366 C.L.H.Hord·H.Ma
redundantly in a way similar toEMS1/EXS. The expression patterns ofSERK1
andSERK2in the anthers largely coincide with the expression pattern of
EMS1/EXS. Fluorescence Resonance Energy Transfer (FRET) analysis showed
that SERK1 and SERK2 can form homo- or hetero-dimers in the plasma
membrane, suggesting that SERK1 and SERK2 act interchangeably in the
same complex (Albrecht et al. 2005).SERK1andSERK2transcripts were de-
tected in the L2-derived cells at stages 4 and early 5, including the PMCs,
strongly in the tapetum at late stage 5, and at reduced levels after stage 8 or 9.
The similarity ofems1andserk1 serk2anther phenotypes suggests that
these genes might mediate the same signaling pathway. EMS1/EXS is closely
related to the BRI1 brassinosteroid receptor, except that EMS1 lacks the hor-
mone binding domain (Zhao et al. 2002). In addition, the SERK1/SERK2 close
homolog BAK1 is known to interact with BRI1 (Albrecht et al. 2005; Colcom-
bet et al. 2005; Nam and Li 2002). Therefore, it is possible that EMS1/EXS
forms a receptor complex with SERK1 or SERK2 that binds to one or more
ligands (Albrecht et al. 2005; Colcombet et al. 2005). Currently, the most likely
candidate for the ligand is TAPETUM DETERMINANT1 (TPD1), a small 176
amino acid protein (Yang et al. 2003). Thetpd1mutant exhibits an identi-
cal phenotype (Yang et al. 2003) to those of theems1/exssingle mutant and
serk1 serk2double mutant, with no tapetum and extra PMCs. A double mu-
tant betweentpd1and anotherems1/exsallele has nearly identical anther
phenotypes to those of theems1andtpd1single mutants (Yang et al. 2003),
suggesting that these genes act in the same signaling pathway.
It was shown thatTPD1expression in the anther appears to be in all of
the L2-derived cells at stages 2–4 (Yang et al. 2003). At late stage 4 and early
stage 5,TPD1andEMS1/EXSare expressed in the PMCs and in tapetal precur-
sors, withTPD1being strongest in the PMCs andEMS1/EXSbeing strongest
in tapetal precursors, suggesting that they might mediate a signaling event
between these two cell types. The expression ofTPD1was not reduced in the
ems1/exs1-2mutant background, nor wasEMS1/EXSexpression reduced in
thetpd1mutant background, indicating that they do not regulate one another
at the mRNA level (Yang et al. 2005). Expression ofTPD1under the 35S pro-
moter in transgenic lines caused a range of abnormal phenotypes, including
abnormal tapetal development and degeneration and wide carpels (Yang et al.
2005). In addition, the wide carpel phenotype was dependent on the presence
of a wild type copy ofEMS1/EXS. Together these results indicate thatTPD1re-
quiresEMS1/EXSin order to regulate reproductive development (Yang et al.
2005).
In rice, theMULTIPLE SPOROCYTE1(MSP1)geneencodesanLRR-RLK
that shares63.8% amino acid sequence identity with EMS1/EXS (Nonomura
et al. 2003). Themsp1mutant is phenotypically similar toems1/exs,suggest-
ing thatEMS/EXSandMSP1are functional homologs (Nonomura et al. 2003).
In addition, themsp1mutant ovules also produce multiple sporogenous cells.
Thus, perhaps analogous toEMS1/EXSinArabidopsis,MSP1is important