DAV I D J . S T E PHENS

Together with other extracellular proteins, collagen provides the structural frame-work on which tissues develop and

function. It is synthesized in the endoplas-mic reticulum, an intracellular organelle, as a rigid, rod-like precursor (procollagen) about 300 nanometres in length. Pro collagen — like nearly all secreted proteins — is then pack-aged into transport vesicles for delivery to another organelle, the Golgi apparatus, before its secretion to the cell’s surroundings. Trans-port vesicles, however, are typically smaller than 100 nm, as they are generated from the endoplasmic reticulum by a group of proteins (the COPII coat) that co-assemble as a struc-turally defined polyhedral cage1. On page 495 of this issue, Jin et al.2 reveal that modification of one of the COPII proteins allows the forma-tion of vesicles that are large enough to hold procollagen.

The outer layer of the COPII coat is assem-bled using structural elements comprised of the proteins SEC13 and SEC31 (Fig. 1a). Although it was thought that the hinges between these elements are flexible enough to allow vesicles of various sizes to form3,4, little was known about how vesicle size is controlled. Jin and colleagues2 show that SEC31 can be modified by ubiquitination — the attach-ment of one or more copies of a small protein called ubiquitin. Although ubiquitination can ‘mark’ a protein for degradation, it is becoming increasingly clear that it can also affect protein function5.

Specifically, the authors2 report that, in mouse cells, the enzyme CUL3–KLHL12 adds a single ubiquitin to a small pool of SEC31 molecules, and that this modifica-tion is required to drive the secretion of collagen. Using high-resolution electron micro scopy, they found that overexpression of CUL3–KLHL12 leads to the production of large COPII structures, up to 500 nm in diameter — sufficient to accommodate pro-collagen. The simplest explanation for these observations is that ubiquitin attachment to SEC31 results in a structural change in the COPII cage that alters coat flexibility, and allows procollagen to be encapsulated in a nascent vesicle (Fig. 1b).

Jin and colleagues’ observation that only some SEC31 molecules are modified indicates strongly that the addition of ubiquitin does not directly modulate the mechanics of COPII coat assembly. Instead, SEC31 ubiquitination might lead to recruitment of an additional, unknown protein to perform this role — for example, by further stabilizing lateral SEC13–SEC31 inter-actions. Identification of the additional factor and a more detailed molecular explanation of the modified geometry of the vesicle coat are challenges for the future.

Ubiquitination of some SEC31 molecules

could be an ongoing process that facilitates the formation of large COPII vesicles as a routine cell function; alternatively, large vesi-cles might be formed only on demand. In the latter case, however, it is not immediately obvious how CUL3–KLHL12, located in the cytoplasm, would sense the presence of newly synthesized procollagen in the endoplasmic reticulum. A potential candidate for relay-ing this information across the endoplasmic reticulum membrane is the transmembrane protein TANGO1, which forms part of a packaging receptor that is essential for

CELL B IOLOGY

Collagen secretion explained Cells package proteins into vesicles for secretion to the extracellular milieu. A study has now identified an enzyme that modifies the packaging machinery to encapsulate unusually large proteins, such as collagen. S A .

Figure 1 | Big vesicles for collagen secretion. a, Soluble proteins targeted for secretion, together with small transmembrane proteins, are packaged at the endoplasmic reticulum into vesicles that are coated by the COPII protein cage. Proteins that will form the inner layer of the COPII coat associate in an ordered fashion and then recruit the proteins SEC13 and SEC31, which form the outer layer. This leads to membrane deformation and ultimately to scission of 60–80-nm transport vesicles. b, Large proteins such as procollagen (the collagen precursor) do not fit into these typical vesicles. Jin et al.2 report that, to encapsulate such large cargoes, the enzyme CUL3–KLHL12 attaches one copy of the small protein ubiquitin to SEC31 within the SEC13–SEC31 complex, and that this process facilitates collagen export. An additional, unknown protein might further stabilize lateral SEC13–SEC31 interactions. Although it is not known whether collagen synthesis directly triggers CUL3–KLHL12 activity, the transmembrane protein TANGO1 — which couples collagen in the endoplasmic reticulum to the assembling coat on the cytosolic face — might have a role in the process.

a

Membrane

Endoplasmicreticulum

Inner-layercomponents

b

SEC13–SEC31complex

Largetransportvesicle

Transport vesicleCOPII cage

60–80 nm

TANGO1

Ubiquitin

Procollagen

CUL3–KLHL12

Unde!nedprotein

Cytosol

Small transmembraneand soluble proteins

4 7 4 | N A T U R E | V O L 4 8 2 | 2 3 F E B R U A R Y 2 0 1 2

NEWS & VIEWS

procollagen secretion6,7. TANGO1, however, does not make contact with SEC31 directly, nor is it found in fully formed vesicles, and so its possible connection to CUL3–KLHL12 is unclear.

Other questions remain. Does collagen become entirely encapsulated in a large COPII cage during vesicle formation (Fig. 1b), or does COPII somehow aid collagen export indirectly, without the need for a complete cage? And how does the addition of ubiqui-tin change the geometry of the COPII coat? Jin and colleagues’ findings2 might aid the development of a cell-free system for study-ing COPII-dependent packaging of collagen that would help to address these issues.

More over, is SEC31 ubiquitination relevant to the packaging of other large secreted macromolecules, such as lipoproteins?

These questions are relevant to our understanding not only of the fundamental mechanisms of cellular secretion, but also of diseases in which secretion (particularly of collagen) is defective because of gene muta-tion8. Furthermore, manipulation of the CUL3–KLHL12 ubiquitination pathway might be used to increase collagen secre-tion from cells for applications in stem-cell culture, for growth of tissue components in regenerative medicine, or perhaps for amelio-rating age-related degeneration of connective tissue. ■

David J. Stephens is in the Cell Biology Laboratories, School of Biochemistry, University of Bristol, Bristol BS8 1TD, UK. e-mail: david.stephens@bristol.ac.uk

1. Zanetti, G., Pahuja, K. B., Studer, S., Shim, S. & Schekman, R. Nature Cell Biol. 14, 20–28 (2011).

2. Jin, L. et al. Nature 482, 495–500 (2012).3. Stagg, S. M. et al. Cell 134, 474–484 (2008).4. Fath, S., Mancias, J. D., Bi, X. & Goldberg, J. Cell 129,

1325–1336 (2007).5. Komander, D. Biochem. Soc. Trans. 37, 937–953

(2009).6. Saito, K. et al. Cell 136, 891–902 (2009).7. Wilson, D. G. et al. J. Cell Biol. 193, 935–951

(2011).8. De Matteis, M. A. & Luini, A. N. Engl. J. Med. 365,

927–938 (2011).

called the cosmic infrared background, has now been observed2 by a large team of astrono-mers working with data from the Planck space observatory. The results are part of a series of studies that form a collection of 26 papers, published by the Planck team in Astronomy & Astrophysics (see go.nature.com/au8vap).

The Planck satellite’s measurement of the cosmic infrared background2 improves on pre-vious measurements, including data3 obtained by Herschel, a twin observatory to Planck launched by the European Space Agency aboard the same rocket in 2009. The rocket carried them to the Earth–Sun Lagrangian point L2 (1.5 million kilometres from Earth in the opposite direction from the Sun), where the satellites can be stationary relative to both the Sun and Earth, allowing for shielding from the Sun’s radiation.

Planck detects microwave light in several wavelength bands in which the warm-dust emission can be observed (Fig. 1). Because the Universe is expanding and the wavelength of light stretches with the expansion, the light that we observe has a longer wavelength than it

ASTROPHYS ICS

First results from Planck observatoryEarly data from the Planck space satellite provide information about dust in distant galaxies, as well as in the Milky Way, and on the properties of gas in some of the largest clusters of galaxies in the Universe.

UROŠ S E L JAK

Astronomers have long known1 that most of the stars in the Universe are born in messy environments containing

dusty clouds. Young stars in such dust-enshrouded regions are not visible to optical telescopes; thus, multi-wavelength studies, from the radio to the X-ray regime, are used to better understand how stars form in our Gal-axy. But for more distant galaxies, including

some of the first galaxies in the Universe, such dusty expanses are essentially invisible across most wavelengths. One exception is the wave-lengths in the far-infrared and microwave regimes, which are roughly 1,000 times longer than those of visible light. Stars heat up the dust surrounding them to temperatures of roughly 20 kelvin — much lower than that of the stars themselves, but nevertheless high enough for the dust to radiate microwave and far-infrared light. This warm-dust signature,

Figure 1 | The cosmic infrared background. The images show the anisotropies, or irregularities, of the cosmic infrared background in three of the frequency channels (217 gigahertz, 353 GHz and 857 GHz) probed by the Planck observatory2

over a 26° × 26° patch of the sky. The anisotropies are visible as globular structures and correspond

to dusty galaxies clumped together on large scales. As we move across frequency channels, different epochs of cosmic time become visible: observations at 217 GHz offer a glimpse of some of the oldest galaxies in the Universe, which formed when the Universe was less than 2 billion years old.

217 GHz 857 GHz353 GHzES

A/P

LA

NC

K C

OLLA

BO

RATIO

N

2 3 F E B R U A R Y 2 0 1 2 | V O L 4 8 2 | N A T U R E | 4 7 5

NEWS & VIEWS RESEARCH