The APC Subunit Doc1 Promotes Recognition of the Substrate Destruction Box
Accurate chromosome segregation during mitosis requires the coordinated destruction of the mitotic regulators securin and cyclins. The anaphase-promoting complex (APC) is a multisubunit ubiquitin-protein ligase that catalyzes the polyubiquitination of these and other proteins and thereby promotes their destruction. How the APC recognizes its substrates is not well understood. In mitosis, the APC activator Cdc20 binds to the APC and is thought to recruit substrates by interacting with a conserved target protein motif called the destruction box. A related protein, called Cdh1, performs a similar function during G1. Recent evidence, however, suggests that the core APC subunit Doc1 also contributes to substrate recognition. To better understand the mechanism by which Doc1 promotes substrate binding to the APC, we generated a series of point mutations in Doc1 and analyzed their effects on the processivity of substrate ubiquitination. Mutations that reduce Doc1 function fall into two classes that define spatially and functionally distinct regions of the protein. One region, which includes the carboxy terminus, anchors Doc1 to the APC but does not influence substrate recognition. The other region, located on the opposite face of Doc1, is required for Doc1 to enhance substrate binding to the APC. Importantly, stimulation of binding by Doc1 also requires that the substrate contain an intact destruction box. Cells carrying DOC1 mutations that eliminate substrate recognition delay in mitosis with high levels of APC substrates. Doc1 contributes to recognition of the substrate destruction box by the APC. This function of Doc1 is necessary for efficient substrate proteolysis in vivo.
Current Biology, Vol. 15, 11–18, January 11, 2005, ©2005 Elsevier Ltd. All rights reserved. DOI 10.1016/j.cub.2004 . 1 2 .066
The APC Subunit Doc1 Promotes
Recognition of the Substrate Destruction Box
activator protein, either Cdc20 or Cdh1 in mitotically
dividing cells [1, 2]. Cdc20 and Cdh1 associate with the
APC in a cell cycle-dependent manner: Cdc20 binds
Christopher W. Carroll, Maria Enquist-Newman,
and David O. Morgan*
Departments of Physiology and Biochemistry
and Biophysics and activates the APC at the metaphase-to-anaphase
transition, whereas Cdh1 maintains activity during lateUniversity of California, San Francisco
San Francisco, California 94143-2200 mitosis and G1 [5, 6]. Cdc20 and Cdh1 are also thought
to interact directly with target substrates [7–10]. In most
but not all cases, the binding of activator to substrate
is reduced or eliminated by mutations in the substrate
D-box or KEN-box, with Cdc20 often displaying specific-
ity for the D-box and Cdh1 displaying specificity for
Background: Accurate chromosome segregation during
the KEN-box. This is consistent with a model in which
mitosis requires the coordinated destruction of the mi-
activator proteins bind substrates through conserved
totic regulators securin and cyclins. The anaphase-pro-
motifs and present them to the APC for ubiquitination.
moting complex (APC) is a multisubunit ubiquitin-protein
In many cases, however, the behavior of a substrate
ligase that catalyzes the polyubiquitination of these and
in vivo cannot be explained simply by its interaction with
other proteins and thereby promotes their destruction.
the activator alone. For example, deletion of the D-box
How the APC recognizes its substrates is not well under-
from the budding yeast cyclin Clb2 or the polo-like ki-
stood. In mitosis, the APC activator Cdc20 binds to the
nase Cdc5 does not abolish binding of these substrates
APC and is thought to recruit substrates by interacting
to Cdh1 (presumably because binding is mediated by
with a conserved target protein motif called the destruc-
the KEN box) but does reduce substrate proteolysis
tion box. A related protein, called Cdh1, performs a
in vivo . Similar observations were made with bud-
similar function during G1. Recent evidence, however,
ding yeast Hsl1 and the human protein Skp1 [7, 11]. One
suggests that the core APC subunit Doc1 also contrib-
explanation for these results is that D-box recognition is
utes to substrate recognition.
not mediated solely by the activator protein. Consistent
Results: To better understand the mechanism by which
with this possibility, a role for the APC itself in substrate
Doc1 promotes substrate binding to the APC, we gener-
recognition has emerged [12–14]. Direct binding has
ated a series of point mutations in Doc1 and analyzed
recently been observed between a D-box peptide and
their effects on the processivity of substrate ubiquitina-
APC isolated from mitotic Xenopus egg extracts .
tion. Mutations that reduce Doc1 function fall into two
This interaction is not dependent on activator, sug-
classes that define spatially and functionally distinct
gesting that core APC subunits play a direct and specific
regions of the protein. One region, which includes the
role in substrate recognition.
carboxy terminus, anchors Doc1 to the APC but does
An intriguing candidate for mediating this interaction
not influence substrate recognition. The other region,
is Doc1/Apc10, the only core APC subunit thus far impli-
located on the opposite face of Doc1, is required for
cated in substrate binding [13, 14]. Biochemical analysis
Doc1 to enhance substrate binding to the APC. Impor-
of budding-yeast APC shows that Doc1 increases the
tantly, stimulation of binding by Doc1 also requires that
processivity of substrate ubiquitination by enhancing
the substrate contain an intact destruction box. Cells
the affinity of the APC-substrate complex . Impor-
carrying DOC1 mutations that eliminate substrate rec-
tantly, the interaction between APC and the activators
ognition delay in mitosis with high levels of APC sub-
Cdh1 and Cdc20 is unaffected by loss of Doc1 function,
suggesting that Doc1 promotes substrate binding di-
Conclusions: Doc1 contributes to recognition of the
rectly or in concert with other core APC subunits. This
substrate destruction box by the APC. This function
possibility is supported by the crystal structures of yeast
of Doc1 is necessary for efficient substrate proteolysis
and human Doc1 [15, 16]. Both structures show that
Doc1 contains a single conserved central domain with
an overall geometry strikingly similar to that of a number
of functionally unrelated proteins. Interestingly, these
proteins, including galactose oxidase and sialidase, use
Substrates of the anaphase-promoting complex (APC)
the same surface to bind diverse ligands, suggesting
[1, 2] generally contain conserved sequence elements
that this interface is also used by Doc1 to bind ligands.
called the destruction box (D-box) and KEN-box [3, 4].
We further explored the role of Doc1 in substrate rec-
Mutation of these elements stabilizes the substrate
ognition by the APC. Our data demonstrate a require-
in vivo and reduces its ubiquitination in vitro, suggesting
ment for the putative ligand binding interface of Doc1
that these motifs are necessary for recognition by the
in D-box-specific substrate binding to the APC. Doc1
mutants that reduce D-box binding to the APC are un-
Substrate ubiquitination by the APC also requires an
able to efficiently target substrates for degradation
in vivo, suggesting that the contribution of Doc1 to
D-box recognition is critical for APC function in the cell.
Results the amount of the C⌬ mutant in these reactions because
our antibody specifically recognizes the C-terminal re-
gion of Doc1. However, a thorough analysis of the bind-Identification of Residues Required
for Doc1 Function ing of these mutants to APC is described below.
The amino acids required for processive substrateThe APC can processively ubiquitinate substrates by
transferring multiple ubiquitin molecules to a target pro- ubiquitination are found in spatially distinct regions of
the Doc1 protein; N238, H239, K243, and D244 are alltein during a single binding event . The degree of
processivity depends on the balance between substrate located on a conserved loop on the putative ligand bind-
ing interface of Doc1, whereas K162, R163, and the Cdissociation and enzyme turnover. The APC subunit
Doc1 affects this balance by limiting substrate dissocia- terminus are located on the opposite face (Figure 1A).
We therefore tested the effects of combining mutationstion [13, 14]. Defects in substrate binding to the APC
because of compromised Doc1 function can therefore that cluster in similar regions of Doc1 (Figures 1B and
1C). Two additional mutants were made: a quadruplebe detected by a reduction in the processivity of sub-
strate ubiquitination. point mutant of N238, H239, K243, and D244 to alanine
(referred to hereafter as 4A) and a combination of theTo better understand how Doc1 contributes to sub-
strate recognition, we used targeted mutagenesis to K162A/R163A mutant and the C⌬ mutant. In both cases,
these mutant combinations were unable to stimulateidentify functionally important regions of Doc1. Se-
quencing efforts have identified clear Doc1 orthologs in processivity above levels displayed by APC lacking
Doc1. As observed with the pairs of point mutations,numerous organisms, as well as domains related to
Doc1 in other putative ubiquitin ligases that are other- the 4A mutations did not affect the binding of Doc1 to
the APC.wise unrelated to the APC [15, 16]. We therefore focused
our analysis on amino acid positions that are highly
conserved in Doc1 orthologs but less well conserved in
Doc1 Mutations that Affect Processivity
related “doc-domains” because these residues are the
Fall into Two Classes
most likely to mediate the APC-specific functions of
The mutations described above define two spatially dis-
Doc1. In all, we changed 11 charged or polar residues,
tinct regions of Doc1. We hypothesized that these re-
either individually or in pairs, to alanine.
gions might affect Doc1 function in different ways. On
Interestingly, the majority of residues selected by
the basis of our initial characterization and the pre-
these criteria map to the putative ligand binding inter-
viously demonstrated interaction between the C termi-
face of the protein, consistent with the idea that this
nus of Doc1 and Cdc27 , the region of Doc1 defined
region is important for Doc1 function (Figure 1A). How-
by mutations in K162, R163, and the C terminus is likely
ever, two of the highly conserved positions, K162 and
to influence substrate ubiquitination through its interac-
R163, map to the opposite face, adjacent to where the
tion with the APC; mutations that inhibit this interaction
C terminus of Doc1 exits the core domain. Previous
would result in reduced binding of Doc1 to the APC
work showed that the C terminus of Doc1 interacts di-
and lead to an apparent reduction in processivity. In
rectly with the APC subunit Cdc27 , but the func-
contrast, mutations in the putative ligand binding region
tional significance of this interaction has yet to be deter-
of Doc1 appear to specifically reduce the ability of Doc1
mined. We therefore also characterized a truncation
to stimulate substrate recognition.
mutant of Doc1 lacking the C-terminal 28 amino acids.
The requirement for these amino acids in Doc1 func-
APC was immunoprecipitated from extracts of yeast
tion was analyzed with purified components in vitro.
cells expressing wild-type or mutant versions of DOC1
Wild-type and mutant versions of Doc1 were expressed
and was tested for its ability to ubiquitinate a radio-
in bacteria as maltose binding protein (MBP) fusion pro-
labeled N-terminal fragment of sea urchin cyclin B (Fig-
teins and purified. Increasing concentrations of recom-
ure 1B). The processivity of the reaction was determined
binant proteins were then added to APC purified from
by quantifying the molar ratio of ubiquitin to cyclin in
doc1⌬ cells (APC
). Wild-type Doc1 was able to fully
the mono-, di-, and tri-ubiquitinated reaction products
restore processivity to APC
with a half-maximal con-
(Figure 1C). As shown previously , APC lacking Doc1
centration of ⵑ40 nM, whereas MBP alone had no effect
(vector) was almost completely nonprocessive.
(Figure 2A). Doc1 proteins with mutations in residues
Two pairs of point mutations, N238A/H239A and
K162 and R163 or lacking the C terminus were also
K243A/D244A, displayed significant defects; in particu-
able to fully restore processivity to APC
, but both
lar, the N238A/H239A mutant was able to stimulate pro-
required increased concentrations to do so. Combining
cessivity only slightly when compared to APC lacking
these mutations had an even greater effect: It increased
Doc1. In addition, point mutations in K162 and R163 or
the half-maximal concentration ⵑ30-fold when com-
deletion of the C terminus (C⌬) of Doc1 resulted in mod-
pared to wild-type Doc1. Nevertheless, at saturating
est but reproducible defects in processivity.
concentrations, even this mutant could stimulate pro-
Western blotting of the Cdc16 subunit in the immuno-
cessivity to a level comparable to that of wild-type Doc1
precipitated APC indicated that equivalent amounts of
(Figure 2A, bottom right). These data confirm that the
enzyme were included in each reaction (Figure 1B). Im-
ubiquitination defects associated with mutations in
portantly, Western blotting of Doc1 showed that the
K162, R163, and the C terminus are exclusively related
N238A/H239A and K243A/D244A mutants bound APC
to APC binding and not to the ability of Doc1 to stimulate
as well as wild-type Doc1. Levels of the K162A/R163A
processive substrate ubiquitination. These residues
mutant appeared similar, though slightly reduced, com-
pared to wild-type Doc1. We were unable to determine therefore define an APC-interaction surface within Doc1.
Destruction Box Recognition by the APC
Figure 1. Doc1 Mutants Are Defective in Processivity
(A) The position of highly conserved residues identified by structure-based sequence alignments is shown on budding-yeast Doc1 [15, 16].
Residues that we show are required for Doc1 function are labeled in red. This image was generated with CHIMERA .
(B) Doc1 mutants have defects in processivity. APC was immunoprecipitated with anti-Cdc26 antibodies from ⵑ1 mg of extract prepared
from strain TC92 (MAT a, bar1, doc1⌬::URA3) expressing wild-type DOC1 (Wt) or various mutants from the DOC1 promoter on a CEN/ARS
plasmid (see Experimental Procedures). Control cells (vector) express no Doc1. Immunoprecipitates were washed three times in lysis buffer
and divided in two. One half (top panel) was used for measurement of cyclin ubiquitination activity and visualized with a PhosphorImager.
The other half (bottom panels) was used for Western blotting of the APC subunits Cdc16 and Doc1.
(C) Quantification of ubiquitination reactions like those in (B). The numbers of mono (cyclin-ub
)-, di (cyclin-ub
)-, and tri (cyclin-ub
species in each lane were quantified, and the processivity of the reaction was determined by calculating the ratio of ubiquitin to cyclin in the
reaction products. Because samples containing fully active Doc1 generate some products above the tri-ubiquitinated species, the data
presented are likely an underestimate of the actual ratio in some cases. Error bars represent the standard deviation from three independent
concentrations, the ratio of ubiquitin to cyclin in the
reaction products was significantly reduced compared
to wild-type Doc1 (Figure 2B). Half-maximal stimulation
occurred at ⵑ70 nM Doc1, a value comparable to that
of wild-type and consistent with the idea that these
residues do not influence association with the APC but
are specifically required to enhance processivity. The
N238A/H239A mutant was a poor activator of APC
making it difficult to accurately determine the concentra-
tion required for half-maximal activation. Nevertheless,
at no concentration tested (up to 2 M) was the N238A/
H239A mutant able to stimulate processivity above lev-
els seen in Figure 1C (data not shown).
The doc1-4A mutant was unable to activate the APC
at any concentration (Figure 2B). To ensure that the
mutant protein does bind APC
in these experiments,
we asked whether doc1-4A could compete with wild-
type Doc1 for APC binding. Indeed, the mutant protein
was able to inhibit the ability of wild-type Doc1 to stimu-
late processivity (Figure 2C). Half-maximal inhibition oc-
curred at ⵑ400 nM doc1-4A, a concentration higher than
might be expected if the mutant protein binds APC as
well as wild-type Doc1. We suspect, however, that our
recombinant doc1-4A protein is not fully active: The
yield of mutant protein after purification is lower than
that of the wild-type protein, suggesting that the 4A
mutations affect the folding or stability of the mutant
protein when it is expressed in bacteria. Importantly, at
saturating concentrations, the doc1-4A mutant was able
to fully inhibit activation of APC
by wild-type Doc1,
confirming that the doc1-4A mutant binds but does not
activate the APC.
Doc1 thus contains at least two functionally distinct
regions. The region associated with the C terminus and
residues K162 and R163 is important for the interaction
of Doc1 with other APC subunits. In contrast, residues
N238, H239, K243 and D244, which map to the putative
ligand binding interface of Doc1, appear to be required
for the ability of Doc1 to stimulate substrate binding to
Figure 2. Functional Analysis of Doc1 Mutations that Affect Pro-
(A and B) Doc1 mutants fall into two classes. Increasing concentra-
tions (up to 12 M) of purified wild-type or Doc1 mutant proteins
were added to ⵑ1 nM APC
in complete ubiquitination reactions.
The Putative Ligand Binding Interface of Doc1
MBP alone was used as a negative control. Reactions were allowed
Is Required for Substrate Recognition
to proceed until ⵑ5% of the total substrate was converted to prod-
To further explore the requirement for the putative ligand
uct, and the processivity of the reaction was quantified as described
binding interface in Doc1 function, we purified APC from
in Figure 1C. Data were fitted to a rectangular hyperbola (solid line).
cells containing wild-type Doc1 or the doc1-4A mutant
Error bars represent the standard deviation from three independent
protein with a tandem affinity purification (TAP)-tagged
experiments. ([A], bottom right) The small graph is a scaled version
of the large graph and includes additional data at high (⬎1 M)
Cdc16 subunit  (Figure 3A). The presence of equal
amounts of Doc1 in the two complexes confirmed that
(C) doc1-4A can bind but not activate the APC. Increasing concen-
the doc1-4A mutation does not affect association with
trations of the doc1-4A mutant protein were added to ⵑ1nM
in the presence of ⵑ10 nM wild-type Doc1. Reactions were
As shown previously , APC
was defective in
analyzed as above. Solid line is the best fit of the data. Error bars
binding the sea urchin cyclin B substrate, resulting in
represent the standard deviation from three independent experi-
ments. The dashed line represents the average processivity of
the slow production of monoubiquitinated reaction
in ubiquitination reactions lacking recombinant Doc1.
products (Figure 3B). Consistent with our immunopre-
cipitation data (Figure 1), the activity of purified APC
was indistinguishable from that of APC
presence of normal levels of doc1-4A protein.In contrast, Doc1 proteins carrying mutations in the
putative ligand binding interface were unable to restore To test the generality of the processivity defects
caused by doc1-4A, we analyzed a variety of well-char-full processivity to the reaction. Mutation of residues
K243 and D244 to alanine resulted in a Doc1 protein that acterized yeast substrates, including Pds1 (securin), the
mitotic cyclin Clb2, and a fragment of Hsl1 (amino acidswas only partially able to activate the APC; at saturating
Destruction Box Recognition by the APC
Figure 3. Doc1-4A Mutants Are Defective in Ubiquitination of Multiple Substrates
(A) The ligand binding region of Doc1 is not required for APC assembly. APC was purified in parallel from strains TC130 (MAT ␣, CDC16::CDC16-
TAP:HIS3, doc1⌬::URA3, trp1::TRP1:DOC1) and TC131 (MAT ␣, CDC16::CDC16-TAP:HIS3, doc1⌬::URA3, trp1::TRP1:doc1-4A) with the TAP
method . Equal volumes of the final eluate (ⵑ20 l) from the wild-type (Wt) or mutant (4A) APC purifications were separated on a 12.5%
polyacrylamide gel, and the proteins were visualized by silver staining. The identity of individual APC subunits is shown on the right; molecular
weight is indicated to the left. The doc1-4A mutant protein migrates slightly faster than wild-type Doc1 under these conditions.
is defective for processive substrate ubiquitination in vitro. Complete ubiquitination reactions containing equal amounts (ⵑ1nM
final concentration) of wild-type APC (Wt), APC
(doc1⌬), or APC
(doc1-4A) were performed in parallel with the indicated substrates;
control reactions (–) have no APC. All substrates were produced in vitro by coupled transcription and translation and are labeled with
S-methionine. Reactions containing sea urchin cyclin B were resolved on a 15% acrylamide gel. All others were resolved on 7.5% acrylamide
667–882). Interestingly, these substrates, particularly substrates still requires intact APC targeting motifs and,
therefore, reflects specific APC activity.
Hsl1, were ubiquitinated much more processively than
We asked whether the D-box or KEN-box is required
sea urchin cyclin B, suggesting that yeast APC may
for Doc1 to promote substrate recognition. As described
bind endogenous substrates with higher affinity than
above, Hsl1 was less processively ubiquitinated by
the model cyclin substrate. The ubiquitination of all sub-
than by the wild-type enzyme (Figure 4A; com-
strates by APC
was significantly reduced when com-
pare lanes 2 and 3). Similarly, ubiquitination of the
pared to wild-type APC. Importantly, the activity of
KEN-box mutant substrate by APC
was identical to that of APC
relative to wild-type APC (lanes 8 and 9). In contrast,
and Clb2, suggesting that the doc1-4A mutant is com-
ubiquitination of the Hsl1 D-box mutant was unaffected
pletely deficient in the recognition of these substrates
by loss of Doc1 activity, suggesting that Doc1 stimulates
(Figure 3B). APC
did have a modest effect on the
processivity only if the substrate contains an intact D-box
ubiquitination of Pds1 when compared to APC
(lanes 5 and 6).
ever, the activity of APC
was still well below levels
We also examined the behavior of Clb2 and Pds1
exhibited by the wild-type enzyme. Similar results were
(Figures 4B and 4C). Although Pds1 and Clb2 substrates
obtained with all substrates when Cdc20 was used as
with mutations in the D-box or KEN-box were ubiquiti-
the APC activator (data not shown). The APC therefore
nated inefficiently, we found that the ability of Doc1 to
appears to recognize all substrates in a manner that
stimulate ubiquitination was absolutely dependent upon
depends on a conserved loop on the ligand binding
the substrate having an intact D-box; Pds1 and Clb2
interface of Doc1.
that rely only on a KEN-box for recognition were not
influenced by Doc1 activity. Doc1 therefore enhances
Doc1 Mediates Substrate Recognition
processivity by a mechanism that requires a D-box but
through the Destruction Box
not a KEN-box.
Ubiquitination of Hsl1 by wild-type APC was reduced
Our data also indicate that Doc1 is not solely responsi-
by introducing point mutations in the substrate D-box
ble for recognition of the D-box in these experiments. In
or KEN-box (Figure 4A; compare lane 2 with lanes 5
all cases, APC
-dependent ubiquitination of a wild-
and 8), with the D-box making a greater contribution to
type substrate was further reduced by introducing point
ubiquitination than the KEN-box. Importantly, ubiquiti-
mutations in the substrate D-box (compare, for example,
nation of Hsl1 containing mutations in both the D-box
lane 3 and lane 6 in each panel), suggesting that the
and KEN-box was further reduced (lanes 5, 8, and 11),
D-box promotes substrate ubiquitination through multi-
ple mechanisms, perhaps involving Cdh1.indicating that the ubiquitination of the single mutant
Figure 4. The Substrate D-box Is Required for Recognition by Doc1
Hsl1 (A), Pds1 (B), and Clb2 (C), each containing intact targeting motifs or mutations in the D-box, KEN-box, or both, were included in complete
ubiquitination reactions containing no APC (–), wild-type APC (Wt), or APC
(4A). Reactions were performed as in Figure 3, and identical
results were obtained in three independent experiments. (⫹) represents wild-type sequence; (–) represents a mutation. D-box mutations are
(RxxLxxxxN → AxxAxxxxA, except Pds1, which is RxxL → AxxA); KEN-box mutations are (KEN → AAA for Clb2 and Pds1, KENxxxE →
AAAxxxA for Hsl1).
The Putative Ligand Binding Interface of Doc1 subunits in this process. Specifically, we identified a
conserved loop, on the putative ligand binding interfaceIs Required for Efficient Proteolysis
of APC Substrates In Vivo of Doc1, that is required for D-box-dependent substrate
binding to the APC and for efficient proteolysis of APCTo determine the specific requirement for the D-box
recognition function of Doc1 in controlling mitotic pro- substrates in vivo.
The essential function of the D-box and KEN-box ingression, we compared doc1⌬ cells expressing wild-
type Doc1 or the doc1-4A mutant through a single, syn- targeting APC substrates for proteolysis is well estab-
lished [3, 4]. How these motifs are recognized by the APCchronous cell cycle. When released from a G1 arrest,
wild-type and doc1-4A cells both initiated the cell cycle is less clear. Abundant evidence from diverse systems
suggests that the activators Cdc20 and Cdh1 contributeand entered mitosis with similar kinetics, as suggested
by bud emergence and mitotic spindle assembly (Figure to substrate recognition and specificity [7–10], and the
direct binding of Cdc20 to numerous substrates is D-box5A; data not shown). However, whereas the wild-type
cells rapidly progressed through mitosis and completed dependent. In addition, a direct interaction between pu-
rified mitotic Xenopus APC and a D-box peptide hascytokinesis about 2 hr after release, the doc1-4A mutant
population was largely unable to initiate sister chromatid recently been demonstrated , consistent with the
idea that core APC subunits also contribute to substrateseparation or exit mitosis (Figures 5A and 5B). Four
hours after release from the ␣ factor arrest, ⵑ70% of the recognition.
Our data now demonstrate a function for a specificdoc1-4A cells remained large budded with unseparated
nuclei, indicative of a delay at the metaphase-to-ana- APC core subunit in D-box recognition. The mechanism
by which Doc1 promotes D-box binding remains un-phase transition.
Western-blotting of the APC substrates Pds1, Clb2, clear. Because Doc1 does not affect the affinity of acti-
vator binding to the APC [13, 14], it is unlikely that Doc1and Ase1 revealed that these proteins were stabilized
dramatically in cells that contain the doc1-4A mutation influences substrate binding through some effect on the
(Figure 5C), consistent with a severe defect in APC activ-
conformation of activators or their binding to the APC.
ity. The inability of the doc1-4A mutant APC to efficiently
A more likely possibility is that Doc1 acts by enhancing
target Pds1 for degradation is likely the cause of the
substrate binding to the APC core.
extended metaphase delay. Doc1-dependent recogni-
The simplest explanation for our results is that the
tion of the substrate D-box is therefore critical for APC
putative ligand binding interface of Doc1 interacts di-
activity and mitotic progression in vivo.
rectly with the substrate D-box. However, using a variety
of approaches—from large-scale immunoprecipitation
to more sophisticated fluorescence and crosslinking
techniques—we failed to detect binding between puri-
fied Doc1 and APC substrates. In a heteronuclear singleOur results further our understanding of substrate rec-
ognition by the APC and support a role for core APC quantum correlation (HSQC) NMR experiment, for exam-
Destruction Box Recognition by the APC
Figure 5. Doc1 Is Required for APC Activity
(A and B) The metaphase-to-anaphase transi-
tion is delayed in doc1-4A cells. Strains
TC136 (MAT a, bar1, doc1⌬::URA3, leu2::
LEU2:DOC1, PDS1::PDS1-13myc:HIS3) and
TC137 (MAT a, bar1, doc1⌬::URA3, leu2::
LEU2:doc1-4A, PDS1::PDS1-13myc:HIS3) were
arrested with ␣ factor (1 g/ml) and released.
Samples were taken every half hour and ana-
lyzed for cell cycle progression. Budding in-
dex (A) and DNA segregation (B) in wild-type
(Wt)ordoc1-4A mutant cells were quantified
as a function of time after release from G1
arrest. Data are from a representative experi-
ment that was repeated three times. Over 100
cells were counted for each time point.
(C) APC substrates are stabilized in doc1-4A
cells. 10 g of cell lysate from each time point
was resolved on a 10% acrylamide gel and
transferred to nitrocellulose for Western blot-
ting of the indicated APC substrate.
ple, 250 M Doc1 was incubated in the presence of 500 manner, such that the function of the activator is to load
substrates onto the APC for ubiquitination. Alternatively,M sea urchin cyclin B, and no interaction was detected
(X. Luo and H. Yu, personal communication). Thus, if the interaction of substrates with the APC may involve
multiple low-affinity binding sites that display somewhatDoc1 does bind the substrate D-box directly, the inter-
action is of extremely low affinity. overlapping specificity.
A more likely scenario is that Doc1 alone is not suffi-
cient to mediate target substrate binding. Rather, we
suspect that Doc1 requires additional core APC subunits
Strains and Plasmids
for its activity. For example, Doc1 may form part of a
All yeast strains used in this study are derivatives of W303. Strains
larger binding site that is only present in the context
used in each experiment are described in the corresponding figure
of the intact enzyme. Alternatively, Doc1 may control
legends. For site-directed mutagenesis, full-length Doc1 or Doc1
substrate binding indirectly, perhaps by driving APC
lacking the C terminus (amino acids 1–255) plus 500 bp upstream
of the translation start site were cloned into pRS314 (CEN/ARS,
structural changes that expose a cryptic D-box binding
TRP1). The Quik-change method (Stratagene, La Jolla, CA) was then
site. To begin to address these issues, we engineered
used to generate the indicated point mutations. All mutants were
a version of Doc1 that contained a radio-labeled, photo-
sequenced to confirm that the desired mutation was the only one
activatable crosslinker on a defined position in the ligand
present. In some cases, wild-type Doc1 or mutants were subcloned
binding interface. The protein was then reconstituted
into pRS304 (TRP1) or pRS305 (LEU2) for integration.
, and the crosslinker was activated by pho-
tolysis. We were unable, however, to detect the ligand
for Doc1, be it a target substrate or additional APC
Complete ubiquitination assays contained E1, E2, ubiquitin, ATP,
Cdh1, and the indicated APC and substrate. All ubiquitination assays
subunits, with this method (data not shown). In addition,
were performed as previously described . For immunoprecipita-
experiments similar to those performed by Yamano et.
tion activity assays in Figure 1, cell lysates were prepared from the
al.  did not reveal an interaction between a fragment
indicated strains by bead beating two times for 20 s in lysis buffer
of Hsl1 immobilized on beads and purified budding-
(20 mM HEPES [pH 7.4], 150 mM NaCl, 5 mM EDTA, 5 mM EGTA,
yeast APC (data not shown). Nevertheless, we believe
and 0.1% NP-40). APC was purified by the TAP method, as de-
that approaches like these, which examine the interac-
For Figure 1,
I-labeled sea urchin cyclin was prepared essen-
tion of substrates with intact APC rather than individual
tially as described . In all other experiments, substrates were
subunits, will ultimately yield the best understanding of
produced by coupled in vitro transcription and translation in rabbit
substrate recognition by the APC.
reticulocyte extracts, as described by the manufacturer (Promega,
Taken together, the data suggest that the interaction
Madison, WI). Pds1 and Clb2 are full-length proteins, sea urchin
of substrates with the APC is likely to be complex and
cyclin B is amino acids 13–110, and Hsl1 is amino acids 667–872.
highly dynamic. One possibility is that substrate associ-
All substrates were initially produced with a C-terminal ZZ tag that
contains a tobacco etch virus (TEV) cleavage site between the sub-
ation with the activator and APC occurs in a stepwise
strate and tag. This facilitates a simple and rapid one-step purifica- 5. Fang, G., Yu, H., and Kirschner, M.W. (1998). Direct binding of
CDC20 protein family members activates the anaphase-pro-tion of substrates away from contaminating activities in the reticulo-
cyte lysate. For purification of substrates, one to five reactions were moting complex in mitosis and G1. Mol. Cell 2, 163–171.
6. Zachariae, W., Schwab, M., Nasmyth, K., and Seufert, W. (1998).carried out in parallel according to the manufacturer’s instructions
and pooled. Three volumes of IgG binding buffer (20 mM HEPES Control of cyclin ubiquitination by CDK-regulated binding of
Hct1 to the Anaphase Promoting Complex. Science 282, 1721–[pH 7.4], 150 mM NaCl, and 0.1% NP-40) was added, and this mixture
was incubated with 10–50 l IgG sepharose (Amersham, Uppsala, 1724.
7. Burton, J.L., and Solomon, M.J. (2001). D box and KEN boxSweden) at 4⬚C for 1 hr. The beads were then washed extensively
and cleaved with TEV protease for 10 min at room temperature. motifs in budding yeast Hsl1p are required for APC-mediated
degradation and direct binding to Cdc20p and Cdh1p. GenesSubstrates with D-box or KEN-box mutations were amplified by PCR
and cloned into the pME34 vector for expression in the reticulocyte Dev. 15, 2381–2395.
8. Hilioti, Z., Chung, Y.S., Mochizuki, Y., Hardy, C.F., and Cohen-extracts. Plasmids bearing mutations in the Hsl1 D-box and
KEN-box were a gift from J. Burton and M. Solomon (Yale University, Fix, O. (2001). The anaphase inhibitor Pds1 binds to the APC/
C-associated protein Cdc20 in a destruction box-dependentNew Haven, CT). Plasmids bearing mutations in the Clb2 and Pds1
D-box and KEN-box were a gift from L. Passmore and D. Barford manner. Curr. Biol. 11, 1347–1352.
9. Pfleger, C.M., Lee, E., and Kirschner, M.W. (2001). Substrate(Cancer Research UK, London).
recognition by the Cdc20 and Cdh1 components of the ana-
phase-promoting complex. Genes Dev. 15, 2396–2407.
Expression and Purification of MBP-Doc1
10. Schwab, M., Neutzner, M., Mocker, D., and Seufert, W. (2001).
Wild-type or mutant DOC1 were cloned into pMal-C2 with PCR
Yeast Hct1 recognizes the mitotic cyclin Clb2 and other sub-
products generated from the Doc1 mutant plasmids described
strates of the ubiquitin ligase APC. EMBO J. 20, 5165–5175.
above. Plasmid isolates used to express protein were sequenced
11. Bashir, T., Dorrello, N.V., Amador, V., Guardavaccaro, D., and
prior to protein expression to ensure that the desired mutations
Pagano, M. (2004). Control of the SCF(Skp2-Cks1) ubiquitin
were the only mutations present. Two liters of BL21 bacterial cells
ligase by the APC/C(Cdh1) ubiquitin ligase. Nature 428,
harboring the expression plasmids were grown to mid-log phase at
37⬚C before the temperature was shifted to 18⬚C for 1 hr. Protein
12. Yamano, H., Gannon, J., Mahbubani, H., and Hunt, T. (2004).
expression was then induced by addition of 0.3 mM IPTG, and cell
Cell cycle-regulated recognition of the destruction box of cyclin
pellets were harvested after 16 hr. MBP-Doc1 or the respective
B by the APC/C in Xenopus egg extracts. Mol. Cell 13, 137–147.
mutants were then purified with amylose resin (New England Bio-
13. Carroll, C.W., and Morgan, D.O. (2002). The Doc1 subunit is a
labs, Beverly, MA).
processivity factor for the anaphase-promoting complex. Nat.
Cell Biol. 4, 880–887.
Cell Cycle Analysis
14. Passmore, L.A., McCormack, E.A., Au, S.W., Paul, A., Willison,
Standard yeast techniques were used . Briefly, strains TC136
K.R., Harper, J.W., and Barford, D. (2003). Doc1 mediates the
and 137 (see Figure 5 legend) were arrested in G1 by addition of ␣
activity of the anaphase-promoting complex by contributing to
factor (1 g/ml) for 4 hr at 25⬚C. ␣ factor was then rapidly washed
substrate recognition. EMBO J. 22, 786–796.
out and samples were taken every 30 min for 4 hr. ␣ factor was
15. Au, S.W., Leng, X., Harper, J.W., and Barford, D. (2002). Implica-
added back after 90% of the cells had budded. DNA was visualized
tions for the ubiquitination reaction of the anaphase-promoting
by DAPI staining. Tubulin was visualized by indirect immunofluores-
complex from the crystal structure of the Doc1/Apc10 subunit.
cence with antibody YOL1/34. Protein extracts were prepared, and
J. Mol. Biol. 316, 955–968.
Western blotting was performed as described . Pds1-13myc
16. Wendt, K.S., Vodermaier, H.C., Jacob, U., Gieffers, C., Gmachl,
was detected with a c-Myc (A-14) polyclonal antibody (Santa Cruz
M., Peters, J.M., Huber, R., and Sondermann, P. (2001). Crystal
Biotech, Santa Cruz, CA) or a 9E10 monoclonal antibody. Clb2 and
structure of the APC10/DOC1 subunit of the human anaphase-
Ase1 were detected with a polyclonal antibody against the endoge-
promoting complex. Nat. Struct. Biol. 8, 784–788.
nous protein. Ase1 antibodies were a generous gift from D. Toczyski
17. Rigaut, G., Shevchenko, A., Rutz, B., Wilm, M., Mann, M., and
(University of California, San Francisco, CA).
Seraphin, B. (1999). A generic protein purification method for
protein complex characterization and proteome exploration.
Nat. Biotechnol. 17, 1030–1032.
18. Guthrie, C., and Fink, G.R., eds. (1991). Guide to Yeast Genetics
We thank David Barford, Janet Burton, Lori Passmore, Mark Solo-
and Molecular Biology, Volume 194 (San Diego: Academic
mon, Brian Thornton, and David Toczyski for reagents, X. Luo and
Hongtao Yu for sharing unpublished data, and Mary Matyskiela, Matt
19. Jaspersen, S.L., Charles, J.F., and Morgan, D.O. (1999). Inhibi-
Sullivan, and members of the Morgan laboratory for discussions and
tory phosphorylation of the APC regulator Hct1 is controlled by
comments on the manuscript. This work was supported by funding
the kinase Cdc28 and the phosphatase Cdc14. Curr. Biol. 9,
from the National Institute of General Medical Sciences (GM53270)
and by a predoctoral fellowship from the National Science Founda-
20. Pettersen, E.F., Goddard, T.D., Huang, C.C., Couch, G.S.,
tion (to M.E.N.).
Greenblatt, D.M., Meng, E.C., and Ferrin, T.E. (2004). UCSF Chi-
mera—a visualization system for exploratory research and anal-
Received: October 11, 2004
ysis. J. Comput. Chem. 25, 1605–1612.
Revised: November 1, 2004
Accepted: November 1, 2004
Published: January 11, 2005
1. Harper, J.W., Burton, J.L., and Solomon, M.J. (2002). The ana-
phase-promoting complex: It’s not just for mitosis any more.
Genes Dev. 16, 2179–2206.
2. Peters, J.M. (2002). The anaphase-promoting complex: Proteol-
ysis in mitosis and beyond. Mol. Cell 9, 931–943.
3. Glotzer, M., Murray, A.W., and Kirschner, M.W. (1991). Cyclin
is degraded by the ubiquitin pathway. Nature 349, 132–138.
4. Pfleger, C.M., and Kirschner, M.W. (2000). The KEN box: An
APC recognition signal distinct from the D box targeted by
Cdh1. Genes Dev. 14, 655–665.