Teosinte branched1 and the origin of maize
--John Doebley and Adrian Stec

 Teosinte branched1 (tb1) is a recessive mutant of maize that affects plant architecture and maps to chromosome arm 1L. Mutant plants (tb1-ref) have long lateral branches tipped by tassels at some upper nodes of the main culm and tillers at the basal nodes. We will refer to this syndrome as teosinte branched plant architecture. This contrasts with normal maize plant architecture conferred by the dominant maize allele (Tb1): short lateral branches tipped by ears at some upper nodes and few or no tillers at the basal nodes. Since both tillers (basal lateral branches) and upper lateral branches arise from axillary meristems, in a general sense, tb1 controls the fate of the axillary meristems, although with different effects depending on the position (basal or upper) of the axillary meristem within the plant.

The architecture of tb1-ref maize plants resembles that of the nearest wild relatives of maize, the annual teosintes. Like tb1-ref plants, annual teosinte grown in its native habitat in Latin America "normally" produces long lateral branches tipped by tassels at upper nodes of the main culm. We say "normally" because annual teosinte plants in Latin America may produce short lateral branches tipped by normal teosinte ears or mixed male-female inflorescences in some environments such as dry shallow soils or low light (shading). Also like tb1-ref plants, annual teosinte can tiller profusely. Tillering in teosinte is extreme when the plants are grown at temperate latitudes, apparently in response to the long days of these regions which may prolong the juvenile phase of development during which tillers are formed. Tillering is uncommon for teosinte in Latin America, perhaps because, under the short Latin American days, the plants begin adult development earlier, restricting the opportunity to produce tillers. Thus, while teosinte resembles tb1-ref maize, the extent and nature of the resemblance is dependent on the environment in which the teosinte plants are grown. This situation suggests that the developmental pathway controlling plant architecture in teosinte is responsive to environmental signals, most likely in a way that best adapts the plant to the local environment.

Evidence from our QTL (quantitative trait locus) mapping studies (Doebley and Stec, Genetics 134:559-570, 1993) demonstrated that a QTL on chromosome arm 1L largely controls the differences in plant architecture between maize and teosinte. This QTL is very near (within 10 map units) to the location of tb1. For this reason, we proposed that tb1 is our QTL and that tb1 was largely responsible for the evolution of normal maize plant architecture from the ancestral teosinte plant architecture. To test this hypothesis, we performed a simple complementation test. First, we transferred the region of teosinte chromosome arm 1L encompassing tb1 and our QTL into maize inbred W22 by four generations of backcrossing using molecular markers to retain the teosinte segment of 1L. No phenotypic selection was exercised. A fourth generation backcross plant heterozygous for the teosinte chromosome segment was used to pollinate a maize plant carrying Tb1/tb1-ref (seed obtained from Charles Burnham). We considered two possible outcomes. (1) Our QTL is not allelic to tb1 in which case all plants should have normal maize plant architecture. (2) Our QTL is allelic to tb1 in which case one-fourth of the plants should have teosinte branched plant architecture. These expectations arise because among the F1's there are four genotypic classes, only one of which should give teosinte branched plant architecture:
Genotypes Expected Phenotypes
Tb1 / Tb1+W22 normal maize
Tb1 / tb1-teosinte normal maize
tb1-ref / Tb1+W22 normal maize
tb1-ref / tb1-teosinte teosinte branched

The expectation that one-fourth of the F1 progeny should have the teosinte branched phenotype assumes that tb1-teosinte will behave as a recessive. This assumption is based on our observation that the BC1, BC2 and other backcross generations all exhibited normal maize plant architecture despite the fact that they were heterozygous (Tb1+W22/tb1-teosinte).

Seventy-two F1's from the cross were classified with 57 having normal maize and 15 having a weak teosinte branched plant architecture. Plants with weak teosinte branched plant architecture had elongated lateral branches tipped by mixed male-female inflorescences (Fig. 1). The ratio of 57:15 does not differ significantly (X2 = 0.67, 1 d.f., p > 0.25) from the 3:1 ratio expected if our QTL is an allele of tb1. Although none of the progeny had a strong teosinte branched phenotype, since one-fourth did show a weak teosinte branched phenotype, we conclude that our QTL is allelic to tb1 but that our QTL represents a weak allele (tb1-teosinte) relative to the maize mutant (tb1-ref).

A model for teosinte branched1. We propose the following model for the function of tb1 and how it changed during the evolution of maize. In teosinte, tb1 encodes a repressor of the elongation of the lateral branches or of the development of axillary meristems. In good environmental conditions and full sunlight, this locus is not expressed in teosinte and thus the axillary meristems develop into basal tillers or upper lateral branches tipped by tassels. In poor environmental conditions (low moisture, shallow soil) and/or shading, tb1 is expressed in teosinte and it directly or indirectly represses the development of the axillary meristems such that few (or no) basal tillers and only short upper lateral branches tipped by ears are produced. Thus, tb1 is a locus whose original function was in adapting the teosinte plant to its local environmental situation by altering plant architecture. To explain the evolution of maize plant architecture, we propose that the expression of tb1 is no longer tied to an environmental signal but rather that tb1 in maize is constitutively expressed during the early development of the axillary meristems, keeping both tillering and full elongation of the upper lateral branches repressed. Under this model, both the tb1-teosinte and Tb1+maize alleles would encode functional products, although ones that are differently regulated. tb1-teosinte is recessive to Tb1+maize because the latter will produce the repressor whether or not the former allele is activated by an environmental signal. Finally, under this model, the maize mutant (tb1-ref) can be explained as a recessive loss of function allele. With complete loss of the repressor function, the axillary meristems of homozygous tb1-ref plants elongate to produce either basal tillers or elongate upper lateral branches tipped by tassels.

Figure 1. Inflorescences terminating the primary lateral branch from the complementation test discussed in the text: a female inflorescence (ear) showing the normal maize phenotype of the wild type maize Tb1 allele (left); and a mixed male-female inflorescence showing the phenotype of a tb1-ref / tb1-teosinte plant (right).

Please Note: Notes submitted to the Maize Genetics Cooperation Newsletter may be cited only with consent of the authors

Return to the MNL 68 On-Line Index
Return to the Maize Newsletter Index
Return to the Maize Genome Database Page