C-TERMINI OF PROTEASOMAL ATPases PLAY NON-EQUIVALENT ROLES IN CELLULAR ...

27 Pages·2011·1.13 MB·English

JBC Papers in Press. Published on May 31, 2011 as Manuscript M111.246793
The latest version is at http://www.jbc.org/cgi/doi/10.1074/jbc.M111.246793
C-TERMINI OF PROTEASOMAL ATPases PLAY NON-EQUIVALENT ROLES IN
CELLULAR ASSEMBLY OF MAMMALIAN 26S PROTEASOME
Young-Chan Kim and George N. DeMartino
Department of Physiology, University of Texas Southwestern Medical Center
5323 Harry Hines Boulevard, Dallas, TX 75390-9040
Running Head: ATPase subunits and 26S proteasome assembly
Address correspondence to: George N. DeMartino, 214.645.6024 (T); 214.645.6019 (F);
[email protected] (E-mail)

The 26S proteasome comprises two distinct subcomplexes: a protease, termed 20S
multisubunit subcomplexes: 20S proteasome proteasome or Core Particle, and an ATPase
and PA700/19S regulatory particle. The cellular regulator, termed PA700 or 19S Regulatory
mechanisms by which these subcomplexes Particle (3). The 700,000-dalton 20S proteasome
assemble into 26S proteasome and the consists of four axially-stacked heptameric rings.
molecular determinants that govern the Each of the two identical inner rings of eukaryotic
assembly process are poorly defined. Here we 20S proteasome contains seven different but
demonstrate the non-equivalent roles of the C- homologous -type subunits (1-7) and each of
termini of six AAA subunits (Rpt1-Rpt6) of the two identical outer rings contains seven D
o
w
PA700 in 26S proteasome assembly in different but homologous -type subunits (1-7) n
lo
mammalian cells. The C-terminal HbYX motif (4;5). The resulting 7-7-7-7 cylindrical ad
e
othf aeta cohf ao ft thwirod s usubbuunnitist,, RRpptt32 ,a wnda sR epsts5e,n btiuatl nfoort cpolamnpe ltehxo ufgeha ttuhree s rCin2g ss y(6m).m etry about an axial d from
h
assembly of 26S proteasome. The C-termini of In eukaryotic cells, the 20S proteasome is ttp
none of the three non-HbYX motif Rpt subunits assembled with the aid of at least five dedicated ://w
w
were essential for cellular 26S proteasome assembly chaperones (7;8). The process proceeds w
assembly, although deletion of the last three .jb
through multiple intermediate subassemblies and c
.o
residues of Rpt6 destabilized the 20S-PA700 concludes when two immature half-proteasomes rg
interaction. Rpt subunits defective for (1-7, 1-7 and associated chaperones) combine. by g/
assembly into 26S proteasome due to C- u
The final assembly step results in exposure of e
s
terminal truncations were incorporated into active site threonine residues by autolytic removal t on
intact PA700. Moreover, intact PA700 of pro-sequences from the N-termini of three Jan
accumulated as an isolated subcomplex when ua
catalytic  subunits (1, 2, and 5) and ry
cellular 20S proteasome content was reduced 5
degradation of the associated chaperones. , 2
by RNAi. These results indicate that 20S 0
Formation of intact 20S proteasome, however, 1
9
proteasome is not an obligatory template for
seals the catalytic threonines in an interior
assembly of PA700. Collectively, these results
chamber that physically sequesters them from
identify specific structural elements of two Rpt
protein substrates (5;9). Two other structural
subunits required for 26S proteasome
features of the 20S proteasome further limit its
assembly, demonstrate that PA700 can be
catalytic potential. First, the only route for
assembled independently of the 20S
substrates to reach the interior degradation
proteasome, and suggest that intact PA700 is a
chamber is through 13Å pores in the center of the
direct intermediate in the cellular pathway of
outer  rings (6). Second, the pores are reversibly
26S proteasome assembly.
gated by alternative conformations of the N-

termini of  subunits (10). Thus, substrate access
Cellular assembly of the 26S proteasome, a
to the central degradation chamber requires both
2,500,000-dalton protease complex responsible for
an open gate conformation of the proteasome and
the selective intracellular degradation of
unfolded polypeptide substrates capable of
polyubiquitylated proteins, requires coordinated
traversing the narrow access portal. As described
association of over 60 subunits representing the
below, the PA700 subcomplex functions to satisfy
products of ~35 genes (1);(2). These subunits are
each of these requirements.
arranged as two structurally and functionally
1
Copyright 2011 by The American Society for Biochemistry and Molecular Biology, Inc.
PA700 is composed of about 20 different complex remain unclear and are the subject of
subunits, including six homologous AAA (ATPase conflicting results. Support for an important role
associated with various cellular activities) proteins of HbYX residues in these processes comes from
(Rpt1-Rpt6) arranged as a heterohexameric ring both in vitro and cellular studies. For example, in
that forms the binding interface of PA700 with vitro reconstitution of 26S proteasome from
either or both of the heteroheptameric  rings of purified 20S proteasome and PA700 is abrogated
the 20S proteasome (3). Four additional subunits when the HbYX residues of Rpt2 and Rpt5 are
(Rpn1, Rpn2, Rpn13, and in many species a fifth, enzymatically removed (24). Moreover, Rpt3
Uch37) associate with the ATPase ring to lacking its HbYX residues is defective for
collectively constitute a PA700 substructure incorporation into 26S proteasome in mammalian
termed “base” (11). The remaining PA700 cells (25). However, other results appear to
subunits (Rpn3,5-13) constitute the “lid,” a contradict these findings and seem inconsistent
substructure positioned distally to 20S proteasome. with a model in which HbYX motifs are the
With the exception of peptide bond hydrolysis principal binding elements for the PA700-20S
per se, all functions necessary for 26S proteasome- proteasome interaction. For example, 26S
catalyzed degradation of polyubiquitylated proteasome assembly is not appreciably
proteins are provided by PA700 (12). Thus, diminished in yeast when HbYX-motif tyrosine
PA700: i) selects substrates by binding residues are mutated in Rpt2, Rpt3, or Rpt5 (20), D
o
polyubiquitin chains (13); ii) prepares the client and deletion of the last residue of each Rpt subunit wn
lo
protein for passage through the substrate access severely diminishes 26S proteasome assembly for a
d
e
pchoarein sb y(1 4u-n1f6o)l,d ianngd ;i t iiai)n dtr adnestlaocchaitnegs tihtes uunbfioqludietidn, o(2n6ly). th eS uncohn -HapbpYarXe nmt odtiisfc sreupbaunncitise sR mpta4y o rr eRflpetc6t d from
deubiquitylated polypeptide to the catalytic differences in experimental design and detail http
chamber for hydrolysis (17). These processes among studies or reflect roles for Rpt C-terminal ://w
w
appear to be mechanistically coupled to one residues in cellular assembly processes other than w
another by PA700-catalyzed ATP hydrolysis, binding per se to the proteasome (see below). In .jb
c
.o
which is required for overall proteolysis of any case, these inconsistent results highlight the rg
ubiquitylated substrates (18). lack of understanding of the specific and relative by/
g
In addition to substrate binding and roles of the C-termini of Rpt subunits in u
e
s
processing, PA700 also mediates 26S proteasome proteasome assembly. t o
n
function by inducting the open gate conformation In contrast to the relatively advanced Ja
n
of the substrate access pore (19). Recent work has knowledge of 20S proteasome assembly, much ua
ry
revealed critical roles of the extreme C-termini of less is known about cellular mechanisms of PA700 5
, 2
certain Rpt subunits in this process. Three of the assembly and the relationship of this process to 0
1
9
six Rpt subunits (Rpt2, Rpt3, and Rpt5) feature C- 26S proteasome assembly. Recent studies have
terminal residues that conform to an HbYX motif, provided important but incomplete and, in some
where Hb is a hydrophobic residue, Y is tyrosine, instances, conflicting data regarding cellular
and X is any amino acid (20). These residues bind assembly of PA700. Multiple studies in yeast and
to pockets between specific  subunits of 20S more limited studies in mammalian cells show that
proteasome (21-24). Binding of HbYX residues of the six Rpt subunits initially associate as three
either of two Rpt subunits (Rpt2 and Rpt5) or of separate but discrete subunit pairs in association
the HbYX residues of PAN, an Rpt subunit with pair-specific chaperones (26-33). The three
ortholog from archaea, induces gate opening, even Rpt subassemblies subsequently combine to form
when these residues are in the form of isolated a structure equivalent to the base. Conflicting
short peptides corresponding to the C-terminus of data, however, have been obtained regarding the
respective subunits (20;24;25). role of the 20S proteasome in both this and
Despite the documented physical interactions subsequent steps of PA700 formation. These
between HbYX residues of PA700 and the 20S conflicts include a possible requirement for 20S
proteasome, the roles and contributions of these proteasome as a template for base assembly, the
interactions to cellular 26S proteasome assembly mechanism of lid assembly, and the physical and
and to the structural stability of the assembled temporal relationship between assembly of the
2
base and the lid (26;34;35). Thus, it remains clones exhibiting stable expression of respective
uncertain whether PA700 is formed as a separate proteins. Each stable cell line was grown to ~95%
intact entity prior to its binding to 20S proteasome confluence, harvested and washed with phosphate-
or whether it is assembled sequentially from buffered saline. Cells were disrupted in buffer ice-
subassemblies using 20S proteasome and perhaps cold buffer consisting of 50 mM Tris-HCl, pH 7.5
other PA700 subassemblies as templates. at 4oC, 5 mM MgCl , 1 mM -mercaptoethanol, 1
2
The purpose of the current work was to mM ATP and depending on the experiment,
determine the relative roles of the C-termini of Rpt Protease Inhibitor Cocktail (Roche) by 15
subunits for 26S proteasome assembly in intact passages through a 27-gauge needle. The lysates
mammalian cells. We sought to test the were centrifuged for 20 min to obtain a crude
hypothesis, based in part on previous in vitro soluble fraction. Expression of Rpt proteins was
biochemical studies, that intact HbYX residues of determined by western blot analysis using anti-
Rpt2, Rpt3, and Rpt5, but not corresponding non- FLAG M2 antibody (Sigma, St. Louis, MO) and
HbYX residues of Rpt1, Rpt4, and Rpt6, would be corresponding anti-Rpt antibodies.
essential for cellular 26S proteasome assembly,
and to establish whether the HbYX motifs of Glycerol Density Gradient Centrifugation.
different subunits played independent or Glycerol density gradient centrifugation was
cooperative roles in the assembly process (24;25). conducted as described previously using 10-40 % D
o
We also sought to evaluate the physiologic linear glycerol gradients in buffer containing 50 wn
lo
significance of our in vitro reconstitution assay of mM Tris-HCl, pH 7.5, 20 mM NaCl, 1mM - ad
e
2P6AS7 0p0ro tesausbocmome pfrloexme si soblayt edd e2t0eSrm pirnoitnega sowmhee tahnedr m(3e6r)c ap toCeothnatrnool l, c1e nmtrMifu gAaTtioPn sa nde s5ta bmliMsh eMd gCthle2 d from
PA700 is formed as an intact complex prior to its sedimentation positions of purified bovine 26S http
cellular assembly into 26S proteasome. Our results proteasome, PA700, and 20S proteasome (37). ://w
w
reveal that features of cellular 26S proteasome w
assembly mirror some but not all of those expected Affinity purification of proteasome .jbc
.o
from biochemical studies and provide new insights complexes. Cells were cultured, harvested and rg
to the mechanisms of this process. lysed as described above. Approximately 20 mg of by/
g
each cell extract normalized for the level of ue
s
EXPERIMENTAL PROCEDURES FLAG-Rpt protein was mixed gently for 2 hrs at t o
n
4°C with 100 µl of anti-FLAG M2 agarose beads Ja
n
DNA constructs. cDNAs encoding each full (Sigma, St. Louis, MO) in 50 mM Tris-HCl, pH ua
ry
length (WT) human Rpt subunit and each subunit 7.5 at 4oC, 100 mM NaCl, 1 mM - 5
, 2
lack three C-terminal amino acids (-C3) were mercaptoethanol, 1 mM ATP, 5mM MgCl2, 10% 019
subcloned into pIRESpuro3 expression vector glycerol and 0.1% NP-40. The beads were
(Clonetech) featuring N-terminal FLAG epitope. harvested by centrifugation and washed three
Each construct was confirmed by DNA times with the same buffer. Bound proteins were
sequencing.
eluted at 4°C with 150 l of elution buffer
containing 200 µg/ml FLAG peptide (Sigma, St.
Preparation of HEK293 cell lines with stable Louis, MO). Identical volumes of each affinity-
expression of FLAG-tagged Rpt subunits.
purified protein sample (10 l) were characterized
HEK293 cell lines were cultured in Dulbecco's
by proteasome activity, SDS-PAGE, western
modified Eagle's medium (Gibco) containing high
blotting, and native PAGE, as described in the text
glucose and glutamine, supplemented with 10%
and individual figure legends.
fetal bovine serum in the presence of 5% CO at
2
37°C. HEK293 cells were transfected at
Proteasome Activity. Proteasome activity
approximately 60% confluence with respective
was measured fluorescently by determining the
vectors using FuGene 6 (Roche). Forty-eight hrs
rate of hydrolysis of AMC from the peptide Suc-
after transfection, the media was replaced with
Leu-Leu-Val-Tyr-AMC, as described previously
media containing 5 µg/ml of puromycin and cells
(18). All values represent the mean of triplicate
were incubated further for 4 weeks for selection of
3
assays and are expressed as arbitrary fluorescent Normal HEK293 cells, like most other commonly-
units (AFU) produced per min. studied mammalian cells, feature a bimodal
distribution of proteasome activity in gradient
PA700 activity. PA700 activity was fractions. Most proteasome activity sediments at a
determined the ATP-dependent activation of 20S position characteristic of 26S proteasome whereas
proteasome activity (38). In brief, PA700 was a smaller amount of slower-sedimenting activity is
preincubated with 100 ng purified bovine 20S found in a position characteristic of 20S
proteasome, prior to addition of Suc-Leu-Leu-Val- proteasome (Figure 1A). Western blotting of
Tyr-AMC and proteasome assay, as described gradient fractions reveals a bimodal distribution of
above. Assays were performed in triplicate and 20S proteasome subunits corresponding to the
activities are expressed as arbitrary fluorescent proteasome activity profiles Figure 1B). Western
units (AFU) produced per min. blotting also showed that the majority of each Rpt
subunit was present in 26S proteasome. Thus,
Native polyacrylamide gel electrophoresis these subunits co-sedimented with one another,
(PAGE). Non-denaturing PAGE was conducted with other subunits of the 26S proteasome, and
with 4% polyacrylamide gels, as described with 26S proteasome activity (Figure 1). Lesser
previously (18). Gels were either stained with amounts of the Rpt subunits sedimented at a
silver, blotted for proteins of interest, or assayed position characteristic of free intact PA700. Small D
o
for in-gel proteasome activity by overlaying a 50 but detectable amounts of several subunits wn
lo
M solution of Suc-Leu-Leu-Val-Tyr-AMC. including Rpt3, Rpt4 and Rpt5 sedimented a
d
e
Aprfotedru citnicounb wataiosn v ifsoura l1i5ze-3d0 b my ienxsp oats u3r7e o oCf ,t hAe MgeCl spirgonbiafibclayn tlyre prselsoewnet r stuhbaans seimntbalciet s PAof7 00P A7an0d0 d from
to fluorescent light. described previously (29;30;39). http
To analyze the relative roles of the C-termini ://w
w
RNA interference (RNAi) with siRNA. RNAi of Rpt subunits in 26S proteasome assembly, we w
of the β5 subunit of 20S proteasome was created a series of HEK293 cell lines that stably .jbc
.o
conducted by transfection of double-stranded express N-terminally FLAG-tagged versions of a rg
siRNA oligonucleotides (5’- given Rpt subunit either as a wild-type (WT) by/
g
GAAGGUGAUAGAGAUCAAC-3’, Dharmacon, protein or as a protein whose three C-terminal ue
s
Boulder, CO), as described previously (36). residues are deleted (-C3). We utilized the t o
n
Control transfections used a proven non-targeting epitope tag to monitor the steady-state Ja
n
siRNA (D-001210-01-05) provided by incorporation of these proteins into 26S ua
ry
Dharmacon. The HEK293 cells expressing proteasome and/or other protein complexes by 5
, 2
FLAG-Rpt6 (WT) were plated, grown to 40-50 % glycerol density gradient centrifugation, as 0
1
9
confluence and transfected with 20 nM siRNA described above for endogenous proteins. We also
using Lipofectamine 2000 (Invitrogen, Carlsbad, exploited the epitope tag to affinity-purify and
CA). 72 hrs after transfection, cells were harvested characterize protein complexes into which these
and cell lysates were prepared and analyzed by proteins were assembled. In this manner, we
Western blotting. The remaining cell extracts could determine the influence of the C-terminus
were subjected to FLAG affinity-purification and on assembly of a given Rpt subunit into 26S
characterized as described in the text. proteasome, compare the relative roles of C-
termini of different Rpt subunits for this process,
RESULTS and determine the fate of Rpt subunits that were
defective for 26S proteasome assembly. To avoid
Exogenously expressed Rpt subunits of potential complications of protein overexpression
PA700 are incorporated into intact 26S in this analysis, we initially selected cell lines that
proteasome in mammalian cells. Structural and expressed the FLAG-tagged proteins at levels
functional states of proteasome complexes in approximately equal to or less than those of their
cultured cells can be analyzed simply and respective endogenous counterparts.
reproducibly by glycerol density gradient With the exception of Rpt2 (see below), each
centrifugation of corresponding cell-free extracts. wild-type FLAG-tagged Rpt protein was
4
incorporated into intact 26S proteasome as judged For example, the distributions of FLAG-Rpt3 and
by the distribution of FLAG-Rpt protein in FLAG-Rpt6 closely mirrored those of their
fractions from the glycerol density gradient respective endogenous counterparts such that most
centrifugation of cell extracts (Panel B in Figures of these proteins were found in 26S proteasome.
2,3,5-7). Additional structural and functional In contrast, significant proportions of FLAG-Rpt5,
analysis of protein complexes purified with anti- Rpt1, and Rpt4 were present in fractions that
FLAG beads confirmed this conclusion. For sedimented slower than 26S proteasome. Most
example, SDS-PAGE of the affinity-purified FLAG-Rpt5 not associated with 26S proteasome
proteins revealed a set of stained proteins protein sedimented in fractions characteristic of
characteristic of purified 26S proteasome subunits. intact PA700 (Figure 1B), whereas significant
Western blotting with antibodies against selected amounts of non-26S proteasome-associated
26S proteasome subunits revealed the presence of FLAG-Rpt1 and FLAG-Rpt4 sedimented both in
subunits for both the base and lid subassemblies of fractions characteristic of PA700 and in gradient
PA700 as well as for 20S proteasome (Panel C in fractions smaller than PA700 (Figures 5B and 6B,
Figures 2,3,5-7). The affinity-purified proteins respectively). These results suggest that different
also featured proteasome-specific peptidase FLAG-Rpt proteins were incorporated into 26S
activity (Panel D in Figures 2,3,5-7) and proteasome with different efficiencies. Despite
electrophoretic migration positions characteristic these differences, the incorporation of wild-type D
o
of singly and doubly capped 26S proteasome on FLAG-tagged Rpt subunits into 26S proteasome wn
lo
native polyacrylaminde gels (Panel E in Figures indicates that these proteins can be used to monitor a
d
e
2st,a3i,n5i-n7g)., wTheset elrant terb lwotetirneg detwecitthe d abnyt i-pFrLotAeiGn featu r e s of the assembly process. d from
antibodies, and proteasome activity using an in-gel C-terminal HbYX motifs of two Rpt http
peptidase overlay assay. subunits, Rpt5 and Rpt3, are essential for ://w
w
In contrast to other FLAG-Rpt wild-type assembly of 26S proteasome. Previous work by w
proteins, FLAG- Rpt2(WT) appeared to undergo us and others has established an obligatory role for .jb
c
.o
C-terminal proteolytic processing, as judged by its the C-terminal HbYX motifs of eukaryotic Rpt2, rg
increased mobility on SDS-PAGE when compared Rpt3, and Rpt5 subunits and the related archaeal by/
g
to endogenous Rpt2 (Figure 4, Panel A). The PAN in binding of these proteins to the 20S u
e
s
processing, which decreased the apparent size of proteasome. This analysis was conducted t o
n
the protein by about 10 kDa, most likely occurred principally in vitro using isolated peptides J
a
n
in intact cells rather than during preparation and corresponding to the proteins’ C-termini or with ua
ry
analysis of the extract and also was observed when various homomeric protein complexes (20;21;23- 5
, 2
the protein was expressed transiently (data not 25). Therefore, to test the relative roles of the three 0
1
9
shown). Most FLAG-Rpt2(WT) protein HbYX motifs of the Rpt subunits in cellular
sedimented slowly during glycerol density assembly of mammalian 26S proteasome, we
gradient centrifugation, indicating that the repeated the analysis described above with FLAG-
processed protein was present mainly in low Rpt proteins whose HbYX motifs were deleted.
molecular weight complexes rather than in 26S FLAG-Rpt5 and FLAG-Rpt3 lacking their HbYX
proteasome or intact PA700 complexes (Figure 4, motifs failed to be incorporated to appreciable
Panel B). Affinity purification of Flag-Rpt2(WT) extents into 26S proteasome, as judged by their
complexes displayed low levels of both 26S greatly reduced presence in glycerol gradient
proteasome activity and intact 26S proteasome fractions characteristic of 26S proteasome (Figures
protein (Figure 4, Panels C-E). Additional 2B and 3B). Analysis of the corresponding
characterization of FLAG-Rpt2(WT) protein is affinity-purified protein complexes confirmed this
described below. conclusion and demonstrated clear distinctions
The FLAG-Rpt(WT) proteins that successfully between the respective wild-type and mutant
assembled into 26S proteasome featured proteins. Thus, affinity-purified proteins from
distribution profiles on glycerol density gradients neither FLAG-Rpt3(-C3) nor FLAG-Rpt5(-C3)
that differed from one another, and in some cases, cell extracts displayed significant proteasome
from their respective endogenous counterparts. activity or featured complexes characteristic of
5
26S proteasome after native PAGE (Figures 2C-E profile of endogenous Rpt2, and was confirmed by
and 3C-E, respectively). Instead, these affinity- isolation of intact, catalytically active 26S
purified protein complexes featured several proteasome by anti-FLAG affinity purification
smaller catalytically inactive complexes including (Figure 4C). Thus, unlike the HbYX motifs of
increased amounts of PA700. SDS-PAGE Rpt3 and Rpt5, the HbYX motif of Rpt2 does not
followed by protein staining or western blotting appear to be essential for assembly of the protein
for 26S proteasome subunits revealed that the into 26S proteasome. As described above, the
main difference between the wild type and mutant proteolytic processing of FLAG-Rpt2(WT) greatly
proteins was the absence of 20S proteasome reduced the efficiency of incorporation of the
subunits in complexes of the latter. To confirm protein into 26S proteasome, thereby preventing a
the exclusion of the mutant proteins from 26S direct comparison of the deletion mutant with its
proteasome complexes, we also subjected affinity- wild-type counterpart. It is unclear why FLAG-
purified complexes from FLAG-Rpt5(WT) and Rpt2(WT) protein, but neither FLAG-Rpt2(-C3)
FLAG-Rpt5(-C3) cell extracts to glycerol density nor endogenous Rpt2 is susceptible to
gradient centrifugation. Only complexes purified modification.
from extracts of expressing wild-type proteins Previous work has established the selective
displayed 26S proteasome activity that co- roles of the HbYX residues of Rpt2 and Rpt5 in
sedimented with proteins detected with antibodies proteasome activation by gating (20;24;25). D
o
against FLAG- and 20S proteasome subunits Purification of 26S proteasome containing Rpt2 wn
lo
(Figure 2F and data not shown). lacking its HbYX motif allowed us to determine a
d
e
wildU-tnylpiek ec eoiuthneterr iptasr tesn, dmoguetnanotu sR oprt 3F LaAppGe-atraegdg etdo tchoem pcaornintrgi bpuetipotnid asoef aRctpivt2it ietso oft htihs e pmrouctaensst anbdy d from
undergo C-terminal proteolytic processing. Thus, wild-type proteins. Because of the proteolytic http
in addition to the intact mutant protein (~ 45 kDa), modification of FLAG-Rpt2(WT) we utilized ://w
w
two truncated FLAG-Rpt3 proteins (~40 kDa and other wild-type 26S proteasomes, including both w
~30 kDa) also were detected by anti-FLAG FLAG-tagged 26 proteasome affinity purified .jb
c
.o
western blotting (Figure 3A and 3B). The 30 kDa from HEK293 cells and 26S proteasome purified rg
protein differed from the others by sedimenting from bovine red blood cells for this comparison. by/
g
more slowly during glycerol density gradient 26S proteasome containing FLAG-Rpt2(-C3) u
e
s
centrifugation and not assembling significantly displayed approximately 20% of the specific t o
n
into PA700 (Figure 3B). Although the reason for peptidase activity of various wild-type 26S J
a
n
the susceptibility of mutant Rpt3 to proteolysis proteasomes (Supplemental Figure S1). This ua
ry
may be a direct consequence of the HbYX result indicates that the HbYX motif of Rpt2 plays 5
, 2
deletion, we note that the same mutant protein was a significant role in proteasome gating and 0
1
9
not processed when expressed transiently at about distinguishes this effect from its non-essential role
the same level in HEK293 cells (25). in proteasome assembly.
Nevertheless, the features of unprocessed mutant Collectively, these results demonstrate
Rpt3 demonstrate that the loss of only the HbYX differential effects of the three HbYX-containing
residues was sufficient to prevent incorporation of Rpt subunits on their incorporation into 26S
this subunit into 26S proteasome but not into intact proteasome. Remarkably, for each of the two
PA700. subunits rendered assembly-defective by this
deletion, the presence of five other wild-type Rpt
The C-terminal HbYX motif of Rpt2 is not subunits is insufficient to overcome this defect.
essential for assembly of 26S proteasome but Moreover, the two mutant subunits incapable of
promotes proteasome gating. In contrast to Rpt3 incorporation into 26S proteasome are
and Rpt5, deletion of the HbYX motif from Flag- nevertheless competent for assembly into intact
Rpt2 did not prevent its assembly into 26S PA700.
proteasome. This result was readily apparent from
the distribution profile of the mutant protein after C-terminal residues of none of the three non-
glycerol density gradient centrifugation (Figure HbYX motif Rpt subunits are essential for
4B), which closely resembled the distribution assembly of 26S proteasome. C-terminal peptides
6
of the three non-HbYX motif Rpt subunits (Rpt1, cellular 26S proteasome assembly occurs by
Rpt4, and Rpt6), unlike those of subunits binding of intact PA700 to the 20S proteasome.
containing the HbYX motif, do not bind detectably Such a model, however, differs from one
to the 20S proteasome in vitro (20;25). This presented by others, in which 26S proteasome is
feature, however, does not exclude these residues formed by the sequential binding of PA700
from critical roles in cellular assembly of 26S subassemblies to the 20S proteasome (26;34). The
proteasome because their action may require mechanisms of these alternative models should
cooperation with other PA700 components or may result in different fates of Rpt subunits in the
be exerted in assembly processes other than direct absence of 20S proteasome. Thus, if the 20S
binding to 20S proteasome. Nevertheless, as proteasome is a required template for PA700
shown in Figures 5, 6, and 7, FLAG-Rpt1(-C3), assembly, Rpt subunits should accumulate as
FLAG-Rpt4(-C3), and FLAG-Rpt6(-C3) were subassemblies in the absence of 20S proteasome.
assembled into intact 26S proteasome similarly to In contrast, if 20S proteasome is not an obligatory
their wild-type counterparts. Affinity purification template for PA700 assembly, intact PA700
of FLAG-proteins from extracts of cells should accumulate in its absence. To distinguish
expressing mutant Rpt1 and Rpt 4 confirmed the between these models, we reduced 20S
presence of structurally intact and functionally proteasome content in FLAG-Rpt6(WT) cells
active 26S proteasome (Figure 5C-E and Figure using siRNA against the 5 subunit. RNAi D
o
6C-E). In contrast, much reduced active 26S reduced 5 content by more than 90% and wn
lo
proteasome was detected in affinity-purified proteasome activity in cell extracts by ad
e
sRapmt6p le(Fs ifgruorme 7exCt-rFac).t s oInf scteeallds, emxporsets soinf ga fmfiuntiatyn-t agprapdrioexnitm acteenltyr i7fu5g%at i(oFni guorfe e8xAtr)a. cGtsl yfcreormol dcoennstritoyl d from
purified Rpt6(-C3) protein was in the form of and siRNA-treated cells confirmed that RNAi of http
PA700 and variable amounts of dissociated 20S 5 significantly reduced 26S proteasome content ://w
w
proteasome. We interpret these results to indicate and activity, and promoted a relative redistribution w
that FLAG-Rpt6(-C3) is assembled normally into of the FLAG-Rpt6 subunit from fractions .jbc
.o
26S proteasome in cells but destabilizes the characteristic of 26S proteasome to those rg
resulting complex. Thus, the high salt conditions characteristic of PA700 (Figure 8B). These results by/
g
used during affinity purification likely promotes indicate that intact PA700 accumulated in cells ue
s
either loss of 20S proteasome during the binding when the level of 20S proteasome was reduced. To t o
n
and washing phases of the purification or confirm the nature of the FLAG-Rpt6-containing Ja
n
dissociation of any remaining holoenzyme during complexes present in extracts of siRNA-treated ua
ry
and after elution from the affinity matrix. To cells, we purified these complexes on anti-FLAG 5
, 2
illustrate this feature, we repeated the initial beads and analyzed them by native-PAGE and 01
9
glycerol density gradient centrifugation of cell glycerol density gradient centrifugation. FLAG-
extracts in the presence of the same salt Rpt6 purified from control extracts was present
concentrations used during affinity purification. almost exclusively as a subunit of 26S proteasome,
Increased salt concentrations had no effect on the as demonstrated by both methods (Figure 8C and
distribution profile of the FLAG-Rpt6(WT) 8E). In contrast, FLAG-Rpt6 purified from
protein but significantly shifted FLAG-Rpt6(-C3) extracts of siRNA-treated cells was present both in
to a position characteristic of PA700 the remaining but significantly reduced 26S
(Supplemental Figure S2). proteasome, and in PA700 (Figure 8C and 8E).
We also subjected the proteins separated on native
20S proteasome is not required for assembly PAGE to second-dimension SDS-PAGE followed
of PA700. The results presented above indicate by silver staining. The band derived from the
that intact C-termini of certain Rpt subunits are siRNA-treated cells that accumulated in the native
required for assembly of those subunits into 26S gel with a migration position characteristic of
proteasome. In each case, however, the assembly- PA700 featured a second-dimension subunit
defective subunit was present in a complex pattern characteristic of purified PA700 and
structurally indistinguishable from intact PA700. differed from the 26S proteasome only by the
This finding is consistent with a model in which absence of 20S proteasome subunits (Figure 8D).
7
A similar subunit pattern was observed when these proteins as monitors of the assembly process.
siRNA glycerol gradient fractions corresponding Not surprisingly, varying proportions of some
to the position of PA700 were subjected to SDS- exogenously expressed Rpt proteins also
PAGE and silver staining (Supplemental Figure accumulated in complexes smaller than 26S
S3). Interestingly, after siRNA-mediated proteasome to a greater extent than that observed
knockown of β5 subunit, residual 26S proteasome for the corresponding endogenous proteins. This
was enriched in the doubly-capped form, as could reflect accumulation of the FLAG-protein in
detected by both a slower migrating band on normal assembly pathway intermediates that
native-PAGE (Figure 8C) and a faster become stalled due to an excess of that protein
sedimenting species during glycerol density over endogenous interacting partners required for
gradient centrifugation (Figure 8E). This feature continued progress along the assembly pathway.
is a likely consequence of increased binding of In fact, our initial characterization of selected low
accumulated PA700 to both ends of the remaining molecular weight complexes indicates that they
20S proteasomes as the number of total available likely represent authentic assembly intermediates
binding sites on 20S proteasomes declines during (see below).
RNAi. A similar effect has been noted with yeast The C-terminal HbYX motifs of PA700 Rpt
mutants defective in 20S proteasome assembly subunits have been studied previously for their
(35). Finally, we tested whether PA700 that roles in proteasome gating. The structural basis of D
o
accumulated in siRNA-treated cells featured this process has been advanced significantly by wn
lo
PA700 activity, defined as ATP-dependent analogous studies of the homohexameric PAN and a
d
e
aaccttiivvaittyio. n oFf r2a0ctSio pnrso te8a-s1o1m eo-fc atgallyyczeerdo lp edpetindsaistye trhege ulatmorosn, omwehroisce HPbAY2X00 /Brelsmid1u0e s aplrsoot ealsikoemlye d from
gradient centrifugation of affinity-purified proteins provide the main determinants for proteasome http
from control and siRNA-treated cells contained no binding and subsequent complex stability (20- ://w
w
significant proteasome activity themselves in the 23;40;41). However, unlike PAN and w
presence or absence of ATP nor did they stimulate PA200/Blm10, the six Rpt subunits are .jb
c
.o
the activity of exogenous purified 20S proteasome structurally distinct proteins that feature both rg
in the absence of ATP (Figure 8F). However, in HbYX and non-HbYX motifs at their C-termini. by/
g
the presence of ATP, fractions from the siRNA Thus, the relative roles Rpt C-termini among u
e
s
gradients significantly stimulated 20S proteasome individual subunits and between the HbYX and t o
n
activity. Notably, this activity was about 3-fold non-HbYX subunits in various functions have Ja
n
greater in gradient fractions derived from siRNA- been unclear and the source of conflicting results. ua
ry
treated cells compared to those from control cells Our results demonstrate that deletion of the 5
, 2
(Figure 8F). These results confirm the increased HbYX residues from either Rpt3 or Rpt5 was 0
1
9
presence of intact PA700 in cells subjected to sufficient to prevent incorporation of the truncated
RNAi of 20S proteasome. In sum, these results subunit into 26S proteasome and indicate that in
are consistent with a model in which PA700 is each case the contributions of other intact binding
assembled as an intact complex which then binds elements of PA700 were not able to overcome the
to 20S proteasome to form 26S proteasome loss of a single critical binding interaction.
(Figure 9, lower panel, upper pathway). Because assembly of intact PA700 was not
affected by these deletions, it seems unlikely that
DISCUSSION the defects in 26S proteasome assembly were a
consequence of disrupted early steps of PA700
To investigate the roles of the C-termini of assembly. In fact, because of C-terminal
Rpt subunits in cellular 26S proteasome assembly proteolytic processing of a fraction of the FLAG-
we utilized HEK293 cells stably expressing Rpt3(-C3) protein, we could directly discriminate
epitope-tagged forms of wild-type and C- between the effect of deletion of the HbYX
terminally truncated Rpt proteins in a background residues and the effect of more extensive C-
of endogenous wild-type proteins. The successful terminal deletions on PA700 assembly. Thus,
incorporation of FLAG-Rpt(WT) proteins into the unlike the unprocessed Rpt3(-C3), larger C-
26S proteasome validates the strategy of using terminal truncations inhibited PA700 assembly per
8
se, perhaps as a consequence of lost interactions crucial for structural stability of the assembled 26S
between Rpt3 and cognate PA700 assembly proteasome.
chaperones, such as p28 (42). The exact reason Various aspects of the effects of Rpt subunit
for the lack of processing of wild-type Rpt3 C-termini on 26S proteasome assembly reported
compared to mutant Rpt3 is uncertain but seems here are both compatible and in apparent conflict
unlikely to be related to an inherent structural role with analogous but not identically designed studies
of the deleted residues since no processing in yeast. For example, yeast expressing Rpt1,
occurred when the same mutant protein was Rpt2, Rpt3, or Rpt5 lacking a single C-terminal
expressed transiently (25). residue displayed normal or near-normal 26S
In contrast to the essential roles of the HbYX proteasome levels, whereas yeast expressing Rpt4
residues for assembly of Rpt3 and Rpt5 into 26S or Rpt6 with this deletion showed severely
proteasome, deletion of HbYX residues from Rpt2 decreased levels (26). Similarly, intact 26S
had no effect on incorporation of this subunit. This proteasomes were affinity-purified from yeast
effect was clearly evident in the case of genetically strains in which the tyrosine of the HbYX-motifs
truncated Rpt2, but direct comparison to the of Rpt2, Rpt3 or Rpt5 was mutated (20). Other
assembly of FLAG-Rpt2(WT) was complicated by than the obvious differences in experimental
apparent C-terminal processing of the latter details of the studies, these apparent discrepancies
protein. Processing of wild-type Rpt2 appreciably have unclear bases. D
o
decreased the extent and efficiency of Recently, Goldberg and colleagues used wn
lo
incorporation of the protein into 26S proteasome. features of ATP binding by PAN and Rpt subunits a
d
e
Aofs Rwpitt2h ethxec lFusLiAveG o-fR tphte3 (H-Cb3Y)X, C m-toetrimf ipnraolb raebsliyd uaeres tHob YprXo poasned an omn-oHdbelY Xth atr esdiidvuidese s bfeutwnceteionn s2 0oSf d from
involved in protein-protein interactions required in gating and 20S binding, respectively (45). This http
early steps of the assembly pathway. Nevertheless, model posits that the fixed-order arrangement of ://w
w
despite the common in vitro proteasome binding alternating HbYX and non-HbYX Rpt subunits w
capacities of C-terminal peptides of the HbYX-Rpt (Rpt1, Rpt2, Rpt6, Rpt3, Rpt4, and Rpt5) (46) .jb
c
.o
subunits, HbYX residues of Rpt2 are not essential ensures simultaneous interactions of at least one rg
for cellular incorporation of this subunit into the para-positioned pair of gating and binding Rpt by/
g
26S proteasome. 26S proteasome with Rpt2 subunits with the 20S proteasome at all times u
e
s
lacking an HbYX displayed significantly lower during ATP-dependent hydrolysis of proteins. t o
n
peptidase activity than wild-type 26S proteasome, Such an arrangement could reflect the need to J
a
n
demonstrating an important role of this motif in maintain overall structural stability of the 26S ua
ry
proteasome gating. An important role for Rpt2 in proteasome while allowing dynamic 20S-Rpt 5
, 2
proteasome gating was demonstrated previously subunit interactions that likely occur during 0
1
9
using yeast Rpt2 mutants defective in ATP ordered cycles of nucleotide binding, hydrolysis,
binding and hydrolysis (43;44). and release by different subunit pairs. An
No peptide corresponding to the C-terminus of appealing aspect of this model is its explanation
any non-HbYX Rpt subunits binds detectably to for the symmetry mismatch of the opposing
20S proteasome (25). Nevertheless, it seems likely hexameric Rpt subunit and heptameric  subunit
that these residues must interact with proteasome rings. This mismatch may also be related to
 subunits when present in assembled 26S various imaging results that reveal asymmetric
proteasomes. In support of this assertion, we contacts between the rings (47-50). We note that
found that deletion of non-HbYX residues from the important determinants of proteasome
Rpt6 had no effect on cellular proteasome assembly and stability identified here are found in
assembly, but destabilized the structural integrity subunits oriented asymmetrically within the Rpt
of 26S proteasome by promoting dissociation of ring (Figure 9, upper panel).
the holoenzyme into 20S proteasome and PA700 In addition to defining the relative roles
subcomplexes. The former result, but not the of C-termini of Rpt subunits for 26S proteasome
latter, was observed when analogous deletions assembly, our results provide new information
were made in other non-HbYX subunits. Thus, C- about intermediates of the assembly pathway.
terminal non-HbYX residues of Rpt6 appear Each of the two Rpt subunits (Rpt5 and Rpt3) not
9
assembled into 26S protease due C-terminal reconstitution requires intact C-termini of HbYX
truncations was assembled normally into intact residues shown here to mediate cellular 26S
PA700. Furthermore, reduction of cellular 20S proteasome assembly (24;25). Finally, PA700 can
proteasome content by RNAi promoted be reconstituted in vitro from purified
accumulation of PA700, indicating that intact subassemblies that may represent cellular
PA700 was an immediate precursor to the 26S assembly intermediates (56). Reconstitution of
proteasome. Inhibition of proteasome function by PA700 from subassemblies requires neither intact
RNAi of 20S proteasome subunits can potentially C-termini of Rpt subunits nor 20S proteasome, and
cause compensatory upregulated expression of reconstituted PA700, like independently purified
other proteasome subunits (51;52). However, this PA700, can subsequently bind to 20S proteasome
effect was minimal under the conditions employed to form 26S proteasome. Thus, our collective data
here and cannot account for the observed increase support a model in which intact PA700 is formed
in PA700 content. The incorporation of mutant independently of 20S proteasome and is an
Rpt subunits into intact PA700 demonstrated here immediate precursor to 26S proteasome (Figure 9,
indicates that their failure to assemble into 26S lower panel, upper pathway).
proteasome is not a consequence of a defect in an Despite these cellular and biochemical data,
early stage of the assembly process. For example, evidence for an alternative model for PA700 and
previous work has documented the roles of 26S proteasome formation has been presented D
o
specific assembly chaperones in the formation of previously in both yeast and mammalian cells. wn
lo
multiple Rpt subassemblies and in the subsequent These studies show that Rpt-containing a
d
e
cboamseb (i2n7at;2io8n) . o Af ltthheosueg hsu obuars seexmpberliiemse ntots fhoarvme nthoet spurobtaesasseommbel iepsr ioorf toP Afo7r0m0a taiossno ocifa tthee winitthac tt hPeA 2700S0 d from
systematically evaluated the effect of C-terminal and indicate that the 20S proteasome provides a http
mutations on the formation of Rpt subassemblies, template function for PA700 formation (26;34) ://w
w
assembly of these mutant Rpt proteins into intact (Figure 9, lower panel, lower pathway). Moreover, w
PA700 argues that the last three residues are not subassemblies of PA700 accumulate in yeast .jb
c
.o
critical for early stages of the assembly process genetically defective for normal 20S proteasome rg
and instead exert their effects at the level of 20S assembly (35). The basis for discrepancies among by/
g
proteasome binding. Our initial analysis of the these experiments is not readily apparent. u
e
s
low molecular weight complexes observed for However, these alternative assembly models need t o
n
FLAG-Rpt5 and FLAG-Rpt1 indicate that they not be mutually exclusive. Different assembly J
a
n
represent complexes similar to those previously pathways may be preferentially favored by ua
ry
identified as assembly intermediates (24;30;39) different cells or by given cells under different 5
, 2
(Supplemental Figures S4 and S5) . physiologic states. Nevertheless, the data 0
1
9
Our cellular data also support multiple types presented here for mammalian cells under normal
of in vitro results consistent with the view that growth conditions favors a model in which PA700
cellular 26S proteasome is assembled by the can be assembled independently of the 20S
binding of 20S proteasome with intact PA700 via proteasome, and suggest that intact PA700 is a
essential Rpt C-terminal residues. First, intact direct intermediate in the pathway of 26S
PA700 exists in vitro and can be purified from a proteasome assembly (Figure 9, lower panel,
variety of cells and tissues (53). Although it is upper pathway).
possible that free PA700 is derived artifactually by
dissociation of 26S proteasome during the
purification process, free intact PA700 also exists
in crude extracts of mammalian cells lysed under
gentle conditions that favor maintenance of intact
26S proteasome. Second, 26S proteasome can be
reconstituted in vitro from purified PA700 and
20S proteasome, suggesting that this process
mirrors one that occurs in intact cells (53-55).
Significantly, this ATP-dependent in vitro
10