Hide and Seek in the Peptidome

See allHide authors and affiliations

Science  05 Sep 2003:
Vol. 301, Issue 5638, pp. 1334-1335
DOI: 10.1126/science.1089553

By now we should be used to the wiliness of evolution,” begins a review on the many ways in which living organisms translate genetic information into polypeptides (1). Truer words were never written. Yet nature is still capable of surprises, much to the delight of those who study it. Two reports in this issue by Schwab et al. on page 1367 (2) and Yin et al. on page 1371 (3) illustrate the skill of the vertebrate immune system and its viral adversaries in detecting and concealing the peptide translation products of their genomes.

The adaptive cellular immune system is obsessed with peptides, which enable it to detect the presence of viruses and other intracellular pathogens. The job of detecting foreign peptides falls to CD8+ T cells. These cells express a clonally restricted receptor that recognizes 8- to 11-residue peptides nestling in the groove of major histocompatibility complex (MHC) class I molecules. Virtually all cell types in humans and other jawed vertebrates continually send class I molecules loaded with cellular peptides to the cell surface. In a process termed “tolerance,” T cells directed against class I molecules bearing self peptides are either killed or functionally silenced, preventing autoimmune responses. Following a viral infection, however, newly synthesized class I molecules carry viral peptides to the surface of infected cells. Here they are recognized by nonself-reactive T cells specific for the given peptide-class 1 complex. Activated T cells then deliver a cocktail of immune effector molecules that interfere with viral replication either by brute force (killing the virally infected cell) or by more subtle means (reprogramming the virally infected cell to disfavor viral replication). One of the key features of T cells is their remarkable sensitivity. The most efficient T cells approach the ultimate design limit of recognizing a single peptide-class I complex on the surface of a target cell.

Faced with such an adroit opponent, many viruses adopt a hit-and-run strategy, moving from host to host with such rapidity that they don't have to confront T cells, which need time to expand their numbers to cope with the burgeoning infected cell population. Epstein-Barr virus (EBV), however, like other herpesviruses, chooses a different strategy. This virus infects individuals for their entire lives, hiding in certain cell types and only infrequently becoming reactivated to produce new viral progeny that then infect other individuals. EBV evades T cells largely by suppressing the expression of its genes. To maintain latency, however, it must express EBNA1 (Epstein-Barr virus nuclear antigen 1). This protein has an amino-terminal domain composed of a glycine-alanine repeat region (GAr) that prevents its degradation by proteasomes, the macromolecular assemblies that dispose of damaged or unwanted cellular proteins. Most of the peptides presented by class I molecules to T cells are a by-product of this destruction. By blocking proteasome destruction of EBNA1, the GAr region may prevent presentation of EBNA1 peptides to T cells (4).

Yin et al. (3), however, show that EBNA1 must do more than block its destruction by the proteasome to avoid detection by the immune system's radar. As EBNA1 emerges from the ribosome during its synthesis, the GAr region interferes with translation of the protein in a manner yet to be elucidated. Why the virus would choose this method of reducing EBNA1 expression remains unclear. One possibility is that modifying translation in this manner increases the effectiveness of the GAr region in preventing peptide production, perhaps by decreasing the number of defective ribosomal products (DRiPs) generated.

DRiPs are the consequence of unavoidable imperfections in the process of translating genetic information into functional proteins (see the figure). They may account for the bulk of the surprisingly large fraction (about one-third) of newly synthesized proteins that are degraded by proteasomes within 30 minutes of their synthesis (5, 6). Evidence is accumulating that DRiPs are the major source of peptides associated with class I molecules. A critical question is the composition of the rapidly degraded cohort of newly synthesized proteins. It is still not clear how much of this cohort represents truly short-lived proteins (SLiPs) and how much represents various forms of DRiPs, which include properly translated but misfolded proteins and misbegotten polypeptides resulting from mistakes in the fidelity of transcription or translation.

Generating MHC class I peptides.

Ribosomes convert genetic information conveyed by messenger RNA into proteins. Newly synthesized proteins are divided into two cohorts with distinct life-spans. About 70% of proteins are stable with an average half-life of 3000 min prior to degradation by proteasomes, whereas the rest have an average half-life of 10 min prior to proteasomal degradation. It is uncertain what fraction of the rapidly degraded pool are proteins intended to be short-lived (SLiPs) versus defective proteins (DRiPs). In all cases, proteins are degraded to free amino acids that are recycled into new proteins. A very small fraction of proteasome-generated peptides (perhaps 1 in 106) are presented by MHC class I molecules to T cells, frequently after further trimming by aminopeptidases. Most antigenic peptides are derived from DRiPs. The relative contributions of errors in folding, translation, and transcription to the DRiP pool are unclear. Yin et al. show that EBNA1 of EBV has an amino-terminal sequence that disfavors proteasome degradation and also reduces translation of its own message (3). Together, these features reduce the generation of EBNA1 peptides, enabling cells harboring EBV to escape immune surveillance. Schwab et al. show that ribosomes initiate translation in unexpected places, generating unintended translation products that contribute to DRiPs (2).


In relatively artificial viral and tumor systems, antigenic peptides can arise from “noncoding” regions of host and viral genomes (7). Schwab et al. (2) bring these findings a step closer to real life by generating transgenic mice that encode a defined antigenic peptide in an “untranslatable” region of an inserted gene. T cells specific for this peptide kill a variety of distinct cell types from the transgenic mice in vitro. Such T cells have to be raised in nontransgenic animals, because the transgenic animals become tolerant to the peptide, clearly demonstrating its biological significance. Amazingly, synthesis of the peptide is initiated not by a standard AUG codon specifying the amino acid methionine or even by misreading of a non-AUG codon as methionine. Rather, translation is initiated in an unprecedented manner by a CUG codon that specifies the amino acid leucine.

These findings, together with earlier clues (8), have important implications for both cellular immunity and for the general process of decoding genetic information. The “peptidome,” that is, the set of all peptides from host and pathogenic genomes presented by class I molecules to the immune system, may be much larger than previously thought. It now seems that peptides can potentially be generated from all DNA sequences in each of the six potential reading frames (three frames on each strand of DNA), and not just from standard genes in standard reading frames. Moreover, because of ribosomal skipping (1), peptides might be derived from noncontiguous sequences in the same or different reading frames. Immunologists, particularly those searching for elusive peptides recognized by tumor-specific T cells, will be disheartened by the new results. Their job of finding a peptide needle in the growing genetic haystack has just become a lot more difficult. Biologists now have to consider the possibility that cells generate polypeptides of potential evolutionary importance from a much larger information store than previously considered.

Typically, these findings raise many more questions than they answer. Is there something peculiar about the experimental strategy used by Schwab et al. that predisposes genes to aberrant expression, or do real genes behave this way? How much translation is devoted to these nonstandard genes? How many of the translation products have a real function in cells? Are these peptides translated for the purpose of tumor immune surveillance (9), or, as seems more likely, is the immune system taking advantage of a preexisting process? These are tough questions, and answering them will require all of the wiliness that evolution has provided our species.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.

Navigate This Article