The surprising complexity of signal sequences

Most secreted and many membrane proteins contain cleavable N-terminal signal sequences that mediate their targeting to and translocation across the endoplasmic reticulum or bacterial cytoplasmic membrane. Recent studies have identified many exceptions to the widely held view that signal sequences are simple, degenerate and interchangeable. Growing evidence indicates that signal sequences contain information that specifies the choice of targeting pathway, the efficiency of translocation, the timing of cleavage and even postcleavage functions. As a consequence, signal sequences can have important roles in modulating protein biogenesis. Based on a synthesis of studies in numerous experimental systems, we propose that substrate-specific sequence elements embedded in a conserved domain structure impart unique and physiologically important functionality to signal sequences.

Introduction

The discovery that secreted proteins are encoded with short, removable N-terminal ‘signal sequences’ that earmark them for export [1] was one of the great breakthroughs in cell biology in the 1970s. Much work in the subsequent 30 years has shown that signal sequences are recognized by dedicated factors that catalyze the transport of secretory precursors (preproteins) across the endoplasmic reticulum (ER) or bacterial cytoplasmic membrane. The key molecules in this pathway are found in all living cells and include a cytosolic targeting factor called the signal recognition particle (SRP), the membrane-bound receptor for SRP, an integral membrane protein conducting channel called the Sec61p complex (in eukaryotes) or the SecY complex (in prokaryotes), and a membrane-bound peptidase that removes signal sequences from preproteins on the lumenal (or periplasmic) face of the membrane.

The signal hypothesis originally predicted that all signal sequences would share a distinctive sequence motif; however, the absence of any ‘consensus’ quickly became apparent when several preproteins were sequenced. Nevertheless, the concept of ‘signal sequence equivalence’ prevailed. An early comparative sequence analysis [2] showed that signal sequences have a typical size (∼20–30 residues) and a recognizable three-domain structure (a basic ‘N domain’, a ∼7–13 residue hydrophobic ‘H domain’ and a slightly polar ‘C domain’), but otherwise lack any significant homology. By the early 1980s, it had become clear that signal sequences are often readily interchangeable, tolerant of a wide range of mutations [3] and even capable of directing secretion in evolutionarily distant organisms 4, 5. A remarkable study published in 1987 showed that ∼20% of random sequences can promote the secretion of invertase in yeast [6], strengthening the conclusion that the only essential feature of signal sequences is a hydrophobic core that is uninterrupted by charged residues. These studies implied that the signal recognition machinery has a high degree of tolerance, but they also created the impression that signal sequences have only a very circumscribed role in protein biogenesis. Thus, it has become common ‘wisdom’ that signal sequences are simple, interchangeable domains with a low information content.

Although the discovery of species-specific features of signal sequences [7] and a distinct motif in bacterial lipoprotein signal sequences [8] suggested >20 years ago that not all signal sequences are equivalent, hints that signal sequence diversity reflects an underlying functional complexity have emerged only in more recent studies. These studies, often from seemingly unrelated fields and disparate experimental systems, share one important feature: they have each examined the biosynthesis of a preprotein or membrane protein that differs from the ‘model’ substrates that were originally used to define the basic principles of signal sequence function. Individually each study might seem to describe an ‘exception’ to the well-established paradigm, but viewed together they suggest that differences among signal sequences could ultimately prove to be as physiologically important as their similarities.

Recent experiments have demonstrated that sequence variation among signal sequences can affect protein targeting and translocation, signal sequence cleavage and even postcleavage events. Perhaps more importantly, these studies have shown that the modulation of specific steps in the recognition and processing of signal sequences can have an essential role in protein biogenesis. Here, we review evidence showing that signal sequences encode far more information than is commonly thought, and we synthesize these studies into a cohesive framework to explain signal sequence diversity.