Intrinsic Protein Disorder in Complete Genomes

Difficulties of determining function and structure from sequence:

• Identification of distantly related sequences
• Motifs like the TIM barrel have evolved different functions
• Moonlighting, where one protein can have two or more completely unrelated functions

Functional disorder:

• DNA recognition
• Enzymatic activation through proteolytic digestion
• control of protein lifetimes
• transport of an unfolded chain through a small orifice
• structural uncoupling of two or more domains by flexible linkers

[edit] 1 Materials and Methods

• Disordered data: proteins with at least 40 consecutive disordered residues, 1149 residues in all; X-ray characterized proteins from PDB, NMR characterized from SWISS PROTEIN or PIR
• Ordered data: randomly selected from NRL_3D, PDB_SELECT_25
• Genomic databases: 34 complete or mostly complete from NCBI

• PONDR VL-XT applied to protein sequences longer than 60 residues in length in 34 genomes

[edit] 2 Results

• Wholly disordered proteins should not form crystals; any protein with a high sequence similarity to a PDB protein a candidate for prediction error, but sometimes fragments of proteins rather than whole proteins are crystallized, and many disordered proteins become ordered upon binding, and such proteins can end up in the PDB

• SWISS PROTEIN and PIR contain proteins with higher predictions for long regions of disorder compared to sequences in PDB, suggesting nature is rich in disorder and PDB is biased against it due to the crystallization requirement
• Characterization of disorder on a genome by genome basis better for estimating the commonness of intrinsic disorder due to the biases of sequence databases

Theory: Higher amount of disorder in eukaryotes might be due to a greater need for control and regulation compared to archaea and eubacteria; wide range of predicted disorder in archaea and eubacteria might relate to differences in signalling and controlling.

But: if that were true, the number of regulatory proteins and thus disordered proteins should increase as the total number of proteins increases, but there is no clear relationship between total numbers of proteins and the percent of putative disordered proteins

• Many wholly unfolded proteins and disordered regions fold into specific 3D structures upon formation of a complex with a partner
• Difficulty in recognizing a disordered region if its flavor has never been seen before
• Structural genomics will be incomplete if only ordered proteins are considered