Nature © Macmillan Publishers Ltd 1998
AER excision and bead implantation. Fertilized chicken eggs (SPAFAS) were
the AER from the right forelimb was surgically excised and the embryo was
incubated for the time specified and was then fixed in 4% paraformaldehyde.
Heparin acrylic beads, of ?200?m in diameter, were soaked in 1mgml−1
FGF-4 (R&D Systems) for 1h at ambient temperature. One bead was stapled
Virus preparation and infections. An EcoRI fragment of BS(Sk-)–I?B-?DN
of 0.9kb was subcloned into a Cla12–Nco adaptor plasmid and then into
RCAS–BP(A)27. RCAS viral stocks were produced as described28. The vector
rAd–?DN was constructed by ligating the 0.9-kb Acc651 fragment of pBS(SK-
)–I?B-?DN into the Acc651 site of vector pAC. Identical construct formation
wasusedfor wild-typeI?B-?,pAC–I?B?.Weusedamodified procedurefrom
ref. 29 to prepare adenoviruses; additional details will be supplied upon
request. Fertilized chicken eggs were incubated at 39?C for 36–96h. The
eggs werewindowed andtheregionto beinfectedstainedwith Nile Blue.Virus
was loaded into a capillary pipet, attached to a picospritzer II and delivered at
injected at 12–16 locations, depending on the limb size. For infection of a
specific region, for example, the ZPA, eight injections were performed within
paraformaldehydefor insituhybridizationorin5%trichloroacetic acid (TCA)
for cartilage staining, respectively.
Cartilage staining. TCA-fixed embryos were stained overnight in 0.075%
Alcian Blue in acid alcohol, partially destained in 2% aqueous KOH,
dehydrated in 100% ethanol and cleared in methyl salicylate.
Scanning electron microscopy. Samples were fixed in 3% glutaraldehyde
buffered to pH 7.3 with sodium cacodylate. Following post-fixation in OsO4,
they were dehydrated, critical-point dried, and sputter-coated with gold before
viewing in a Hitachi S-500 scanning electron microscope.
Received 20 October 1997; accepted 2 February 1998.
1. Nu ¨sslein-Volhard,C.,Lohs-Schardin, M., Sander,K.&Cremer, C.Adorso-ventralshift ofembryonic
primordia in a new maternal-effect mutant of Drosophila. Nature 283, 474–476 (1980).
2. Hamburger, V. & Hamilton, H. L. A series of normal stages in the development of the chick embryo.
J. Exp. Morphol. 88, 49–92 (1951).
3. Summerbell, D. A quantitative analysis of the effect of the excision of the AER from the chick limb
bud. J. Embryol. Exp. Morphol. 32, 651–660 (1974).
4. Fallon, J. F. et al. FGF-2: apical ectodermal ridge growth signal for chick limb bud development.
Science 264, 104–107 (1994).
5. Niswander, L., Tickle, C., Vogel, A., Booth, I. & Martin, G. R. FGF-4 replaces the apical ectodermal
ridge and directs outgrowth and patterning of the limb. Cell 75, 579–587 (1993).
6. Vogel,A., Rodriguez, C.&Izpisua-Belmonte,J.C. Involvementof FGF-8in initiation, outgrowth and
patterning of the vertebrate limb. Development 122, 1737–1750 (1996).
7. Crossley, P. H., Minowada, G., MacArthur, C. A. & Martin, G. R. Roles for FGF8 in the induction,
initiation and maintenance of chick limb development. Cell 84, 127–136 (1996).
repeats in the inhibition of DNA binding activity. Proc. Natl Acad. Sci. USA 89, 4333–4337 (1992).
9. Brockman, J. A. et al. Coupling of a signal response domain in I?B? to multiple pathways for NF-?B
activation. Mol. Cell. Biol. 15, 2809–2818 (1995).
10. Treanckner, E. B.-M. et al. Phosphorylation of human I?B-? on serines 32 and 36 controls I?B-?
proteolysis and NF-?B activation in response to diverse stimuli. EMBO J. 14, 2876–2883 (1995).
sites in the murine c-rel promoter are required for constitutive c-rel transcription in B cells. Cell
Growth Differ. 4, 731–743 (1993).
12. Laufer, E., Nelson, C. E., Johnson, R. L., Morgan, B. A. & Tabin, C. Sonic hedgehog and FGF-4 act
through a signaling cascade and feedback loop to integrate growth and patterning of the developing
limb bud. Cell 79, 993–1003 (1994).
13. Niswander, L., Jeffrey, S., Martin, G. R. & Tickle, C. A positive feedback loop coordinates growth and
patterning in the vertebrate limb. Nature 371, 609–612 (1994).
14. Jiang, J., Kosman, D., Ip, Y. T. & Levine, M. The dorsal morphogen gradient regulates the mesoderm
determinant twist in early Drosophila embryos. Genes Dev. 5, 1881–1891 (1991).
15. Pan, D., Huang, J.-D. & Courey, A. J. Functional analysis of the Drosophila twist promoter reveals a
dorsal-binding ventral activator region. Genes Dev. 5, 1892–1901 (1991).
16. Huang, J. D., Schwyter, D. H., Shirokawa, J. M. & Courey, A. J. The interplay between multiple
enhancer and silencerelements defines the pattern of decapentaplegic expression. Genes Dev. 7, 694–
17. Schwyter, D. H., Huang, J. D., Dubnicoff, T. & Courey, A. J. The decapentaplegic promoter region
plays an integral role in the spatial control of transcription. Mol. Cell. Biol. 15, 3960–3968 (1995).
18. Gitelman, I. Twist protein in mouse embryogenesis. Dev. Biol. 189, 205–214 (1997).
mesoderm formation and patterning in the mouse. Genes Dev. 9, 2105–2116 (1995).
20. Yokouchi, Y. et al. BMP-2/-4 mediate programmedcelldeath in chickenlimb buds. Development 122,
21. Duprez, D. et al. Overexpressionof BMP-2 and BMP-4alters the sizeand shape of developing skeletal
elements in the chick limb. Mech. Dev. 57, 145–157 (1996).
22. Saunders, J. W., Jr & Gasseling, M. T. in Epithelial-Mesenchymal Interaction (eds Fleischmayer, R. &
Billingham, R. E.) 78–97 (Williams and Wilkins, Boston, 1968).
23. Ros, M., Lyons, G. & Fallon, J. Spatial and temporal analysis of homeobox genes expressed in chick
limb buds by whole mount in situ hybridization. Prog. Clin. Biol. Res. 383A, 79–87 (1993).
24. Echelard, Y. et al. Sonic hedgehog, a memberof a family of putative signalling molecules, is implicated
in the regulation of CNS parity. Cell 75, 1417–1430 (1993).
25. Winnier, G.E.,Hargett, L.&Hogan,B.L.Thewingedhelixtranscription factor MFH1is required for
proliferation and patterning of paraxial mesoderm in the mouse embryo. Genes Dev. 11, 926–940
26. Wall, N. A. & Hogan, B. L. Expression of bone morphogenetic protein-4 (BMP-4), bone morpho-
genetic protein-7 (BMP-7), fibroblast growth factor-8 (FGF-8) and sonic hedgehog (SHH) during
branchial arch development in the chick. Mech. Dev. 53, 383–392 (1995).
27. Hughes, S. H., Greenhouse, J. J., Petropoulos, C. J. & Sutrave, P. Adaptor plasmids simplify the
insertion of foreign DNA into helper-independent retroviral vectors. J. Virol. 61, 3004–3012 (1987).
Methods Cell Biol. 51, 185–218 (1996).
29. Becker, T. C. et al. Use of recombinant adenovirus for metabolic engineering of mammalian cells.
Methods Cell Biol. 43, 161–176 (1994).
Supplementary information is available on Nature’s World-Wide Web site (http://www.nature.com) or
as paper copy from Mary Sheehan at the London editorial office of Nature.
Acknowledgements. We thank B. L. M. Hogan, J. Barnett, C. Wright, D. Bader, L. Rollins-Smith, and
members of the Kerr laboratory for helpful discussions and critical review of this manuscript.
A. McMahon for avian Shh cDNA; B. Hogan for the M-Twist and Bmp-4 cDNA; C. B. Newgard for the
adenoviral shuttle and wild-type vectors; S. Hughes for RCAS-BP(A) and RCAS-AlkPhos vectors;
C. McCarther for the Fgf-8 cDNA; C. Tabin for cHoxA and cHoxD cDNA series. This work was
supported by an NIH grant, the American Cancer Society, and gene therapy pilot funds from the
Vanderbilt Cancer Center Grant and MSKCC Support Grant. P.B.B.is a predoctoral fellow supported by
an NIH training grant. L.D.K. is a recipient of an ACS Junior Faculty Research Award and a Cancer
Research Institute Investigator Award.
Correspondence and requests for materials should be addressed to L.D.K. (e-mail: KerrLD@ctrvax.
letters to nature
NATURE |VOL 392 |9 APRIL 1998
Rickard Glas*†, Matthew Bogyo*†, John S. McMaster†,
Maria Gaczynska† & Hidde L. Ploegh†
Center for Cancer Research, Department for Biology, Massachusetts Institute of
Technology, 77 Massachusetts Avenue, Cambridge, Massachusetts 02139-4307,
*These authors contributed equally to this work
Proteolysis is essential for the execution of many cellular func-
tions. These include removal of incorrectly folded or damaged
proteins1, the activation of transcription factors2, the ordered
degradation of proteins involved in cell cycle control3, and the
generation of peptides destined for presentation by class I mol-
ecules of the major histocompatibility complex4. A multisubunit
protease complex, the proteasome5, accomplishes these tasks.
Here we show that in mammalian cells inactivation of the
proteasome by covalent inhibitors allows the outgrowth of
inhibitor-resistant cells. The growth of such adapted cells is
apparently maintained by the induction of other proteolytic
systems that compensate for the loss of proteasomal activity.
Proteasomal function can be inhibited pharmacologically by
covalent or non-covalent modification of proteasomal ?-subunits.
Peptide aldehydes4act by forming a reversible hemiacetal adduct
with Thr1 on catalytically active ?-subunits6,7, whereas the natural
product lactacystin irreversibly alkylates Thr18–11. An amino-
terminally modified tri-leucine vinyl sulphone, NIPL3-VS (NLVS),
is a selective, covalent inhibitor that penetrates cell membranes
and can be used to inhibit proteasomes in living cells12. NLVS
and related peptide vinyl sulphones selectively and covalently
modify all three catalytically active (X, Y and Z) ?-subunits, as
well as the ?-interferon-inducible ?-subunits (LMP2, LMP7 and
MECL-1), with little or no crossreaction with non-proteasomal
EL-4 lymphoma cells, maintained in the presence of 10?M
NLVS, have drastically reduced proteasomal activity and die after
24–48 hours (Fig. 1). Homogenates from such EL-4 cells show
†Present addresses: Department of Pathology, Harvard Medical School, 200 Longwood Avenue, Boston,
Massachusetts02115, USA (R.G., M.B. and H.L.P.): Universityof UtahDepartment of Biochemistry, Salt
Lake City, Utah 84132, USA (J.McM.); Department of Molecular Medicine, Institute of Biotechnology,
Univesity of Texas Health Science Center, San Antonio, Texas 78245, USA (M.G.).
Nature © Macmillan Publishers Ltd 1998
letters to nature
NATURE |VOL 392|9 APRIL 1998
reduced hydrolysis of proteasome-preferred fluorogenic peptide
substrates (Fig. 1a). Moreover, their proteasomal ?-subunits are
refractory to modification with125I-labelled NLVS (Fig. 1b). The
lowest concentration of NLVS required to block proteasomal
activity (5–10?M) also inhibits the growth of EL-4 cells (Fig. 1a–
c); the toxic and inhibitoryeffectsofNLVSdepend onits tri-leucine
sequence, indicating that the toxicity is due to inhibition of the
proteasome (data not shown). In contrast, prolonged exposure of
EL-4 cells to the corresponding tri-leucine peptide aldehyde inhi-
bitor (Z-L3-H) is toxic at 0.5–1.0?M, a concentration at which
there is only minimal inhibition of proteasomal proteolysis4,12.
die (Fig. 1), after 2–3 weeks a minor population of EL-4 cells
recovers and grows (termed adapted cells; Fig. 1c, d). Limiting
dilution indicated a frequency of growth of ?1 in 300 cells in the
presence of 10?M NLVS, whereas 50?M NLVS did not allow any
recovery of cell growth (data not shown; Fig. 1c). This frequencyof
adaptation rules out mutation as the cause of resistance to NLVS, as
does the observation that removal of NLVS allows recovery of
proteasomal activity after 24h of culture (data not shown). EL-4
cells,adapted to grow in up to 50?M NLVS by gradually increasing
the concentration of the inhibitor, had a growth rate similar to
normal EL-4 cells. In overgrown cultures, adapted EL-4 cells died
more rapidly than control cells, which could be due to effects on
are limiting (Fig. 1d).
Exposure of intact cells to [125I]NLVS showed labelling of control
but not of adapted EL-4 cells, indicating prior modification of
proteasomal ?-subunits. To rule out the possibility that a change in
membrane permeability of adapted cells prevents access of NLVS to
the proteasome, proteolytic activity was monitored in cellular
homogenates. Hydrolysis of proteasome substrates Z-GGL-MCA,
Suc-LLVY-MCA (both substrates for the chymotrypsin-like activ-
ity) and Z-LLE-?NA (substrate for the PGPH (post glutamyl
peptide hydrolysing activity) was reduced in adapted EL-4 cells to
about 10–25% of control values, and labelling of proteasomal ?-
subunits by [125I]NLVS was largely absent (data not shown). Gel
filtration and the assay of individual fractions for hydrolysis of the
Suc-LLVY-MCA substrate showed a characteristic pattern of 20S
proteasome (Fig. 2a: main peak fractions, 20–25) and 26S protea-
some (Fig. 2a: shoulder fractions 16–19) activities in EL-4 cells,
whereas very little activity (?10%) remained in extracts prepared
from adapted EL-4 cells (Fig. 2a). Upon addition of [125I]NLVS, we
observed labelling of proteasomal ?-subunits in fractions from
control cells, corresponding to the peak of Suc-LLVY-MCA hydro-
lytic activity, but no labelling was detectable in fractions from
adapted cells (Fig. 2b). These combined results indicate that the
reduced proteasomal activity in adapted cells is attributable to
covalent modification of ?-subunits by NLVS.
We immunoprecipitated proteasomes from [35S]methionine-
Figure 1 Covalent modification of proteasomes in
EL-4 cells cultured in the presence of NLVS blocks
enzymatic activity. The block inproteasomal activity
inhibits cellular proliferation, but a minor population
inthe presenceof unlabelledNLVSfor the indicated
times and cellular homogenates of these cells were
tested against the fluorogenic peptide substrates
suc-LLVY-MCA (grey bars) and Z-GGL-MCA (black
bars), showing progressive loss of enzymatic
activity in a concentration- and time-dependent
manner. b, Cellular homogenates in a were labelled
with [125I]NLVS. c, EL-4 cells (100,000 cells per well)
were cultured with the indicated concentrations of
NLVS and counted daily. d, Cells recovered from
10?M NLVS were adapted to grow in the presence
of 50?M NLVS by gradually increasing the concen-
tration of NLVS during culture.
Figure 2 Peptidase activity from EL-4-adapted and control fractions resolved by
gel filtration (see Methods). Cytosolic fractions were centrifugated at high speed
and the pellets were resuspended and fractionated on a Superose-6 column.
Each fraction was tested for peptidase activity (a) and labelled with [125I]NLVS (b).
Nature © Macmillan Publishers Ltd 1998
letters to nature
NATURE |VOL 392 |9 APRIL 1998
labelled control and adapted EL-4 cells with an antiserum against
the proteasomal C9 ?-subunit. Analysis by two-dimensional SDS–
PAGEshowedthe characteristic pattern of ?- and ?-subunits13. The
relative intensities of the spots corresponding to ?-subunits varied
slightly, but the labelling pattern was comparable for EL-4 and
adapted EL-4 cells (Fig 3a, b). The most obvious change was a
comigration of LMP7 and LMP8 in proteasomes fromadapted cells
due to a shift in mobility of NLVS-modified LMP7 (Fig. 3b;
indicated as LMP7*). This was confirmed in vitro by showing that
NLVS-treatment of isolated proteasomal subunits from EL-4 cells
also resulted in comigration of NLVS-modified LMP7 with LMP8
(Fig. 3c, d).
Proteasomal proteolysis has been implicated in the generation of
peptides4,14–17. Exposure of cells to either peptide aldehydes or
lactacystin causes a temporary deficiency in the supply of peptides
for MHC class I molecules4,16, leading to a reduction in the amount
ofassembledMHC classIproducts. InEL-4cells,newlysynthesized
class I heavy chains rapidly form stable complexes with ?2-micro-
globulin (?2m) and acquire complex-type glycans (Fig. 4a, d). In
EL-4 cells treated with NLVS, there is a substantial reduction in the
formation of assembled MHC class I complexes. As reported for
L929 cells18, we do not observe complete inhibition of MHC class I
assembly and transport, suggesting that non-proteasomal proteases
may contribute a fraction of MHC class I-presented peptides in
these cells(Fig. 4). Addition of MHC-binding peptide to cell lysates
restored the recovery of assembled MHC class I molecules that lack
terminal glycan modifications, which indicates an arrest of intra-
cellular transport caused by a shortage of MHC ligands (Fig. 4b, e).
However, NLVS-adapted EL-4 cells displayed almost normal levels
indicates that there is a non-proteasomal source of MHC ligands
In non-adapted EL-4 cells treated with the proteasome inhibitor
NLVS, degradation of ubiquitinated proteins is inhibited, as shown
by an increase in ubiquitin-conjugated material (relative molecular
masses of 50K–200K) revealed by immunoblotting (data not
shown). Staining the DNA of these cells with propidium iodide
of the cell cycle in comparison with control cells (23% in G2),
Figure 3 Synthesis and assembly of proteasome subunits is normal in adapted
cells. a–d, Adapted and control EL-4 cells were pulse-labelled for 1h and chased
for 24h. Proteasomes were immunoprecipitated from either normal EL-4 cells (a)
EL-4 proteasomes were precipitated using a rabbit polyclonal antiserum raised
against purified mouse proteasomes. These proteasomes were either not
treated(c) or treated in vitrowith 10?M unlabelledNLVS for 2h (d). Sampleswere
separated by two-dimensional non-equilibrium pH gradient electrophoresis and
subunits visualized by fluorography. Note that the relative intensities of spots are
the same for control and adapted cells, except for LMP-7 (NLVS-modified LMP7
comigrating with LMP8 is indicated as LMP7*). The spot corresponding to LMP7
was identified by its relative molecular mass and isoelectric point13.
Figure 4 Adapted EL-4 cells generate ligands for stable assembly and
intracellular transport of MHC class I molecules. Cells were pulsed with [35S]
4 (a, d), NLVS-treated EL-4 (b, e) and NLVS-adapted EL-4 (c, f). To assess total
levels of class I heavy chain, regardless of peptide occupancy or state of
assembly, we used an antiserum (anti-p8) against the cytoplasmic tail of H–2Kb
(a–c). To detect properly conformed H–2Kbmolecules, we used the monoclonal
antibody Y3 (d–f). Lysates were exposed to 37?C for 20min in the presence or
absence of 10?M of an H–2Kb-binding peptide (SIINFEKL), or were kept at 4?C.
Only peptide-stabilized class I molecules maintain their conformation (Y3-
reactive) at 37?C. Bands corresponding to the heavy chain (HC) and ?2-m are
Control activity (%)
AAF-cmk added to assay:
EL-4 + 50 µM NLVS
EL-4 + 5 µM AAF-cmk
EL-4ad + 50 µM NLVS
EL-4ad + 50 µM NLVS
+ 5 µM AAF-cmk
Figure 5 AAF-MCA hydrolysis is increased in adapted cells. a, Cytosol prepared
from independent cultures (bars 2–7) of adapted EL-4 cells shows increased
hydrolysisof AAF-MCA peptide substrate and decreased hydrolysisof Cbz-GGL-
MCA(proteasomesubstrate)compared with EL-4 (bar1).b, The cytosolic activity
that cleaves AAF-MCA is blocked by AAF-CMK both in vivo and in vitro. c, d,
control EL-4 cells (c).
Nature © Macmillan Publishers Ltd 1998
letters to nature
NATURE |VOL 392|9 APRIL 1998
consistent with proteasomal proteolysis being required at the
G2/M–G1 transition3,19–21. In adapted EL-4 cells growing in the
presence of NLVS, no cells accumulated in the G2/M phase (25% in
G2), and the levels of ubiquitinated proteins were comparable to
those in untreated EL-4 cells (data not shown). Thus, in adapted EL-4
cells, cell-cycle progression (Fig. 1d) and degradation of ubiquitinated
proteins continue in the absence of normal proteasomal function.
The occurrence of a large (Mr? 1;000K) protease complex in
Thermoplasma, termed tricorn protease22, inspired us to examine
the possibility of similar large complexes in adapted cells that could
functionally replace proteasomal activity. Tricorn protease is resis-
tant to proteasomal inhibitors and, characteristically, hydrolysis of
the AAF-MCA substrate requires a free amino terminus. We indeed
observed a remarkable increase in AAF-hydrolysing activity in
adapted compared with normal EL-4 cells (Fig. 5a).
Although the active site of the tricorn protease has not yet been
identified, it may contain a serine nucleophile22. As chloromethyl-
ketones are potent irreversible inhibitors of serine proteases23, we
synthesized AAF-chloromethylketone (AAF-CMK) and found that
it blocked AAF-hydrolysing activity in a dose-dependent manner,
achieving ?80% inhibition in vitro at 10?M AAF-CMK (Fig. 5b)
without inhibiting proteasomal activity. If the AAF-MCA hydro-
lysing activity in adapted cells can compensate for loss of the
proteasome, then inhibiting this activity should prevent growth of
adapted EL-4 cells. There was no effect of 5?M AAF-CMK on the
growth of normal EL-4 cells, whereas proliferation of adapted EL-4
cells was inhibited (Fig. 5c, d). The AAF-MCA cleaving activity
must therefore be part of a protease or protease complex that
contributes to or is required for survival of adapted cells.
To distinguish the induced protease activity, from the pro-
teasome, we isolated material by centrifugation of cytosol for 5
hours at 100,000g, fractionated this bygel filtration, and tested each
fraction for its capacity to hydrolyse AAF-MCA (Fig. 6; solid lines).
The peak of AAF-MCA hydrolysing activity resolved from the void
AAF-hydrolytic activity is thus associated with a complex larger
than the proteasome. This activity cannot be inhibited by NLVS,
whereas addition of 10?M AAF-CMK blocks it completely (data
not shown, and Fig. 5). Analysis by SDS–PAGE and silver staining
showed that in control cells there was a cluster of proteasomal ?-
subunit-sized polypeptides of 20K–30K in fractions that cleave the
proteasome substrate LLVY-MCA, and in adapted cells there was a
complex set of polypeptides in the fractions that cleave AAF-MCA.
inhibiting proliferation5,24(Fig. 1). We have described the out-
growth of cells that, upon prolonged exposure to the proteasome
inhibitor NLVS, appear to be resistant to the inclusion of otherwise
toxic concentrations of this inhibitor. In these adapted cells there is
an increase in enzymatic activity capable of hydrolysing AAF-MCA
activity is blocked. This activity is separable from both 26S and 20S
proteasomes in specificity, size and inhibition profile. We propose
that a minority of cells expressauxiliary proteases. In the absence of
proteasomal inhibition, these cells have no obvious growth advan-
tage and are not represented in appreciable numbers. In the
continuous presence of NLVS, however, the rare cells that express
more AAF-hydrolysing activity are spared and recover to expand
after two weeks of culture, relying at least partly on AAF-hydrolys-
ing activity to do so.
We conclude that proteasomes can be replaced functionally by
other protease activities, which raises questions about the fate of
ubiquitin-conjugated proteins in adapted cells, on the mechanisms
used by adapted cells to regulate their cell cycle, and on the possible
involvement of non-proteasomal cytosolic proteases in MHC class
I-restricted antigen presentation. Destruction of cyclins and cyclin-
dependent-kinase inhibitors has almost invariably been attributed
to the ubiquitin–proteasomal system3,25–27. The detailed molecular
characterization of the activities that allow adapted cells to survive
will be a challenge.
Cells and cellculture.Themouse lymphoma celllineEL-4was maintained in
Dulbecco’s modified Eagle’s or RPMI 1640 medium supplemented with 10%
(v/v) fetal calf serum. Adapted EL-4 cells were maintained in the presence
of 50?M 5-iodo-4-hydroxy-nitrophenyl acetyl-leucyl-leucyl-leucine vinyl
Inhibitors. [125I]NIP-L3-VS and NIP-L3-VS (NLVS) were synthesized as
described12. All inhibitors were added to lysates and to cells by dilution of a
dimethylsulphoxide (DMSO) stock solution such that the final concentration
of DMSO was 1%.
Antibodies. Anti-C9 is a monoclonal antibody against a proteasomal a-
subunit in assembled proteasomes; rabbit anti-mouse proteasome is an
antiserum directed against assembled mouse proteasomes (gift from J.
Monaco)13; Y3 (ref. 28) is a monoclonal antibody that recognizes the ?1/?2
domain of properly folded mouse class I heavy chains; anti-P8 is a rabbit
antiserum specific for exon 8 in the cytoplasmic tail of H–2Kbmouse MHC
class I heavy chains; and anti-ubiquitin is a polyclonal antiserum reactive
against ubiquitin and ubiquitin conjugates (gift from A. L. Haas).
inhibitor or DMSO (1%) in 24-well plates. Cells were resuspended daily, 20-?l
aliquots were removed and mixed with trypan blue dye (0.1%). Cells that
excluded the dye were considered viable and were counted.
Labelling of proteasomes with [125I]NIP-L3-VS. Cells (1:5 ? 106) or lysates
(50?g unless otherwise indicated) were labelled with [125I]NLVS, diluted to a
final concentration of 1:8 ? 104Bqml?1in tissue culture medium (cells) or
buffer I (lysates; 50mM Tris, pH 7.4, 2mM DTT, 5mM MgCl2, 2mM ATP).
Samples were incubated at 37?C for 2h then 1× (cells) or 4× (lysates) SDS-
sample buffer was added. Proteins were separated by SDS–PAGE.
Preparation of lysates from control EL-4, NLVS-treated or adapted cells.
microns, acid-washed; Sigma) equivalent to the volume of the cellular pellet
were added, followed by the same volume of homogenization buffer (50mM
vortexed for 1min. Beads and cell debris were removed by centrifugation at
1,000g for 5min, followed by 10,000g for 20min. Protein concentration was
determined using the BCA protocol (Pierce Chemical).
Fluorogenicpeptide substrate assay.Peptidase activity was measured using
the following fluorescent labelled peptide substrates: Suc-LLVY-MCA, Z-LLG-
MCA and Suc-AAF-MCA for analysis of chymotrypsin-like activity of the
proteasome; Boc-LRR-MCA for analysis of the trypsin-like activity of the
proteasome; Z–LLE-?NA for analysis of the PGPH activity of the proteasome;
and AAF-MCA for analysis of non-proteasomal proteolysis. Partially purified
proteasomes from 5-h pellets (5?g total protein), or total lysates (10?g) were
diluted to 100?l in only buffer I or 50mM Tris, pH 7.4. Inhibitors and
substrates (100?M) were added to samples as DMSO stocks. Samples were
30 20 1010
Control - LLVY
Adapted - AAF
Control - AAF
Figure 6 AAF-MCA hydrolysing activity elutes before the proteasome, indicating
that it has a higher Mr. Protein fractions from differential centrifugation of cytosols
(see Methods) were fractionated on a Superose-6 gel filtration column as for Fig.
2. Fractions were assayed for hydrolysis of the non-proteasomal substrate AAF-
MCA (solid lines) or Suc-LLVY-MCA (dashed line, indicating the elution of
Nature © Macmillan Publishers Ltd 1998
letters to nature
NATURE |VOL 392 |9 APRIL 1998
Fluorescence was measured using a Hitachi F4500 spectrofluorimeter.
Class I assembly. To monitor class I assembly, a pulse-chase experiment was
done using EL-4 cells, EL-4 cells treated with 50?M NLVS, or adapted EL-4
cells. NLVS-treated cells were pretreated with 50?M NLVS for 16h before the
pulse chase4. Stabilization was assayed as described29, using the OVA peptide
SIINFEKL. Class I molecules were isolated by immunoprecipitation using the
Y3 monoclonal antibody or p8 antiserum. Proteins were separated by SDS–
PAGE and visualized by fluorography.
Gel filtration of subcellular fractions. Lysates were prepared from control
EL-4 and EL-4 adapted cells (2–3 ? 108cells) and fractionated by differential
centrifugation. The 5-h 100,000g pellets were resuspended in 0.5ml homo-
genization buffer (50mM Tris, pH 7.0, 20% glycerol) and injected into a
Superose-6 column (1:5cm ? 30cm) at room temperature. The sample was
eluted at a flow rate of 0.2mlmin−1. Fractions of 0.5ml were collected and
of Suc-LLV-MCA and AAF-MCA.
Received 22 September 1997; accepted 2 January 1998.
1. Etlinger, J. D. & Goldberg, A. L. A soluble ATP-dependent proteolytic system responsible for the
degradation of abnormal proteins in reticulocytes. Proc. Natl Acad. Sci. USA 74, 54–58 (1977).
2. Palombella, V. J., Rando, O. J., Goldberg, A. L. & Maniatis, T. The ubiquitin-proteasome pathway is
required for processing the NF-?B1 precursor protein and the activation of NF-?B. Cell 78, 773–785
3. Glotzer, M., Murray, A. W. & Kirschner, M. W. Cyclin is degraded by the ubiquitin pathway. Nature
349, 132–138 (1991).
4. Rock, K. L. et al. Inhibition of the proteasome block the degradation of most cell proteins and the
generation of peptides presented on MHC class I molecules. Cell 78, 761–771 (1994).
5. Coux, O., Tanaka, K. & Goldberg, A. Structure and functions of the 20S and 26S proteasomes. Annu.
Rev. Biochem. 65, 801–847 (1996).
6. Seemuller, E. et al. The proteasome from Thermoplasma acidophilum: A threonine protease. Science
268, 579–582 (1995).
7. Lo ¨we, J. et al. Crystal structure of the 20S proteasome from the Archaeon T. acidophilum at 3.4A˚
resolution. Science 268, 533–539 (1995).
8. Omura, S. et al. Lactacystin, a novel microbial metabolite, induces neuritogenesis of neuroblastoma
cells. J. Antibiot. 44, 113–116 (1991).
9. Fenteany, G. et al. Inhibition of proteasome activities and subunit-specific amino-terminal threonine
modification by lactacystin. Science 268, 726–730 (1995).
10. Dick, L. R. et al. Mechanistic studies on the inactivation of the proteasome by lactacystin. J. Biol.
Chem. 271, 7273–7276 (1996).
11. Groll, M. et al. Structure of 20S proteasome from yeast at 2.4A˚resolution. Nature 386, 463–471
homolog HslV by a new class of inhibitors. Proc. Natl Acad. Sci.. USA 94, 6629–6634 (1997).
13. Nandi, D., Woodward, E., Ginsburg, D. B. & Monaco, J. J. Intermediates in the formation of mouse
20S proteasomes: implications for the assembly of precursor beta subunits. EMBO J. 16, 5363–5375
14. Driscoll, J.,Brown,M.G.,Finley,D.&Monaco,J.J.MHC-linked LMPgeneproductsspecificallyalter
peptidase activities of the proteasome. Nature 365, 262–264 (1993).
15. Gaczynska, M., Rock, K. L. & Goldberg, A. g-Interferon and expression of MHC genes regulate
peptide hydrolysis by proteasomes. Nature 365, 264–267 (1993).
16. Craiu, A.C.et al. Lactacystin and clasto-lactacystin b-lactone modify multiple proteasome b subunits
17. Bai, A. & Forman,J. The effect of proteasome inhibitor lactacystin on the presentation of transporter
associated with antigen processing (TAP)-dependent and TAP-independent peptide epitopes byclass
I molecules. J. Immunol. 159, 2139–2146 )1007).
18. Vinitsky, A. et al. The generation of MHC class I-associated peptides is only partially inhibited by
proteasome inhibitors: involvement of non-proteasomal cytosolic proteases in antigen processing? J.
Immunol. 159, 554–564 (1997).
19. Hilt, W. & Wolf, D. H. Proteasomes of the yeast S. cerevisiae: genes, structure and function. Mol. Biol.
Rep. 21, 3–10 (1995).
20. Ghislain, M., Udvardy, A. & Mann, C. et al. S. cerevisiae 26S protease mutants arrest cell division in
G2/metaphase. Nature 366, 358–362 (1993).
21. Gordon, C., McGurk, G., Dilion, P., Rosen, C. & Hastie, N. D. Defective mitosis due to a mutation in
the gene for a fission yeast 26S protease subunit. Nature 366, 355–357 (1993).
22. Tamura, T. et al. Tricornprotease—the core of a modular proteolytic system. Science 274, 1385–1389
23. Kettner, C. & Shaw, E. Inactivation of trypsin-like enzymes with peptides of arginine chloromethyl
ketone. Meth. Enzymol. 80, 826–842 (1981).
Acad. Sci. USA 94, 855–860 (1997).
26. King, R. W., Jackson, P. K. & Kirschner, M. W. Mitosis in transition. Cell 79, 563–571 (1994).
27. Sheaff, R. & Roberts, J. M. End of the line: proteolytic degradation of cyclin-dependent kinase
inhibitors. Chem. Biol. 3, 869–873 (1996).
28. Hammerling, G. J., Rusch, E., Tada, N., Kimura, S. & Hammerling, U. Localization of allo-
determinants on H-2bdetermined with monoclonal antibodies and H-2 mutant mice. Proc. Natl
Acad. Sci. USA 79, 4737–4741 (1982).
29. Ljunggren, H.-G. et al. Empty MHC class I molecules come out in the cold. Nature 346, 476–480
a fellow of the The Swedish Foundation for International Cooperation in Research and High Education
Randall D. York, Hong Yao, Tara Dillon, Cindy L. Ellig,
Stephani P. Eckert, Edwin W. McCleskey & Philip J. S. Stork*
The Vollum Institute for Advanced Biomedical Research, *Department of
Pathology, Oregon Health Sciences University, Portland, Oregon 97201, USA
Activation of mitogen-activated protein (MAP) kinase (also
known as extracellular-signal-regulated kinase, or ERK)1by
growth factors can trigger either cell growth or differentiation.
may determine the different effects of growth factors: for exam-
ple, transient activation of MAP kinase by epidermal growth
factor stimulates proliferation of PC12 cells1, whereas they differ-
entiate in response to nerve growth factor, which acts partly by
inducing a sustained activation of MAP kinase1. Here we show
distinct pathways: the initial activation of MAP kinase requires
G protein Rap1. Rap1 is activated by CRK adaptor proteins and
the guanine-nucleotide-exchange factor C3G, and forms a stable
complex with B-Raf, an activatorof MAP kinase. Rap1 is required
for at least two indices of neuronal differentiation by nerve
growth factor: electrical excitability and the induction of
neuron-specific genes. We propose that the activation of Rap1
by C3G represents a common mechanism to induce sustained
activation of the MAP kinase cascade in cells that express B-Raf.
PC12 cells are a well studied model of growth-factor specificity.
Treatment of PC12 cells with nerve growth factor (NGF) triggers
differentiation into sympathetic-like neurons, characterized by
electrical excitability, the induction of a set of neuron-specific
genes, and neurite outgrowth2,3. The ability of NGF to induce
sustained activation of the ERK family of MAP kinases has been
implicated in PC12 cell differentiation1,4–7. The molecular mechan-
isms responsible for sustaining ERK activation are not known.
However, signals that limit ERK activation have been identified.
In some cells, ERK directs the phosphorylation of the Ras guanine-
nucleotide-exchange factor, Sos, to terminate Ras-dependent ERK
activation8. Therefore, Ras-independent pathways may be required
to maintain ERK activation for sustained periods. As Rap1 can also
stimulate ERK in PC12 cells9, we examined whether Rap1 con-
tributes to NGF action in PC12 cells.
Using aninterferingmutantofRap1, RapN17 (ref.9), weshowed
that Rap1 is required for maximal activation of ERK by NGF in
PC12 cells. RapN17 blocked the ability of NGF to stimulate the
sustained phase of ERK activation in these cells, without inhibiting
the initial rapid phase of ERK activation. In contrast, RasN17, a
dominant-negative mutant of Ras, blocked only the initial phase of
ERK activation by NGF (Fig. 1a). It has been suggested that Ras is
essential for NGF signalling to ERKs10,11, but these studies investi-
gated only early time points of stimulation by NGF or used stably
transfected clonal variants of PC12 cells rather than wild-type cells.
Our results indicate that Ras and Rap1 mediate the initial and the
sustained phases of NGF-induced activation of ERK, respectively,
and that Ras and Rap1 act largely independently of each other.
As one of the nuclear targets of ERK is Elk-1, a transcrip-
tion factor of the Ets family9, NGF-induced activation of ERK
can be monitored by measuring the rate of Elk-1-dependent
transcription9. In PC12 cells, NGF stimulated Elk-1 more than did
EGF(Fig. 1b). Expression of RapN17 reduced NGF-induced activa-
tion of Elk-1 to levels observed following stimulation by epidermal