In setting out to locate the “lock” into which brassinosteroids fit, the researchers looked first at a leucine-rich receptor kinase called BRI1.  Previous research indicated that this particular substance was a central part of a plasma membrane receptor for brassinolide, but whether it directly bonds with brassinolide was unknown before this experiment.  This paper presented the first evidence for direct binding of active brassinosteroids to BRI1.  The researchers were able to isolate a group of 94 amino acids in the extracellular domain of BRI1 that directly bind to brassinosteroids. 

The first part of the research consisted of the analysis of proteins immunoprecipitated from plants with high levels of a protein linked to BRI1, BRI1-GFP, or Green Flourescent Protein.  There was a high level of binding activity with brassinolide in these immunoprecipitates, strongly suggesting that some protein in these immunoprecipitates was binding with brassinosteroids. 
To identify this protein, the researchers prepared a probe, similar to brassinolide but with a biotin tag attached to it.  This biotin tag can be detected using specific antibodies.  Thus, the researchers were able to simulate brassinolide bonding using this probe, and use the antibodies to detect and analyze the binding.  The results of this analysis were conclusive: the probe bound exclusively to BRI1.  A second, similar study was conducting using microsomal membranes, with similar results.  Thus, the researchers had determined that BRI1 does, in fact, bind to brassinosteroids.  The “doorknob” had been found.  However, the researchers still were not satisfied, and thus conducted more research to determine which part of this doorknob constituted the lock. 

The results of previous studies suggested that the extracellular domain of BRI1, particularly the 70-amino-acid island domain (or ID) located between the 21st and 22nd leucine-rich repeats (LRR21 and LRR22), might play a crucial role in brassinosteroid binding.  The researchers used five BRI1 fragments to perform photoaffinity crosslinking and search for binding signals.  These binding signals were found in LRR21-ID-LRR22 and in ID-LRR22, but not in either LRR21-ID or LRR22, suggesting that both the island domain and the 22nd leucine-rich repeat are required for brassinosteroid bonding. 

The researchers then set out to ascertain whether the probe binding was specific for brassinosteroid receptors.  They tested whether the probe would bind to ID-LRR fragments from three proteins similar to BRI1: BRL1, BRL2, and BRL3.  They found that the probe would bind to ID-LRRs from BRL1 and BRL3, but not to BRL2.  BRL2 also fails to bind brassinolide with high affinity as BRL1 and BRL3 do.  The results of this phase of the research suggest that brassinosteroid binding is specific and that small differences from BRL2 in BRL1, BRI1, and BRL3 are the deciding factor in brassinosteroid binding ability. 
To determine the minimum region required for brassinosteroid binding, the researchers deleted amino acids from the N terminus and the C terminus of ID-LRR22.  Almost immediately, these modified binding sites ceased to function effectively, demonstrating that the whole ID-LRR22 region is necessary for brassinosteroid binding.  

The results of the overall research indicate that the minimal binding domain for brassinosteroids is a 70-amino-acid island domain located between the 21st and 22nd leucine-rich repeats of BRI1, coupled with the carboxy-terminal flanking the 22nd leucine-rich repeat, that this binding domain is very specific, and that the entire domain is necessary for brassinosteroid binding.