Low complexity regions in proteins

LCRs were originally thought to be unstructured and flexible linkers that served to separate the structured (and functional) domains of complex proteins, but they are also capable of forming secondary structures, like helices (more often) and even sheets. They may play a structural role in proteins such as collagens, myosin, keratins, silk, cell wall proteins. Tandem repeats of short oligopeptides that are rich in glycine, proline, serine or threonine are capable of forming flexible structures that bind ligands under certain pH and temperature conditions. Proline is a well-known alpha-helix breaker, however, amino acid repeats composed of proline may form poly-proline helices. == Functions ==

LCRs were originally thought as 'junk' regions or as neutral linkers between domains; however, experimental and computational evidence increasingly indicates that they may play important adaptive and conserved roles, relevant to biotechnology, heterologous protein expression, medicine, as well as to our understanding of protein evolution. LCRs of eukaryotic proteins have been involved in human diseases, especially neurodegenerative ones, where they tend to form amyloids in humans and other eukaryotes. They have been reported to have adhesive roles, function in excreted sticky proteins used for prey capture, or have roles as transducers of molecular movement, e.g. in the prokaryotic TonB/TolA systems. LCRs may form surfaces for interaction with phospholipid bilayers, or as positive charge clusters for DNA binding, or as negative or even histidine-acidic charge clusters for coordinating calcium, magnesium or zinc ions. They may even function as frame-shift checkpoints, by shifting to an unusual amino acid content that makes the protein highly unstable or insoluble, which in turn triggers fast recycling, before any further cellular damage. Analyses on model and non-model eukaryotic proteomes have revealed that LCRs are frequently found in proteins involved in binding of nucleic acids (DNA or RNA), in transcription, receptor activity, development, reproduction and immunity whereas metabolic proteins are depleted of LCRs. A bioinformatics study of the UniProt annotation of LCR containing proteins observed that 44% (9751/22259) of Bacterial and 44% (662/1521) of Archaeal LCRs are detected in proteins of unknown function, however, a significant number of proteins of known function (from many different species), especially those involved in translation and the ribosome, nucleic acid binding, metal-ion binding, and protein folding were also found to contain LCRs. == Properties ==

LCRs are more abundant in eukaryotes, but they also have a significant presence in many prokaryotes. Similarly, extremely hydrophobic regions can form non-specific protein–protein interactions among themselves and with other moderately hydrophobic regions in mammalian cells. Thus, their presence may disturb the balance of protein-protein interaction networks within the cell, especially if the carrier proteins are highly expressed. A third explanation may be based on micro-evolutionary forces and, more specifically, on the bias of DNA polymerase slippage for certain di- tri- or tetra-nucleotides . Amino acid enrichment for certain functional categories of LCRs A bioinformatics analysis of prokaryotic LCRs identified 5 types of amino acid enrichment, for certain functional categories of LCRs: • Proteins with GO terms related to polysaccharide binding and processing were enriched for serine and threonine in their LCRs. • Proteins with GO terms related to RNA binding and processing were enriched for arginine in their LCRs. • Proteins with GO terms related to DNA binding and processing were especially enriched for lysine, but also for glycine, tyrosine, phenylalanine and glutamine in their LCRs. • Proteins with GO terms related to metal binding and more specifically to cobalt or nickel-binding were enriched mostly for histidine but also for aspartate in their LCRs. • Proteins with GO terms related to protein folding were enriched for glycine, methionine and phenylalanine in their LCRs. Based on the above observations and analyses, a Neural Network webserver named LCR-hound has been developed to predict LCRs and their function. == Evolution ==

LCRs are very interesting from a micro and macro evolutionary perspective. Thus, they are linked to recombination hotspots and may even possibly facilitate cross-over. By originating from genetic instability, they may cause, at the DNA level, a certain region of the protein to expand or contract and even cause frame-shifts (phase-variants) that affect microbial pathogenicity or provide raw material for evolution. Most intriguingly, they may provide a window into the very early evolution of life. During early evolution, when only few amino acids were available and the primary genetic code was still expanding its repertoire, the first proteins were assumed to be short, repetitive and therefore, of low complexity. Thus, modern LCRs could represent primordial aspects of the evolution towards the protein world and may provide clues about the functions of the early proto-peptides. Thus, any prokaryotic LCRs that constitute evolutionary accidents with no functional significance should not be fixed by genetic drift and consequently should not demonstrate any levels of conservation among moderately distant evolutionary relatives. On the contrary, any LCR found among homologs of several moderately distant prokaryotic species should very probably reserve a functional role. == LCRs and the protopeptides of the early genetic code ==

The amino acids with the highest frequency in LCRs are glycine and alanine, with their respective codons GGC and GCC being the most frequent, as well as complementary. Intriguingly, it has also been suggested that they represent the very first two amino acids and codons of the early genetic code. Thus, these two codons and their respective amino acids must have been constituents of the earliest oligopeptides, with a length of 10–55 amino acids and very low complexity. Based on several different criteria and sources of data, Higgs and Pudritz suggest G, A, D, E, V, S, P, I, L, T as the early amino acids of the genetic code. Trifonov's work largely agrees with this categorization and proposes that the early amino acids in chronological order are G, A, D, V, S, P, E, L, T, R. An evolutionary analysis observed that many of the amino acids of the suggested very early genetic code (with the exception of the hydrophobic ones) are significantly enriched in bacterial LCRs. Most of the later additions to the genetic code are significantly under-represented in bacterial LCRs. They thus hypothesize and propose that, in a cell-free environment, the early genetic code may have also produced low complexity oligo-peptides from valine and leucine. However, later on, within a more complex cellular environment, these highly hydrophobic LCRs became inappropriate or even toxic from a protein interaction perspective and have been selected against ever since. In addition, they further hypothesize that the very early protopeptides did not have a nucleic acid binding role, because DNA and RNA-binding LCRs are highly enriched in glycine, arginine and lysine, however, arginine and lysine are not among the amino acids of the proposed early genetic code. == Detection methods ==

Low complexity regions in proteins can be computationally detected from sequence using various methods and definitions, as reviewed in. == References ==