19 DECEMBER 2014 • VOL 346 ISSUE 6216 1455 SCIENCE sciencemag.org
The latter may benefit from adjuvant-like
signals generated by viral elements (14), including endogenous retroviruses. Even under steady-state conditions, the continuous
stimulation of mucosal B cells by commensal antigens might increase the expression
of endogenous retroviruses.
Similar to bacterial polysaccharides, carbohydrates linked to large viral glycoproteins
may generate TI-2–like B cell responses (15).
Given that certain viral infections stimulate
the expression of endogenous retroviruses,
adjuvants containing endogenous retroviral
sequences may enhance both the magnitude and breadth of neutralizing antibody
responses to viruses, including HIV. Thus,
a full understanding of our virome may
catalyze more efficient strategies to enhance
protective humoral responses. However, targeting these viruses with vaccines should be
approached with caution, as certain endogenous retroviruses could precipitate or accelerate the development of autoimmunity and
cancer as well as the spreading of a preexisting HIV infection. In this context, defining the “virotype” of each individual may be
preferable to defining how individual viral
strains regulate the immune response. Individual virotypes would define a distinct set
of endogenous retroviruses that share target
cells, sensors, and downstream transcription
factors in an individual (2).
The study by Zeng et al. is a reminder that
the virome is an integral part of our genetic
identity. Thus, deciphering human virotypes
may help elucidate the interindividual differences than cannot be explained solely by genetic mutations. Together with the bacterial
microbiome, the retrovirome likely shapes
both the magnitude and quality of the immune response. Future advances in our understanding of host-virome relationships are
bound to generate new insights into the role
of these peaceful genome inhabitants in infection, vaccination, inflammatory disease,
autoimmunity, and cancer. ■
1. E. C. Holmes, Cell Host Microbe 10, 368 (2011).
2. H. W.Virgin, Cell 157,142(2014).
3. K. Honda, D. R. Littman, Annu. Rev. Immunol. 30, 759
4. S. Fagarasan et a l., Annu. Rev. Immunol. 28, 243 (2010).
5. J.Z.Oh et al., Immunity 41,478(2014).
6. M. Zeng et al ., Science 346, 1486 (2014).
7. G. D. Victora, M. C. Nussenzweig, Annu. Rev. Immunol. 30,
8. M. McHeyzer-Williams et al ., Nat. Rev. Immunol. 12, 24
9. A. Cerutti et al., Nat. Rev. Immunol. 13, 118 (2013).
10. J. J. Mond et al ., Annu. Rev. Immunol.13, 655 (1995).
11. R. Guinamard etal., Nat.Immunol.1, 31 (2000).
12. G. Magri et al ., Nat. Immunol.15, 354 (2014).
13. F. Martin etal.,Immunity14, 617 (2001).
14. B. Romeu etal .,J.Drug Target. 20, 502 (2012).
15. E.Szomolanyi-Tsuda,R.M. Welsh, Curr.Opin.Immunol. 10,
Life today depends on the chemical activity of proteins. Scientists have attempted for decades to understand their intricate structures and diverse chemical activities and to emulate their properties by design. However,
designing structured polypeptide chains—
a prerequisite for creating functional proteins—has proven extremely challenging.
Recent years have seen progress in the design of folded proteins in aqueous solution
(1) and of proteins that are catalytically active (2), but the design of membrane proteins remains in its infancy ( 3). A landmark
study by Joh et al. on page 1520 of this issue
(4) meets two major challenges in the quest
to engineer new proteins: the design of a
folded membrane protein that performs a
By the time of the last universal common
ancestor of all life on Earth, some 3.5 billion years ago, a tripartite division of labor
had emerged among life’s macromolecules,
with DNA assuming the role of information
repository, proteins providing catalytic activity, and RNA mediating between them.
All three require defined three-dimensional
structures to fulfill their biological roles.
But whereas nucleic acids fold spontane-
ously and recover their structure robustly
after denaturation, protein folding is a com-
plicated process that is easily derailed; after
denaturation, proteins typically aggregate
and have to be degraded and resynthesized.
Rapid advances in engineering nucleic
acids have made genetic engineering a
routine technology, with a broad range of
applications and predictable outcomes.
Engineering proteins, on the other hand,
turned out to be an altogether more difficult proposition due to what has become
known as the protein folding problem: How
does an amino acid sequence determine a
Two aspects in particular have made
this problem intractable. First, most amino
acid chains do not have a folded structure.
This may seem counterintuitive, because
the chains we typically encounter are those
of natural proteins, and most of these are
folded. However, screens of polypeptide libraries have shown that fewer than one in
a billion exemplars is folded. Second, for
the few chains that fold, the free energy of
folding is equivalent to just a few hydrogen
bonds. Most folded proteins are thus energetically quite close to the unfolded state—a
fact illustrated by the disruption that heat
shocks of just a few degrees above normal
growth temperature can cause. Because
What I cannot create,
I do not understand
A designed ion channel. The ability to reproduce a biological activity in a designed membrane protein, reported
by Joh et al., is an essential step in establishing that the underlying principle is understood, as stated succinctly in
Richard Feynman’s well-known dictum of the title.
By Andrei N. Lupas
A designed protein transports ions across a membrane
Department of Protein Evolution, Max Planck Institute for
Developmental Biology, 72076 Tübingen, Germany. E-mail: