and the ns-O mutations as given in Tables 2 and 3. The green numbers in parentheses indicate RFLP recombination data between that RFLP locus and the ns2-Mu77 allele, as given in Table 3. All recombination values are given in centimorgans. The dark boxes indicate approximate regions of centromere. Not drawn to scale.
Previous analyses (Scanlonet al. 1996) have identified a maize stock derived from the original ns mutant family, which was designated the ns1:1 line (described in materials and methods). When ns mutant plants are backcrossed to nonmutant siblings from the ns1:1 line, the progeny segregate ns mutant and nonmutant progeny in a ratio approximating 1:1 (Scanlonet al. 1996). These genetic data demonstrate that in the ns1:1 line, nonmutant plants are homozygous for one ns mutation and heterozygous for the other ns mutation. To determine which ns locus is homozygous in the ns1:1 line, true-breeding, nonmutant plants (i.e., homozygous for one ns mutation and homozygous nonmutant at the other ns locus) were generated from the ns1:1 line by two successive generations of self-pollination of nonmutant plants. Once identified, true-breeding nonmutant plants were crossed to plants of the genotype ns1-O/Ns1-Mu; ns2-O/ns2-O (i.e., heterozygous for ns1-O, homozygous for the ns2-O mutation as determined by RFLP analyses). Following self-pollination of 10 F1 progeny plants, 4 plants segregated ~25% F2 mutant progeny (putative F1 genotype ns1-O/Ns1; ns2-O/ns2-O), whereas 6 plants segregated no F2 mutant progeny (putative F1 genotype Ns1-Mu/Ns1; ns2-O/ns2-O). These data reveal that the ns1:1 line is homozygous for the ns2-O mutation.
Analysis of pollen viability in F1 plants crossed to W23 and Mo17
Genetic and cytogenetic analyses of pollen formation in ns heterozygous plants: Genetic and cytogenetic investigations were performed to determine the cause of the aberrant RFLP recombination data obtained for the ns2-O allele (Figure 2, Table 3). Plants of four different genotypes [(1) ns1-O/ns1-O; ns2-O/ns2-O (original narrow sheath mutant), (2) Ns1-O/ns1-O; ns2-O/ns2-O (nonmutant sibling in 1:1 line), (3) Ns1-O/Ns1-O; ns2-O/ns2-O (true-breeding nonmutant generated by self-pollination of nonmutant siblings in 1:1 line above), and (4) ns1-O/ns1-O; ns2*Mu77/ns2-O (narrow sheath mutant that is heterozygous for two independently isolated mutations at ns2 (see materials and methods)] were crossed to the nonmutant inbred lines W23 and B73. Pollen abortion was scored in at least six F1 plants of all individual genotypes, as well as in Mo17, B73, and W23 inbred plants. As shown in Table 4, all F1 individuals generated from outcrosses of original narrow sheath mutants, nonmutant siblings, and true-breeding nonmutant parents displayed ~25% pollen abortion. In contrast, half of the F1 plants generated by crossing B73 to ns1-O/ns1-O; ns2*Mu77/ns2-O displayed ~25% pollen death. Further analyses revealed that F1 plants heterozygous for ns2-O (as determined by RFLP analysis) were characterized by 25% pollen abortion, whereas F1 plants heterozygous for ns2*Mu77 were not. Likewise, B73, Mo17, and W23 inbreds did not exhibit appreciable pollen mortality. These data indicate that a genetic factor causing pollen abortion is linked to the ns2-O mutation. Furthermore, this factor segregates independently from both ns1-O and ns2*Mu77. Moreover, semisterility is observed only in plants heterozygous for ns2-O; no abnormal pollen mortality was observed in ns2-O homozygous plants.
Cytogenetic analyses of meiotic anaphase figures
Cytogenetic investigations of meiocytes isolated from ns-O mutant/W23 heterozygous plants were performed to determine the cause of the pollen semisterility in ns2-O heterozygous plants. As shown in Table 5, 26.5% of anaphase I figures examined from ns/W23 heterozygous plants contained chromosome bridges and fragments, or bridges alone (Figure 3). No anaphase bridges were observed in meiocytes examined from ns mutant homozygotes (ns1-O/ns1-O; ns2-O/ns2-O), nonmutant siblings from the 1:1 line (ns1-O/Ns1; ns2-O/ns2-O), or W23 inbred plants. Furthermore, analyses of pachytene figures from ns/W23 heterozygous plants reveal the presence of inversion loops (Figure 3, D and E) interpreted to represent an inverted segment on chromosome 4. Taken together, the RFLP recombination data, pollen viability studies, and cytogenetic analyses provide compelling evidence that a paracentric inversion encompassing ~12–13 cM of chromosome arm 4L is linked to the ns2-O mutation.
Cytogenetic analysis of microsporocytes in nonmutant and narrow sheath heterozygous plants. (A) Pachytene-staged chromosomal figures in nonmutant W23 plants. (B) Anaphase I figures in nonmutant W23. (C) At anaphase I, ns/W23 heterozygous plants display high frequencies of chromosomal bridge formation (arrow). (D) Pachytenestaged chromosomal squash of ns/W23 heterozygote reveals an inversion loop (arrow), which is interpreted to represent inverted portions of chromosome arm 4L. (E) Enlarged image of inversion loop. N, nucleolus.
Leaf phenotypes of narrow sheath mosaic plants. (A) Successive leaves 7, 8, and 9 from mosaic plant 2 alternate between narrow and wide leaf phenotypes. Note that the edges of the sheath (arrow) in wide leaf 8 overlap; arrowheads indicate the left and right blade margins of wide leaf 8, whereas double arrowheads denote the blade margins of narrow leaf 7. (B) Wide leaf 9 (left) and narrow leaf 10 (right) from mosaic plant 6. The margins of the narrow leaf contain yellow-green sectors (arrows). (C) Mosaic plant 1. Wide leaf 9 (left) has a midrib sector (arrows), whereas narrow leaf 8 (right) has a margin sector (arrows).
Mosaic narrow sheath phenotypes generated via X-ray treatment of hypoploid seed: We endeavored to test our model for NARROW SHEATH function during development of a lateral domain of the leaf by inducing the random, sectored loss of the nonmutant Ns1 allele in hypoploid plants already hemizygous for ns2-O. Therefore, narrow sheath plants were first crossed by TB-4Lf hyperploid plants; the F1 seed generated from this cross are expected to segregate for hypoploid progeny that are hemizygous for ns2-O. A total of 425 F1 mature seeds were X-rayed in order to induce random breakage of chromosome arms. Eight remarkable plants (plants 1–8, Figures 4 and 5) in this population exhibited a “narrow sheath mosaic” phenotype, characterized by leaves with the narrow sheath mutant phenotype on one side of the plant and normal leaves on the other side of the plant.
All the mosaic plants recovered following X-irradiation of mature seed displayed completely normal leaf width in at least the first four leaves. Maize seed typically form four to six leaf primordia prior to the onset of meristem quiescence and kernel maturation (Randolph 1936). As shown in Figure 4 and modeled in Figure 5, each of the mosaic plants displayed yellow-green, meristematic sectors that spanned multiple leaves (Steffenson 1968; Scanlon and Freeling 1997). We do not understand why the sectors were yellow-green; however, as described below, the yellow-green phenotype is correlated with dual hemizygosity for genes located on both the long arm of chromosome 2 and the long arm of chromosome 4.
In all of these chimeric plants, the narrow sheath leaves had yellow-green sectors astride the leaf edges (Figure 4, B and C), whereas the interspersed leaves of normal width displayed yellow-green sectors over the midrib region. The midrib sectors were especially apparent in the distal regions of the affected leaves (Figure 4C). Six of the eight plants had two or more narrow sheath phenotypic leaves on one side of the plant, inter-spersed with one or more wide leaves on the opposite side of the plant. In two plants (plants 5 and 8, Figure 5), the narrow leaf phenotype was restricted to one-half of the leaf only. In both of these samples, the wide leaves succeeding the half-narrow leaves contained yellow-green sectors that did not straddle the midrib, but instead marked only one side of the leaf with respect to the central midrib.
Microscopic examination of the mosaic half-plants revealed that the narrow leaves lacked tapered leaf edges and marginal hairs and contained ~50% fewer lateral veins as the wide leaves found on the left side of the plant. Therefore, the narrow leaves produced on these mosaic plants portrayed all aspects of the narrow sheath leaf mutant phenotype (Figure 4; Scanlonet al. 1996).
Normal leaf width was restored in all of the X-irradiated mosaic plants after leaf 10, concurrent with the absence of yellow-green sectored tissues in the upper phytomers of these plants. Plants 7 and 8 displayed necrotic upper leaves and failed to develop functional tassels. Tassel development was normal in mosaic plants 1–6, although all of these plants exhibited substantially higher levels of pollen abortion than expected in euploid or hyperploid plants heterozygous for the ns-O mutations (see materials and methods and above). When crossed as male to ns mutant ears, the progeny of plants 1–6 segregated ~50% ns mutant offspring. These data are consistent with the hypothesis that the nonsectored tissues of mosaic plants 1–6 are heterozygous for ns1-O, hypoploid for chromosome arm 4L (due to nondisjunction of the B-A chromosome 4Lf), and therefore hemizygous for ns2-O.
Cartoons of narrow sheath mosaic leaves. The phenotypes of successive leaves on eight individual plants mosaic for midrib-margin sectors. Leaf numbers were allocated by assigning the leaf nearest the base of the plant the designation L1. See text for further explanations of the generation of mosaic plant phenotypes.
DNA gel-blot analyses indicate that X-irradiated plants are mosaic for ns1 alleles: DNA was extracted from plant tissues dissected from mosaic plant 1 and examined in DNA gel-blot analyses using an RFLP probe linked to the ns1 locus (Figure 6). Two different domains of successive leaves from plant 1 were sampled: the green, unsectored margin of wide leaf 9 and the yellow-green, sectored margins of mutant leaf 8. These data indicate that plant 1 is mosaic for alleles of ns1. Specifically, the unsectored margin of leaf 9 was heterozygous for the ns1-O mutation (genotype ns1-O/Ns1; ns2-O/−− TB-4Lf), whereas the marginal sector was hemizygous for the ns1 mutation (genotype ns1-O/−; ns2-O/−− TB-4Lf).
Southern gel-blot analyses of mosaic plant tissue. Samples isolated from mosaic plant 1. Lane 1 contains DNA prepared from the nonsectored margin of leaf 9. Lane 2 contains DNA prepared from the sectored margins of leaf 8. Samples were digested with EcoRI and hybridized to probe prepared from umc5a. This blot was later reexposed to film for 14 days; no additional hybridizing bands were observed in lane 2.
DISCUSSION
The narrow sheath mutations map to ancestrally duplicated regions of the maize genome: The narrow sheath mutant phenotype is characterized by the deletion of a lateral domain of the leaf that includes the margin. Although the marginal domain of the leaf is deleted in mutant leaves, the central region of the mutant leaf is unaffected (Figure 1, A and B). Inheritance of the narrow sheath trait is determined by two unlinked, recessive mutations: ns1 and ns2 (Scanlonet al. 1996; Scanlon and Freeling 1998).
The ns1 and ns2 loci map to chromosomes 2L and 4L, respectively (Figure 2, Tables 1–3). Moreover, the ns genes map in the vicinity of white pollen1 (whp1) and colorless2 (c2), duplicated genes that encode chalcone synthase activity in the anthocyanin pigment pathway (Coeet al. 1981; Coe 1985; Wienandet al. 1986; Frankenet al. 1991). Analyses of genome evolution in maize have revealed that chromosome arms 2L and 4L contain homologous genes resulting from an ancient, allotetraploid event (Helentjariset al. 1988; Ahn and Tanksley 1993; Gaut and Doebley 1997). Previous studies demonstrated that plants homozygous for the ns2 mutation, and which contain at least one nonmutant allele of ns1, display completely nonmutant morphology in all leaf homologous organs (Scanlon and Freeling 1998). Our data indicate that the ns loci map to ancestrally duplicated regions of the maize genome and encode products of overlapping function in leaf development (discussed below).
The ns2-O mutation is linked to a paracentric inversion: The molecular mapping data (Table 3), genetic data (Table 4), and cytogenetic analyses (Table 5, Figure 3) presented herein provide corroborative evidence that the ns2-O mutation is tightly linked to a paracentric inversion encompassing ~13 cM on chromosome arm 4L. Likewise, these data indicate that neither the ns1-O mutation nor the ns2-Mu77 allele is linked to similar cytogenetic phenomena. In the absence of information concerning the molecular basis of the ns2-O mutation, we cannot speculate as to whether the inversion linked to ns2-O preceded, followed, or was causal to the origin of the ns2-O mutation.
Aneuploid sectors indicate that the narrow sheath genes encode overlapping functions during the development of a marginal leaf domain: Seed obtained from narrow sheath plants crossed to TB-4Lf were subjected to high energy radiation in an attempt to generate random loss of the nonmutant Ns1 gene in hypoploid plants already hemizygous for ns2. Eight mosaic plants were identified that contained wide, yellow-green sectors spanning the midrib region of wide leaves emanating from one axis of the plant; leaves from the opposite side of the plant were phenotypically narrow sheath and contained yellow-green sectors on the leaf edges. Combined genetic and RFLP (Figure 6) analyses reveal that the sectored tissue in mosaic plant 1 is hemizygous for both ns-O mutations (genotype ns1-O/−−; ns2-O/−−), whereas the nonsectored green tissue in these plants is heterozygous for ns1 and hemizygous for ns2-O (genotype ns1-O/Ns1; ns2-O/−−).
Intriguingly, in all cases where the sectored tissue encompassed several lateral veins near the midrib domain of a leaf, there was no effect on leaf width. However, when the hemizygous sectors passed over a broad domain that included the leaf margin, an ns phenotype was associated with the sectors (Figures 4 and 5). We conclude from these analyses that the activity of the NS1 gene product is domain specific and domain autonomous. NS1 gene function is not required for development of the middle leaf domain wherein founder cell recruitment initiates (Figure 7), but may be required to form a lateral domain of the leaf that gives rise to both leaf margins.
Previous analyses indicated that downregulation of KNOX accumulation, a marker of leaf founder cell recruitment (Smithet al. 1992; Jacksonet al. 1994), is not observed in the margin domain of ns mutant meristems (Scanlonet al. 1996). Furthermore, fate mapping studies showed that founder cells tracked to this same meristematic domain are not included during development of ns mutant leaves (Scanlon and Freeling 1997). The data presented here provide additional evidence that the ns duplicate loci perform nondivergent functions during development of the margin domain of the maize leaf. Alternatively, it is formally possible that the yellow-green phenotype alone, which is not caused directly by the ns mutations, generates a specific defect in the development of maize leaf margins without affecting the central leaf domain.
Models of mosaic meristem formation and NAR-ROW SHEATH function during founder cell recruitment. (A) On the left is a maize shoot apical meristem containing a clonal sector (yellow stripe) on one flank. The relative positions of three successive leaf founder cell populations (dark green) are shown. Note that the wider end of the founder cells corresponds to the incipient midrib domain of the leaf, whereas the narrow end on the opposite flank of the meristem corresponds to founder cells that will form both the left and right leaf margins. A maize shoot that is hemizygous for the ns2 locus, due to B-A translocation-induced hypoploidy, under-goes random X-ray-induced loss of chromatin encompassing the nonmutant ns1 locus in a cell on the flank of the shoot apical meristem. This generates a clonal sector of shoot tissue that is devoid of NS function (yellow stripe). The clonal sector on the meristem flank will intersect the founder cells of successive leaves in different domains, such that mature leaves emanating from the left side of the plant contain midrib sectors and leaves on the right side contain margin sectors. (B) Mature leaves containing alternating midrib-margin sectors of NS loss of function are predicted to display alternate wide leaf/narrow leaf phenotypes if NS function is restricted to the margin domain of the developing leaf. An alternative phenotype (C) is predicted in successive leaves if NS function is not specific to the margin domain, but is required in the central domain as well.
The whp1 and c2 gene pair are homologous loci that map near ns1 and ns2, respectively, and result from an ancestral duplication of the maize genome (Helentjariset al. 1988; Ahn and Tanksley 1993; Gaut and Doebley 1997). Although the WHP1 and C2 gene products perform identical biochemical functions, their expression patterns have diverged over ~11 million years of evolutionary time. Both the whp1 and c2 genes are expressed during pollen development, such that anthocyanin production in pollen displays duplicate factor inheritance. In aleurone tissue, however, homozygosity for the c2 mutation results in colorless kernels, despite the presence of a nonmutant whp1 allele(s) (Coe 1985; Frankenet al. 1991). Therefore, the ancestrally duplicated whp1 and c2 genes have evolved partially nonoverlapping functions in the development of the maize plant. Currently, 462 recessively inherited mutants are included in the maize genome database. Of these, only 18 mutants are reported to display duplicate factor inheritance of the full range of their particular mutant phenotypes. Consequently, of ~480 described loci, 18 gene pairs, including the ns genes, have retained functional redundancy since the ancient duplication of the maize genome (Gaut and Doebley 1997).
In addition, the aneuploid-mosaic plants constitute a dosage analysis of the ns-O mutations. Our genetic and molecular data indicate that the mosaic plants contained clonal sectors that were hemizygous for both ns1-O and ns2-O, due to loss of one dose of each chromosomal region harboring these duplicated loci. Intriguingly, the phenotype of the narrow, foliar leaves produced by these aneuploid plants is nearly identical (except for the yellow-green sectors) to the leaf phenotype observed in plants homozygous for both the ns1-O and ns2-O mutations (Figure 4). Therefore, a single dose of each ns-O mutation results in the same mutant phenotype as two doses of each ns-O mutation. These data indicate that the ns-O mutations are null alleles; the narrow sheath leaf phenotype represents the loss of NS function in maize.
Furthermore, because the sectors observed in mosaic plants 1–8 were wide enough to encompass several lateral veins of the leaf, we conclude that the sectors included a relatively large patch of meristematic cells (Poethig 1984). Also, the presence of yellow-green sectors on both margins of many narrow sheath mosaic leaves indicates that these sectors encompassed a leaf domain that is larger than that deleted by the narrow sheath mutations. The data suggest that sectors of double hemizygosity for both 2L and 4L generate yellow-green leaf tissue against a nonmutant background. Such yellow-green tissue is not observed in plants hemizygous for either chromosome 2L or 4L alone (M. J. Scanlon, unpublished observation), nor in ns homozygous plants. The genetic basis for the yellow-green leaf phenotype remains unknown.
The existence of half-leaf phenotypes, i.e., plants with normal margins on one leaf edge and narrow sheath mutant leaf morphology on the opposite edge (Figure 5, plants 5 and 8), indicates that the lateral domain implicated by NS gene function may actually be comprised of two separate leaf domains that must be recruited separately. That is, founder cell recruitment of leaf margins occurs on two fronts; the left leaf margin is formed independently of the right leaf margin. In this interpretation, sectors of NS loss of function that encompass the entire NS domain will yield fully mutant leaves devoid of leaf margins, whereas smaller NS mutant sectors encompassing just one side of the NS domain yield a half-leaf phenotype.
Developmental compartments in plant development: Abundant research conducted on the development of plant flowers has elucidated processes whereby homeotic gene products accumulate in whorl-specific patterns to effect the formation of specific floral organs (Coen and Meyerowitz 1991). Sakai et al. (1995) have proposed that the floral organ whorls are akin to plant developmental compartments and that the superman gene acts to maintain the compartmental-specific expression of floral homeotic genes. Furthermore, the product of the no apical meristem1 locus is purported to delineate the boundary between such floral compartments in petunia (Soeret al. 1996). The data presented here provide additional evidence that developmental compartments may exist within plant organs. We propose that the maize leaf is comprised of at least two distinct developmental domains (compartments; Figure 1), and the data presented here provide further evidence that the ns loci are not required for development of the central compartment(s) (Figure 7). As yet, no mutations have been identified that delete this central domain, although we expect that such a developmental lesion disrupting the initial recruitment of leaf founder cells may result in embryo lethality.
Acknowledgments
We thank M. Freeling for support and guidance during the early stages of this work, and B. Lane for X-ray treatment of corn seed.We thank L. Harper for critical discussions of the data and insight regarding inversion genetics. We thank G. Muehlbauer for assistance in maize cytogenetic techniques. We thank R. K. Dawe for critical reading of the manuscript and discussions of the data. We thank M.Polacco and P. Stinard for assistance with the Maize Genome Database.This research is supported by National Science Foundation grants BIR-9303608 and IBN-9808112 to M.J.S.
Footnotes
-
Communicating editor: J. A. Birchler
- Received February 29, 2000.
- Accepted March 17, 2000.
- Copyright © 2000 by the Genetics Society of America
