How infidelity creates a sticky situation

Abstract

Errors in translation cause cytotoxic protein misfolding and aggregation. In this issue, Ling et al. (2012) show that scavenging or suppression of reactive oxygen species in E. coli reduces the cellular costs of error-induced aggregation.

Infidelity in the translation of genetic information into polypeptides alters the sequences of many proteins—on the order of one in five molecules under normal conditions (Drummond and Wilke, 2009). Because a protein’s sequence carries information about how to fold properly, translation errors tend to disrupt folding and expose hydrophobic residues that would normally be buried safely in the protein core (Fig. 1). Sticky misfolded molecules cling to cellular components and to each other, forming cytotoxic aggregates which are a hallmark of many human neurodegenerative diseases. A new study by Dieter Söll and colleagues in this issue (Ling et al., 2012) provides fascinating new evidence that demonstrates the interplay between translation errors, protein aggregation, and an ancient enemy of proteins: oxygen.

Figure 1.

Oxygen exacerbates mistranslation-induced protein aggregation. Mistranslation at the ribosome (enhanced by streptomycin) destabilizes proteins, disrupting folding and increasing their tendency to aggregate. Oxidation further damages proteins, and can crosslink aggregates, possibly increasing their toxicity. Overexpression of a component of the hydrogen peroxide scavenger AhpCF reduces aggregation and toxicity associated with mistranslation, though how that toxicity arises remains unclear.

To modulate translational fidelity, Söll and colleagues use streptomycin, an aminoglycoside antibiotic which binds to the ribosome and, at low concentrations, causes sub-lethal misreading of messenger RNA (Kramer and Farabaugh, 2007). The precise type and rate of errors induced by streptomycin remain unknown, an unfortunate blind spot; however, precise assays for errors at a single site show that streptomycin elevates misreading of certain codons in E. coli (Kramer and Farabaugh, 2007). Exposing E. coli cells to low levels of streptomycin induced formation of protein aggregates, which cells cleared within an hour, though not without suffering a substantial reduction in growth (Ling et al., 2012). Selection for genes whose overexpression partially suppressed the toxicity of streptomycin led to the identification of the alkyl hydroperoxide reductase subunit AhpF, among others. Critically, AhpF overexpression also suppressed streptomycin-induced aggregation in addition to reducing the toxicity of streptomycin (Fig. 1), whereas overexpression of a catalytically inactive AhpF mutant did not. But why would a hydrogen peroxide scavenger alter protein aggregation at all?

Aerobically growing cells generate reactive oxygen species (hydrogen peroxide, superoxide, and others) as they respire. These species react with cellular components, notably proteins and fatty acids, through multiple pathways, often indirectly. A well-studied example is metal-catalyzed protein oxidation to form carbonyls (Nystrom, 2005). Carbonyl modifications are irreversible and reasonably stable, and thus can be detected by antibodies or mass spectrometry. Decreasing translational fidelity proportionally increases protein oxidation measured by carbonyl formation (Dukan et al., 2000), apparently because error-destabilized proteins present more substrates for oxidative attack (Dukan et al., 2000) (Fig. 1). Whether this is because destabilized proteins simply expose more surface area, or because buried residues are more susceptible to oxidation, remains unclear.

Consistent with an effect from respiration, anaerobically growing cells required higher streptomycin concentrations to produce equal levels of aggregation (Ling et al., 2012), although this may reflect reduced uptake of streptomycin during anaerobic growth (Kogut et al., 1965). Experimental matching of error frequencies would allow comparison of these two conditions. Ling et al. (2012) then take these results several steps further. In a clever and useful contrast, they compare protein aggregates caused by streptomycin and by spectinomycin, which halts rather than scrambles translation. Streptomycin-induced aggregates are sharply enriched in carbonylated proteins relative to soluble proteins and to spectinomycin-induced aggregates (Ling et al., 2012). The question then becomes, why does oxidation worsen aggregation? Perhaps oxidation further destabilizes proteins already made wobbly by mistranslation (Fig. 1), increasing their residence time in an unfolded aggregation-prone state; like mistranslation, oxidation alters the protein’s primary sequence and its encoded folding information. Oxidation also causes protein crosslinking by multiple mechanisms, such as generation of species by lipid oxidation that can contain two separate protein-attacking nucleophiles. Like adding glue to a hairball, crosslinking makes aggregates more difficult to clean up (Fig. 1). Cellular proteases have trouble degrading crosslinked proteins (Grune et al., 1997), and chaperones, most of which operate by holding and pulling, cannot on their own resolve covalent adducts.

The reduced aggregation in the hyper-scavenging AhpF overexpression strain observed by Söll and colleagues (2012) could therefore result from reduced protein destabilization, reduced aggregate crosslinking, or both. These observations raise what remains a central question in studying the cellular consequences of protein misfolding: what causes cytotoxicity? Titrating a functionless aggregation-prone protein into the cytosol of otherwise healthy yeast cells linearly reduces their growth rate (Geiler-Samerotte et al., 2011), suggesting toxicity does not depend on loss of function. Aggregation-prone artificial peptides drag essential proteins out of solution in mammalian cells (Olzscha et al., 2011), suggesting a plausible mechanism for the apparent intrinsic toxicity of misfolded, aggregation-prone proteins (Bucciantini et al., 2002).

But misfolding and aggregation, while often correlated and causally related, represent distinct physical phenomena. Misfolding tends to expose sticky surface area; aggregation tends to bury it, at least as aggregates grow large. The longstanding correlation between aggregates and cytotoxicity, most evident in neurological disease, has fueled the notion that aggregates are the toxic species. An alternate view holds that exposed hydrophobic surface area is the source of bad behavior, that misfolded proteins and small aggregates are the toxic species, and that large aggregates are primarily cytoprotective (Bucciantini et al., 2002); whatever cells must spend to deal with aggregation may be outweighed by their gains from sequestering harmful species. Without resolving this debate, Ling et al.’s isolation of an antioxidant protein which is apparently better at reducing toxicity and aggregation than several prominent chaperones provides a new dimension on which these models can compete: does oxygen cause problems primarily by elevating misfolding, by make aggregates nastier, or by some other means (Fig. 1)?

Efforts to pin down causality will benefit from Söll and colleagues’ use of high-resolution mass spectrometry to thoroughly dissect aggregates, identifying their components and sites of oxidation. The aggregated proteins fall into several into several probable categories: proteins which are misfolded; proteins deployed to rescue or degrade misfolded proteins (chaperones and proteases); and proteins, often abundant species, which come along for the ride due to chance interactions in the cell or during sample processing. Determining which are which will require more targeted efforts, and this study’s large-scale census provides a crucial starting point.

Ultimately, we would like to know to what extent mistranslation contributes to protein misfolding and disease in humans, and how its effects might be suppressed. Do oxygen and mistranslation conspire to hammer neural proteins into neuron-destroying toxins? Might antioxidants confer their noted benefits in part through rendering mistranslation-induced misfolding less harmful, or more easily cleaned up? To answer these questions will require studies in eukaryotic cells, with techniques sensitive enough to detect mistranslation and its effects in the absence of fidelity-altering drugs.