Identification of a Regulatory Loop Controlling Xylem Production

The study of the factors that control patterning and differenti- ation of the vascular system in plants is of great basic and applied biological interest. As the vascular system plays a fundamental role in plant growth and physiology and is the major component of wood and plant-derived fibers, the ability to control vascular devel- opment has important implications for both agriculture and wood and fibers production. The characterization of Arabidopsis mutants with defects in vascular development, and reverse genetic analysis of vascular tissue-related genes revealed by transcriptional profil- ing has been deepening our knowledge of transcriptional regulation during wood formation (Demura and Fukuda, 2007). We have pre- viously shown that ATHB8, a member of the HD-ZIP III family of transcription factors, promotes the proliferation of vascular pre- cursor cells and their differentiation into xylem elements (Baima et al., 2001). ACAULIS5 (ACL5) is also involved in the regulation of the vascular system as acl5 produces an increased number of vascular elements in leaves and stems (Hanzawa et al., 1997, 2000; Clay et al., 2005). Very recently it has been proposed that ACL5 controls xylem specification by preventing premature cell death of tracheary elements (Knott et al., 2007; Muniz et al., 2008). Interestingly, we found that increased levels of ACL5 can delay or even inhibit the differentiation of procambial cells into differ- entiated tracheary elements. At very early stages of procambial cell differentiation, ATHB8 expression precedes that of ACL5 and the timing of ACL5 expression is further delayed by mutations of a cis-regulatory element in the ACL5 promoter recognized by ATHB8. Nonetheless, a careful analysis of the phenotype of multiple mutant combinations of acl5-1 with hd-zipIII mutants has shown that some of the developmental defects caused by the loss of ACL5 function are compensated for by the lack of HD-ZIPIII function. These observa- tions suggest that some HD-ZIPIII proteins act downstream of ACL5, although functional HD-ZIPIII genes are required to sustain the ACL5 autoregulatory negative feedback loop. These findings have been integrated in a model that will be presented.