For this subunit (Figures 2.1d and 2.2b), which is located at the longest distance from the center of the PSI trimer within the stromal ridge, a solution structure was already known [ 72] before its structure could be modeled into an electron density map of PSI crystals. Surprisingly, the structure revealed similarity to Src homology 3 (SH3) domains found in eukaryotic proteins of signal transduction pathways . Only after the resolution and quality of electron density maps determined by crystallographic methods had been improved over many years, could a first model of the polypeptide backbone of PsaE be fitted into an electron density map at 4 A resolution  using the NMR structure ofPsaE from Synechococcus sp. PCC 7002 . This initial structural model was confirmed at higher resolution  and shows good agreement with the solution structure for those parts which fold into a five-strand anti-parallel P-sheet, forming a P-barrel . The largest differences are found for the long CD-loop (connecting strands PC and PD, nomenclature as in Falzone and coworkers 1994 ) which adopts a twisted conformation and is described to be flexible in solution. Obviously, a change in its conformation upon binding to PSI enables the CD-loop to form hydrogen bonds with PsaA, PsaB and PsaC as found in the crystal structure [27, 51]. The CD-loop is the most prominent part of PsaE for the interaction with the PSI core. Compared to the extensive binding networks found between PsaC and PsaD, PsaE is only weakly integrated into the stromal ridge of PSI. This is consistent with the suggested scenario of a sequential assembly of the stromal subunits, where PsaE is the protein which binds last to PSI . Like PsaD, which interacts with the small membrane-integral subunit PsaL, PsaE forms hydrogen bonds via its AB-loop with the membrane [integral subunit PsaF ; An interesting detail of the stromal ridge structure is the interaction between the BC 4oop of PsaE and the inserted loop in PsaC, which contains residue K34 essential for ferredoxin docking. The pairwise stabilization of these two loop conformations could be critical for binding ferredoxin to PSI, especially as the loop in PsaE also contains a residue (R39 in T.elongatus[ which was shown to play an important role for the affinity of PSI to ferredoxin  ; Besides its responsibility for the formation of a stable PSI-fer-redoxin complex, PsaE is not important in electrostatic guiding of ferredoxin to its binding site, in contrast to PsaD, and most probably does not affect the geometry of ferredoxin binding ; for reviews see Setif, 2001  and Setif and coworkers 2002  [ The idea that in cyanobacteria PsaE is involved in NADP+ reduction via interaction with FNR in a ternary PSI-ferredoxin-FNR complex, based on kinetic data , was not confirmed . In contrast, the interaction of PsaE with FNR has been shown in plants . Studies on cyanobacterial mutants lacking PsaE point to another role of this subunit, which is obviously required for cyclic electron transfer around PSI [25, 79].
Was this article helpful?