Rab14-RCP-JBC-AL & MMC & AK-29May2015

36 Pages·2015·5.68 MB·English

by  Amir Khan

JBC Papers in Press. Published on May 31, 2015 as Manuscript M114.612366
The latest version is at http://www.jbc.org/cgi/doi/10.1074/jbc.M114.612366
Structure of Rab14:RCP

Structure-Function Analyses of the Interactions Between Rab11 and Rab14
Small GTPases with their Shared Effector Rab Coupling Protein (RCP)

Patrick Lall1†, Andrew Lindsay2†, Sara Hanscom2, Tea Kecman1, Elizabeth S.
Taglauer1, Una M. McVeigh1, Edward Franklin1, Mary W. McCaffrey2, and Amir
R. Khan1

†These authors contributed equally to this work.

1School of Biochemistry and Immunology, Trinity College, Dublin 2, Ireland

2Molecular Cell Biology Laboratory, School of Biochemistry and Cell Biology,
D
o
Biosciences Institute, University College Cork, Cork, Ireland w
n
lo
a
d
e
Correspondence should be addressed to either: Mary W. McCaffrey or Amir R. Khan d fro
m
h
*Running title: Structure of Rab14 bound to RCP ttp
://w
w
Keywords: Rab14 GTPase, endosomal trafficking; Rab-Coupling Protein, X-ray w
.jb
crystallography c.o
rg
________________________________________________________________________ b/
y
g
Background: Rab14 regulates endosomal and both subclasses contain a highly conserved ue
s
trafficking, however, its binding to Rab11-FIP Rab-binding domain (RBD) at their C-termini. t o
n
effectors is a source of conflicting data. Yeast two-hybrid and biochemical studies have D
e
c
Results: Rab14 interacts with the canonical Rab- revealed that the more distantly related Rab14 em
b
binding domain of RCP and both Rab14 and RCP also interacts with class I FIPs. Here, we er 1
function in neuritogenesis. perform detailed structural, thermodynamic 3, 2
Conclusions: Rab11:RCP complex formation may and cellular analyses of the interactions 01
8
precede recruitment of RCP by Rab14. between Rab14 and one of the class I FIPs – the
Significance: This study provides a conceptual Rab Coupling Protein (RCP) - to clarify the
framework for Rab14 and RCP in health and molecular aspects of the interaction. We find
disease. that Rab14 indeed binds to RCP, albeit with
reduced affinity relative to conventional
ABSTRACT Rab11:FIP and Rab25:FIP complexes.
However, in vivo, Rab11 recruits RCP onto
Rab GTPases recruit effector proteins, via their biological membranes. Furthermore,
GTP-dependent switch regions, to distinct sub- biophysical analyses reveal a non-canonical 1:2
cellular compartments. Rab11 and Rab25 are stoichiometry between Rab14:RCP in dilute
closely related small GTPases that bind to solutions, in contrast to Rab11/25 complexes.
common effectors termed the Rab11-Family of The structure of Rab14:RCP reveals that
Interacting Proteins (FIPs). The FIPs are Rab14 interacts with the canonical RBD, and
organized into two subclasses (class I and class also provides insight into the unusual
II) based on sequence and domain organization, properties of the complex. Finally, we show that
1
Copyright 2015 by The American Society for Biochemistry and Molecular Biology, Inc.
Structure of Rab14:RCP
both the Rab Coupling Protein and Rab14 mammalian Rabs range from 30% to 80% (8),
function in neuritogenesis. with the nucleotide-proximal motifs (switch I/II,
P-loop) bearing the most highly conserved (60-
80%) segments.
Introduction The complexity of Rab–regulated
Rab GTPases are molecular switches that intracellular trafficking is increased by the
belong to the Ras superfamily of small GTPases promiscuity of Rab:effector interactions in
and regulate vesicle trafficking in eukaryotic cells eukaryotic cells. Typically a single Rab binds to
(1). In their active GTP-bound state, membrane- multiple, often unrelated, effectors while a single
localised Rabs exert their biological effects by effector protein can sometimes be recognized by
recruiting cytosolic effector proteins to distinct multiple Rab proteins. For example, Rab6
sub-cellular compartments. Membrane attachment regulates Golgi traffic and interacts with effectors
of active Rabs is facilitated by prenylation of one such as the golgin GCC185 and Rab6-interacting
or two C-terminally situated cysteine residues by protein 1 (R6IP1), which are unrelated proteins
the enzyme geranylgeranyltransferase [GGTase; (9,10). The structures of complexes of
(2)], thus adding a hydrophobic tail to the G- Rab6:GCC185 and Rab6:R6IP1 have revealed that
protein. This tail inserts into the lipid bilayer and GCC185 is a dimeric coiled-coil that binds two
D
the globular domain projects into the cytosol. Rabs, while R6IP1 is a monomeric bundle of o
w
n
Upon hydrolysis of GTP, which is stimulated by seven α-helices with a single interface for Rab6. lo
a
GTPase activating proteins [GAPs; (3)], intact However, the Rab6 epitope is formed along two de
d
Rabs - together with their hydrophobic tails - are parallel α-helices in both complexes, which have fro
m
extracted from membranes by GTP/GDP similar hydrophobic and hydrophilic features at h
fdriascstoicoina tio(4n) . in hTibhiteo r sw(GitDchI ) isin totu rntheed c‘yotno’s obliyc tehfefeirc torer sppreoctteivine ciannte rsfoamceest.i mCeos nbvee rsreeclyo,g nai zseidn gblye ttp://ww
w
GTP/GDP exchange factors (GEFs) which multiple Rabs. An example of effector promiscuity .jb
catalyze the exchange of GDP for GTP. The is Rabenosyn-5, which is recruited by Rab4 and c.o
functional cycle is completed by re-insertion of Rab22 via distinct segments of coiled coil motifs brg/
y
Rab tails into membranes. (11). In all known complexes binding is at least g
u
Mammalian Rabs comprise the largest partly facilitated by switch I and switch II. est o
family of Ras-like GTPases, with over 60 Comparative analyses of the known structures of n D
members in humans. Rabs regulate intracellular Rab:effector complexes have revealed that ec
e
trafficking in their active GTP-bound state by specificity is a complex phenotype that is mb
e
recruiting one or more members of a diverse set of influenced by subtle differences in sequence and r 1
3
effector proteins. Rab structures consist of a 6- conformational diversity in switch I, switch II and , 2
0
1
stranded mixed β-sheet flanked by five α-helices. the interswitch region. 8
The 5’ phosphate arm of the nucleotide is bordered In vitro thermodynamic and kinetic
on one side by the P-loop (Walker A motif), which studies of Rab:effector complexes reveal a variety
is conserved throughout the Ras superfamily, as of affinities in vitro ranging from relatively high
well as in many ATPases (5,6). The [Rab27:Slp2a, dissociation constant K = 13nM;
d
conformations of a pair of adjacent regions termed (12)] to low/weak [Rab8:OCRL, K =5µM; (13)].
d
switch I and switch II are sensitive to the presence Despite numerous structural and thermodynamic
of GTP vs. GDP (7), and thus convey the studies, it is generally difficult to discern why
nucleotide specificity for effector recruitment. some complexes have higher affinity than others,
Additionally, a loop and strand situated between or whether the differences in affinity are
the switch motifs - termed the interswitch - also meaningful. A closely related issue is the question
plays a role in effector recruitment. In particular, of the molecular determinants of Rab:effector
an invariant tryptophan residue that is located specificity. Relating the structural, thermodynamic
within an otherwise variable interswitch sequence and biological properties of Rab:effector
provides a hydrophobic surface for effector complexes is essential for a comprehensive
binding. The overall sequence identities between
2
Structure of Rab14:RCP
understanding of how Rabs co-operate to regulate of the antibodies used by Qi et al. and ourselves
vesicle trafficking. (24). Secondly, we report here detailed structural,
The Rab11 subfamily comprises Rab11A, biophysical and cellular analyses of Rab14 binding
Rab11B and Rab25 (also known as Rab11C). to the class I FIPs. We observe structurally that
Rab11A and Rab11B are ubiquitously expressed Rab14 does indeed bind to the conserved RBD of
while Rab25 expression is restricted to epithelial the Rab Coupling Protein (RCP), albeit as a non-
tissues (14). Rab25 activity has been linked to a canonical Rab:FIP (1:2) assembly in dilute
variety of cancers (15,16). concentrations. Our crystallographic studies
Rab14 is more distantly related to the provide insight into the molecular basis for
Rab11 subfamily, although it also appears to Rab14:RCP binding and clarify several
regulate overlapping endocytic pathways via discrepancies involving Rab11/25 and Rab14
interactions with class I FIPs (17-19). Rab14 also interactions with the FIPs. Our cell biology studies
binds to RUFY1/Rabip4, which is a dual suggest that Rab11 is the major Rab that recruits
Rab4/Rab14 effector that regulates endosomal RCP onto biological membranes, while Rab14 and
trafficking (20). The discrimination by Rab14 of a RCP function in the process of neurite formation
sub-set of FIPs is a unique attribute - Rab11 and in a neuronal cell line.
Rab25 recognize both class I and class II FIPs
D
(21). o
w
n
A high throughput yeast two-hybrid EXPERIMENTAL PROCEDURES lo
a
screen first reported an interaction between Rab14 Mutagenesis and sub-cloning de
d
and FIP2 (19). We subsequently analyzed the Rab14 was expressed by subcloning of the fro
m
ability of all five FIPs to interact with Rab14 and globular region (2-175) into pTrcHisA. PCR h
reported that the class I FIPs (RCP, FIP2 and primers were 5'-AACTCGAGGCAACTG- ttp
Rip11), but not the class II FIPs (FIP3 and FIP4), CACCATACAACTAC (forward) and 5'- ://ww
w
interact with Rab14. We identified the region of AAGAATTCGTAGTTCTGATAGATTTTCTTG .jb
RCP, FIP2 and Rip11 that binds to Rab14 as their GCAG (reverse). The template pEGFP-C1/Rab14 c.o
C-terminus, and in the case of RCP we mapped it harboured the mutation Q70L, the cDNA fragment brg/
y
to between residues 579 – 645 (17). This region was cloned into the XhoI/EcoRI sites of g
u
contains the classical Rab11 binding domain pTrcHisA. A further truncation of Rab14 (7-175) es
t o
(RBD) - an approximately 20 amino acid highly was generated by inserting the corresponding n
D
conserved motif located at the carboxy-terminus of cDNA into pNIC28-Bsa4 (Kan+, N-terminal His ec
e
all the Rab11-FIPs. A single amino acid mutation tag) using a ligation-independent cloning protocol. m
b
e
(I621E) in the RBD abolished the ability of RCP The template for this latter construct was r 1
3
to interact with both Rab14 and Rab11. pTrcHisA-Rab14, and this shorter variant was , 2
0
In contrast to our findings, Jing et al. used for crystallization studies. The primers were 18
subsequently reported that Rab14 does not bind to 5'- TACTTCCAATCCATGaacta-ctcttacatc
any of the FIPs (22). However, more recently, a (forward) and 5'- TATCCACCTTT-
study from the Goldenring and Spearman groups ACTGTTAgttcagatagattttc, which generated the
reported that Rab14 does indeed bind to RCP, but region corresponding to Asn7-Asn175 of Rab14.
via a region upstream of the RBD (23). Following rTEV cleavage, the cloning vector
Furthermore, in this paper Qi et al. did not observe introduced non-native Gly-Met residues at the N-
any interaction between Rab14 and the other class terminus. Similarly, RCP (residues 581-649) were
I FIPs. They identified two serines (S580 and inserted into vector pNIC28-Bsa4 using primers
S582), which are not present in FIP2 and Rip11, 5'- TACTTCCAATCCATGccctcggacc-ctgca
that when mutated to asparagine and leucine, (forward) and 5'-TATCCACCTTTACTG-
respectively, abolished the interaction with Rab14 TTAcatctttcctgcttttttg (reverse).
without affecting the interaction with Rab11. pTrcHisC/RCP was generated by
385-649
In order to resolve these conflicting PCR using the following primers 5’-
reports, we have followed two approaches. Firstly, CCCGGATCCAAAGGCAGCTCTCCGAATCTT
we have recently reported a detailed comparison CC-3’ (forward) and 5’- CCCGAATTCTT-
3
Structure of Rab14:RCP
ACATCTTTCCTGCTTTTTTGCC-3’ (reverse) GGGACAAGGAGGCA-3’ (forward) and 5’-
with pEGFP-C3/RCP as template. The PCR AAAAGAATTCCTACTTGACCTCCAGGATG-
WT
product was purified and cloned into the G-3’ (reverse) with pEGFP-C1/FIP3WT as
BamHI/EcoRI sites of pTrcHisC (Invitrogen). template. The PCR product was purified and
The NF (S580N/S582F), NL cloned into the BamHI/EcoRI sites of pTrcHisB
(S580N/S582L) and ΔRBD (L591/Stop) mutants (Invitrogen).
were generated by QuikChange (Agilent
Technologies) site-directed mutagenesis. The Protein Expression
following sense/antisense primers were used: NF, The following protocol was utilized for
5’-ATGAGGTCATGATGAAGAAATACAAC- expression of all His-tagged proteins. Expression
CCCTTCGACCCTGCATTTGCAT-3’/5’-ATG- was carried out in 2xYT Broth supplemented with
CAAATGCAGGGTCGAAGGGGTTGTATTTCT 34 µg/ml kanamycin (FORMEDIUM™) at 37oC.
TCATCATGACCTCAT-3’; NL, 5’-TGAG- At an OD of 0.7 the culture was induced with
600
GTCATGATGAAGAAATACAACCCCTTGGA 0.5 mM IPTG (FORMEDIUM™), after which
CCCTGCATT-3’/5’-AATGCAGGGTCCAA- cells were grown for a further 3 hours at 37oC.
GGGGTTGTATTTCTTCATCATGACCTCA-3’; Cells were harvested by centrifugation and the
and ΔRBD: 5’- CCCTGCATTTGCATAT- pellets were resuspended in His-tag extraction
D
GCGCAGTAGTAGCACGATGAGCTGATTCA buffer (10 mM Tris-HCl, 300 mM NaCl, 5 mM ow
n
GCTGG-3’/ 5’-CCAGCTGAATCAGCTCAT- MgCl2, 20 mM imidazole and 10mM β- loa
d
CGTGCTACTACTGCGCATATGCAAATGCAG mercaptoethanol, pH 8.0) along with 0.5 mM e
d
GG-3’. pTrcHisC/RCP385-649 and pEGFP-C3 PMSF protease inhibitor (Sigma). Cells were lysed from
R CPWT wAe res eursieeds aos ft emcopnlasttreusc. t s comprising the b1y5, 0so0n0i cxat iogn faonrd 2th5e cmeilnl ulytesas tea wt a4so Cce nttor ifruegmedo vaet http://w
region 559-649 of RCP were generated by PCR cellular debris. The supernatants were filtered and w
w
amplication using the following oligos: 5'- loaded onto a nickel agarose resin (Qiagen). The .jb
c
TACTTCCAATCCATGAGCTTGGGAAC- resin was washed with a 10-fold excess of .o
rg
TGCC (forward primer), and 5'- extraction buffer prior to elution of the bound b/
y
TATCCACCTTTACTGTTACATCTTTCCTGCT proteins using extraction buffer supplemented with g
u
TTTTTGC (reverse primer). The underlined 200 mM imidazole. Overnight incubation at 4 oC est o
codons indicate the start and end of translated with recombinant TEV protease (rTEV) was used n D
products. The template for these reactions were to remove the N-terminal hexahistidine tag from ec
e
m
RCP constructs (residues 385-649) in pTrcHis each protein and the proteins were run through a b
e
vector which already contained the corresponding second Ni2+-agarose column. The ‘flow through’ r 1
3
background (WT, S580N/S582F, and fractions were collected, while the uncut proteins , 2
0
1
S580N/S582L). These WT and mutant proteins and rTEV (which has an uncleavable His-tag) 8
were highly soluble and also encompassed the remained bound to the resin. Soluble aggregates
region around Ser580/Ser582, which has recently were removed by running the eluted protein
been implicated in binding to Rab14. The PCR- through a Superdex 200 (16/60) gel filtration
amplified cDNA corresponding to the region 559- column (GE Healthcare) equilibrated in column
649 of RCP was sub-cloned into pNIC28-Bsa4. buffer (10 mM Tris-HCl, 100 mM NaCl, 5mM
The RCP products contained an N-terminal MgCl , 1 mM DTT, pH 7.5). Expression and
2
hexahistidine tag and an rTEV cleavage site. purification of MBP-FIP2 and Rab25 has been
Table 1 is a list of the various constructs and their described previously (25,26).
applications, as well as corresponding acronyms For crystallization experiments, Rab14
7-
used in this manuscript. The construct MBP-FIP2 :wtRCP complexes were prepared by first
175 581-649
is a fusion of maltose-binding protein (MBP) with combining the individually-purified proteins. The
the indicated C-terminus of FIP2 (410-512), and combined samples were dialyzed against low salt
has been described previously (25). buffer (10mM Tris-Cl, 10mM NaCl, 1mM DTT,
FIP3 was generated by PCR using pH 7.5), followed by ion-exchange affinity
612-756
the following primers 5’-AAAGGATCCACAGA- chromatography on a MonoQTM 5/50 GL column
4
Structure of Rab14:RCP
(GE Healthcare). A salt gradient of 10-500 mM likely affects the backbone conformations in this
NaCl was used over 20mL, and fractions region. The atomic co -ordinates and structure
containing the complex were further purified on a factors have been deposited in the Protein Data
Superdex 200 (16/60) column. The final peaks Bank with the accession code 4d0g.
were collected and concentrated to between 7-15
mg/ml for use in crystal trials, using 10 kDa cutoff
concentrators (Millipore). Prior to crystallization, Isothermal titration calorimetery (ITC)
GppNHp (Jena Bioscience) was added to a final Experiments were performed in duplicate
concentration of 1 mM, since it appeared to on an ITC-200 calorimeter (GE Healthcare). In
improve the reproducibility of crystals. order to generate active Rabs, proteins were
Measurements of polydispersity were performed incubated in 10mM EDTA for 10 minutes at room
using a DynaPro NanoStar dynamic light temperature, followed by incubation with 1 mM
scattering instrument (Wyatt Corp). GppNHp, and supplemented by 15 mM MgCl .
2
Excess nucleotides were removed by running
Crystallization and structure determination samples through a PD10 column (GE healthcare).
Initial screening of Rab14 /wtRCP Protein concentrations were calculated based on
7-175 581-
with commercial sparse-matrix screens was their Abs using a ND-1000 NanoDrop
649 280 D
performed using a MosquitoR robot (TTP Spectrophotometer (Thermo Scientific). Proteins o
w
n
Labtech). Subsequent optimization of crystal were dialyzed together in buffer (10mM Tris-Cl, lo
a
growth was done manually by the hanging drop 100mM NaCl, 5mM MgCl2, and 1mM DTT, pH ded
method using Linbro plates. The complex was 7.5) to minimize heats from buffer mismatch. fro
m
crystallized at a concentration of 8.2 mg/ml in a Samples were centrifuged at 13,200 rpm for 10 h
1:1 ratio with reservoir containing 14% PEG 8000 minutes prior to concentration determination and ttp
and 0.1 M HEPES pH 6.75. ITC analysis. The concentrations of proteins for ://ww
w
Diffraction images were integrated with the injections were: (i) 600 µM MBP-FIP2410-512 .jb
HKL2000 (27) and scaled using SCALEPACK. fusion protein was placed in the syringe, titrated c.o
The structure was refined by first performing into 45 µM Rab25; (ii) 600 µM wtRCP581-649 in the brg/
y
molecular replacement using the structure of syringe, and 30 µM Rab142-175 in the cell; (iii) 600 gu
Rab14(GDP) as a search model (PDB code 4drz). µM Rab142-175 in the syringe, 60 µM MBP-FIP2 in est o
Electron density corresponding to RCP was clearly the cell; and (iv) 600 µM wtRCP581-649 in the n D
visible and was built manually using COOT(28). syringe, titrated into 45 µM Rab25. All titrations ece
Alternating cycles of model building and were performed at 293K. mb
e
refinement using Refmac5 (29) enabled r 1
3
completion of nearly the entire model of Rab14 , 2
0
(residues 7-175) and RCP (595-640), with R/Rf Static and dynamic light scattering 18
values of 0.24/0.31. At this stage, the model and Absolute molar mass was determined by
data were submitted to the PDB_REDO server in static light scattering using a miniDAWN
order to optimize the refinement strategy. Further instrument (Wyatt Technology Corp.) coupled to a
cycles of model building and refinement were Superdex 200 (10/300 GL) column. The elution
performed using Refmac5 and COOT. While the buffer was 10mM Tris-Cl, 100mM NaCl, 5mM
backbone density is continuous in the complex, MgCl , and 1mM DTT, pH 7.5. The
2
several loops and the N/C-termini are poorly concentrations of proteins injected into the system
ordered. The side chains of the following residues were typically 0.5mg/mL. Data were processed
were modelled as alanine: Tyr8-Tyr10, Glu33, using the software programme Astra v5.0 (Wyatt
Lys35, Met37, Lys140, Leu166, Glu167, Lys171, Corp). Rapid data acquisition (1/second) enabled
Gln174 in Rab14; and Leu601, Lys604, Gln605, several thousand independent measurements,
Arg615, Thr640 in RCP. In addition, there is poor which were averaged and used to estimate the
stereochemistry at Met37-Ala38, adjacent to the mean molar mass and associated error (Astra).
disulfide bridge (Cys20-Cys40). Phe36 packs Dynamic light scattering was performed using the
against an intramolecular disulfide bond, which
5
Structure of Rab14:RCP
NanoStar instrument (Wyatt Corp), typically using plasmid using TurboFect (Thermo Scientific)
a 1 mg/mL solution of protein complex in batch according to the manufacturer’s instructions. For
mode. At least 10 independent experiments were silencing experiments Lipofectamine RNAiMax
performed on each sample to obtain a mean (Invitrogen, Carlsbad, CA) was used.
estimated mass from the hydrodynamic radius,
along with an estimate of error. Immunofluorescence microscopy
Cells were seeded on 10 mm glass
Stoppped-flow fluorescence coverslips, fixed with 4% paraformaldehyde and
Data were collected using a PiStar blocked/permeabilized with 0.05% saponin/0.2%
instrument (Applied Photophysics) and processed bovine serum albumin. Cells were incubated with
using associated software. Laser excitation was the indicated primary antibodies in the blocking
set to 280nm for intrinsic tryptophan, and 1000 solution. The secondary antibodies used were
data points were collected over 500 milliseconds Cy3-conjugated donkey anti-rabbit from Jackson
following mixing of solutions. A filter was used ImmunoResearch Laboratories (West Grove, PA)
for detection of fluorescence emissions at and Alexa488–conjugated goat anti-chicken from
λ>320nm. Prior to injections, proteins were Molecular Probes. The cells were washed
dialyzed together in 10mM Tris-Cl, 100mM NaCl, extensively with phosphate-buffered saline (PBS)
D
5mM MgCl2, and 1mM DTT (pH 7.5). The between antibody incubations, and the coverslips ow
n
concentration of Rab14 was either 1µM or 2µM, were mounted in Mowiol. Images were acquired lo
a
while the concentrations of RCP ranged from on a Zeiss LSM510 confocal microscope (Carl de
d
10µM - 40µM. All injections were performed at Zeiss, Jena, Germany). fro
m
273K. These conditions enabled fitting of data to Pearson’s colocalization coefficients were h
ac oncpesneturdaoti-ofnirss,t thoerrdee rw asre aac tiloinne. ar Aret latitohnesseh ip cZaelicsusl Zateend 2u0s0in9g s othftew caorelo. calization module of the ttp://ww
w
between the observed rate constant (kobs) as a .jb
function of RCP concentration. The association Neurite Outgrowth c.o
rate (kon) was derived from the slope of a least- N2a cells were transfected with the byrg/
squares fit (minimum 4 concentrations), while the indicated constructs for 24 hours using TurboFect g
u
y-intercept provided the dissociation rate (koff). and were then induced to differentiate by est o
The dissociation constant (Kd) was calculated from incubation in serum-free medium for a further 24 n D
these parameters. This experimental strategy was hours. The cells were fixed and labelled with ec
e
m
previously established for the study of Rab6 with DAPI and anti-α tubulin. Confocal images were b
e
its effectors (30). recorded at 40X magnification and cells bearing r 1
3
neurites with lengths greater than twofold the cell , 2
0
1
Antibodies body diameter were scored as neurite-bearing cells 8
The antibodies used were rabbit anti- (31). Lengths were measured using ImageJ
Rab14 (R0656), mouse anti-α tubulin (T5168) and v1.49p.
chicken anti-RCP (GW21574A) from Sigma. The
Rab11 antibody is an ‘in house’ antibody raised in Results
a rabbit that had been injected with His -fused Gel filtration analyses of Rab14 complexes -
6
Rab11. A mouse monoclonal anti-Rab11 antibody Rab14 and FIP proteins were subjected to gel
from BD Biosciences (610656) was used for filtration analyses (Figure 1). Purified proteins
Western blots. were eluted on small (10/300) and large (16/60)
Superdex 200 columns to examine the formation
Cell culture and transfection of complexes. Rab147-175 (~20kDa) was shown to
HeLa, N2a and A431 cells were cultured interact with the C-terminal region of FIP2 (MBP-
in DMEM supplemented with 10% fetal bovine FIP2, Figure 1A), as well as the C-terminus of
serum, 100 U/ml penicillin/streptomycin and 2 RCP (wtRCP581-649, Figure 1B). These data show
mM glutamine. These cells were transfected with that Rab14 interacts with sufficient affinity to
6
Structure of Rab14:RCP
enable co-migration of Class I FIPs by gel Isothermal titration calorimetry - Rab GTPases
filtration chromatography. were purified, loaded with GppNHp, and titrated
A competition assay was performed in against various FIPs to determine their
which Rab25(GppNHp) was incubated with thermodynamic parameters by isothermal titration
Rab14 :MBP-FIP2, and the complete solution calorimetry. As described in Experimental
7-175
was analysed by gel filtration chromatography Procedures, titrations were performed with the
(Figure 1Given equivalent amounts of Rab25 carboxy-termini of the effectors, either as isolated
7-180
and Rab14 , we observed that Rab25 can out- (hexahistidine-tagged) proteins, or as fusions with
2-175
compete Rab14 for binding to FIP2. The gel MBP. Injections were performed in both directions
filtration profile reveals that Rab25 is bound to for some binary partners and no significant change
7-180
the effector, and that Rab14 is migrating alone. in the measured thermodynamic parameters was
2-175
This competition assay reveals that Rab25 can out- observed (Fig. 2, Table 4). The results show that
compete Rab14 for binding to FIP2 in vitro. Rab14 binds weakly to FIPs, relative to Rab11/25.

A similar competition assay was Furthermore, the stoichiometry of the interactions
performed for the binding of RCP with either are distinct – Rab11 and Rab25 complexes are 1:1,
Rab11 or Rab14. Excess amounts of Rab11 whereas Rab14:FIP complexes are closer to 1:2.
1-173
and Rab14 (160µM each) were incubated with These observations are consistent with data from
2-175 D
RCP (60µM) in a final volume of 0.5mL, and light scattering. ow
n
subjected to gel filtration chromatography. Under lo
a
these conditions, the majority of complex was Structure of Rab14:RCP - Crystals of the complex de
d
observed to be Rab11:RCP, with small amounts of grew in C2221 space group with one molecule fro
m
Rbianbd1s4 :tRoO CCvPel arisansl l t,hI et Fh'cIeoPsmes, praleelbsxue' liptts e awksih t(hoF wilgo uwtrheear t 1 aDRff)ai.n b i 1t4y eaaassyscehmm mboelfyt r iicnR asbou1lun4ti7it-o.1 7n5 cHoaonnwsdi estvswe rot,Rf Ca tP1h5:e821 -6c4o9b miopilnole gxit choaefl http://ww
w
relative to Rab11. Rab14:RCP, as evidenced by calorimetry, static .jb
and dynamic light scattering. Our interpretation is c.o
Light scattering analyses of Rab:FIP complexes - that the high concentrations of proteins in the brg/
y
Multi-angle light scattering (MALS) analyses of crystallization condition enables the loading of g
u
Rab257-180:MBP-FIP2 revealed that the molecular Rab14 onto the second (symmetric) binding site, est o
mass is consistent with a heterotetramer (Table 3). thus mediating lattice formation and crystal n D
In contrast, the mass of Rab14:RCP was consistent growth. It is not unusual to find non-native ec
e
with a 1:2 complex, as evidenced by MALS. stoichiometry in crystals – the complexes of mb
e
Dynamic light scattering (DLS) can also provide Rab11:FIP2 and Rab11:FIP3 also deviate from the r 1
3
insight into the stoichiometry of protein complexes biological heterotetramer (1:1, 2:2, and 3:3 in the , 2
0
via molecular weight estimates. In this case, the crystalline state (32-34). However, in all cases of 18
mass is inferred from the diffusion co-efficient, conventional Rab:FIP complexes, a 2-fold
which can be influenced by overall shape of symmetric heterodimer (2:2 complex) can be
complexes. The stoichiometry of Rab14 identified from crystallograhic or non-
7-
crystallography symmetry.
:wtRCP from DLS revealed an
175 581-649
The Rab14 interface with RCP is
intermediate ratio, which is possibly a mixture of
topologically identical to Rab11/Rab25
1:2 and 2:2 complexes. In contrast, the size of
interactions with FIPs (Figure 3A,B). Ile44 of
Rab25 :FIP2 was consistent with a 2:2
7-180 410-512 Rab14 resides at a hydrophobic interface, packing
complex by DLS (Table 3), in agreement with the
against the C-terminal half of the α-helical coiled-
stoichiometry from MALS. Thus, the overall
coil of RCP (residues 597-632). However, there
results from light scattering, in which assay
are several features that are distinct. A cysteine
conditions are 0.5-1 mg/mL protein, suggest a 1:2
residue in switch I, Cys40, forms a disulfide
stoichiometry for Rab14:RCP under dilute (non-
bridge with Cys26 in the Rab14:RCP crystal
crystallization) conditions.
complex (Figure 3C). This appears to be a GTP-

dependent disulfide bond, since the GDP-bound
7
Structure of Rab14:RCP
Rab14 (PDB code 1Z0F) is observed in the mutants were separated by SDS-PAGE,
reduced form with the switch I region distant from transferred to nitrocellulose, and overlaid with
the nucleotide (Figure 3B). The electron density GST-fused Rab11A or GST-Rab14 loaded with
for the disulfide bond in the crystal structure is GTPγS. Rab11 and Rab14 that bound to the His-
unambiguous (Figure 3C). In addition, there is a RCP fusions were revealed with an anti-GST
lack of electrostatic parity in switch I – Rab14 antibody. In contrast to the findings of Qi et al.,
contains a proline (Pro41), whereas both Rab11 we observed that both GST-Rab11 and GST-
and Rab25 have a positively charged residue Rab14 bound to the RCP and RCP mutants
NL NF
(Lys41 in Rab11, Arg42 in Rab25) that forms a (Figure 5C). However, consistent with our
salt bridge with FIP2. Thus, the negatively previous findings (17), neither Rab bound to
charged Glu616 of RCP, conserved in class I FIPs, RCP or the RCP mutant. These results
ΔRBD I621E
lacks an electrostatic partner in Rab14 (not confirm that in our in vitro Far Western binding
shown). The next residue in switch I of Rab14 is assays Rab14 does not require the serines at
His42, whose side chain is not positioned for a positions 580 and 582 to interact with RCP.
favourable interaction with FIPs. In fact, this
residue is a small polar side chain in most Rabs Gel filtration Analyses of Rab14:RCP complexes -
that interacts with the γ-phosphate of GTP (Figure In order to further assess the significance of the
D
4). The switch II region of Rab14 in the RCP recently reported serine mutations in RCP on ow
n
complex also reveals significant differences from Rab14 binding, extended C-terminal constructs of lo
a
d
the Rab11/Rab25 complexes with FIPs. Met20 in RCP (residues 559-649) were generated for E. coli e
d
the P-loop of Rab14 protrudes into space that is recombinant protein expression and subsequent in fro
m
generally occupied by the catalytic glutamine in vitro biochemical studies. These fragments of h
active Rabs. Thus, relative to Rab11/25, the wild-type and mutant RCP (wtRCP559-649, RCPNL, ttp://w
segment 70-73 in Rab14 is found in an alternative RCP ; see Table 1) were highly soluble and w
NF w
backbone conformation to avoid steric conflicts encompassed amino acids 580 and 582. The RCP .jb
c
with Met20. polypeptides were purified, incubated with .o
rg
Rab142-175 (Q70L), and analyzed by gel filtration b/
y
Binding assays - Wild-type and mutant versions of chromatography. All of the RCP variants formed a g
u
the C-terminus of RCP (Figure 5A) were stable complex with Rab14, and as a est o
subjected to an overlay assay with GST-Rab11A representative example, the results of the Rab142- n D
and GST-Rab14 in order to further clarify the Rab- :RCP complex are shown (Figure 6). ec
175 NF e
m
binding region. In a recent paper, Ser580 and b
e
Ser582 were mutated to asparagine and leucine, Fluorescence-based affinity studies - The binding r 1
3
respectively, in order to match the residues of the affinities for Rab14 and RCP were measured using , 2
0
1
corresponding positions in FIP2 [(35); Figure 5B]. established fluorescence techniques for 8
However, inspection of this region revealed that Rab:effector complexes (30). The proteins used in
the amino acid in FIP2 which aligns with Ser582 these studies are listed in Table 1. The intrinsic
of RCP is in fact a phenylalanine, not a leucine fluorescence changes (Trp excitation at 280nm,
(Figure 5B). In order to address conflicting results emission filter at 320nm) indicated binding of
between our prior studies (17) and that of Qi et al, Rab14 to the wild-type and mutant RCP variants
we analysed two double serine mutants in RCP. (Figure 7A). Analyses of binding kinetics
These were S580N+S582L (RCP ) mutant used (Figure 7B, Table 5) indicate that RCP binds to
NL NF
by Qi et al. and a S580N+S582F (RCP ) mutant Rab14 with low micromolar K , perhaps with
NF 2-175 d
that correctly matches FIP2. These mutations were slightly greater affinity than wtRCP . The
559-649
first introduced into a truncated form of RCP phosopho-mimetic RCP variant (S580E+S582E)
EE
(amino acids 385 – 649) and expressed as His- and a truncated variant of RCP (wtRCP ) also
590-649
fusions in E. coli. We also generated a mutant with bind effectively to Rab14 . In summary, the
2-175
a premature stop codon at position 591, just various mutations at Ser580 and Ser582 do not
upstream of the RBD (RCP ) (Figure 5A). disrupt binding to Rab14. As a negative control,
ΔRBD
Lysates of E. coli expressing wild-type RCP and RCP - which contains a mutation in the
I621E
8
Structure of Rab14:RCP
hydrophobic binding interface (I621E) - reveals no In order to rule out the possibility that high levels
interactions with Rab14. This provides direct of GFP-RCP in the Rab11-depleted cells masks
WT
evidence that the fluorescence changes are specific a population of RCP localized to membranes the
for the Rab14:RCP interface. experiment was repeated in cells that were
permeablized prior to fixation, in order to wash
Colocalization between RCP and its Rab binding out cytosolic proteins. Under these conditions no
partners - There were no obvious differences in GFP-RCP was observed localizing to
WT
the intracellular localization patterns of cytoplasmic membranes in the Rab11-depleted
exogenously expressed full length GFP-RCP or cells, whereas GFP-RCP clearly labelled
WT WT
the full length mutants GFP-RCP or GFP- vesicles in control and Rab14-depleted cells
NF
RCP All displayed a punctate vesicular pattern (Figure 11B). Surprisingly, a pool of GFP-RCP
NL. WT
distributed throughout the cell (Figure 8). In was observed in nucleoli in the Rab11-depleted
contrast, GFP-RCP and GFP-RCP were cells that had been pre-permeablized. The
ΔRBD I621E
both cytosolic. significance of this has yet to be determined.
Furthermore, considerable colocalization Similar results were observed for
was observed between both double serine mutants endogenous RCP in the A431 cell line. A431 cells
and endogenous Rab11 and Rab14 (Fig. 9A, B). are a human squamous carcinoma cell line in
D
Quantitative colocalization analysis revealed that which RCP-positive membranes primarily localize ow
n
there was no significant difference in the degree of in a compact pericentriolar region. This membrane lo
a
Rab11 and Rab14 colocalization with either localization is considerably reduced in Rab11- de
d
double serine mutants in comparison to GFP- depleted cells, whereas transfection with siRNAs fro
m
RgcorCeloaPtcWearTl izt(ahFtaiiongn u r5be0 e%t9w Cer)ee.dn u IcnRt iaocbno1 n1ti rnaa snttd,h et Rhaeamrbe1o 4uw nawt s iothaf to anr geentdinogg eRnaobu1s4 R, CorP a l occoanltirzoalt isoiRn N(FAig, uhraed 1n1oC e)f.f e ct http://ww
w
GtoF GPF-RPC-RPCΔRPBD a (nFdi gGurFeP 9-RCC).P I621E when compared Rneaubr1it4o gaenndes isR CP are functionally linked in .jbc.o
Of nWoTte, we consistently observed that In order to examine the functional brg/
y
GFP-RCP displayed greater colocalization with interplay between RCP and its Rab partners we g
WT u
endogenous Rab11 than with endogenous Rab14 analyzed the effects of RCP, Rab11, and Rab14 est o
(Figure 9 C, D). Similarly, endogenous RCP also mutants on neurite outgrowth in N2a mouse n D
showed greater colocalization with Rab11 than neuroblastoma cells. Overexpression of dominant- ec
e
with Rab14 (0.25 ± 0.06 and 0.17 ± 0.01, negative mutants of RCP and Rab11, or their mb
e
respectively) (Figure 10 A, B). knockdown, has been shown to inhibit neurite r 1
3
outgrowth in dorsal root ganglion neurons and , 2
0
RCP recruitment to membranes - To examine PC12 cells (36). Neuritogenesis can be readily 18
whether Rab11 and/or Rab14 are required to induced in N2a cells by withdrawing serum. We
recruit RCP to membranes GFP-RCP was transfected these cells with wild-type and
WT
expressed in HeLa cells in which these Rabs had dominant-negative mutants of RCP, Rab11, and
been individually depleted. We could efficiently Rab14 for twenty-four hours and then incubated
deplete both Rabs by greater than 70% (Figure them for a further twenty-four hours in serum-free
11D). The localization of GFP-RCP was medium. Surprisingly, dominant-negative Rab11,
WT
examined in fixed cells by confocal microscopy. (Rab11AS25N) had no effect on the number of N2a
In cells transfected with a control siRNA (a cells that formed neurites. In contrast, there was a
sequence complementary to the firefly luciferase greater than 70% reduction of neurite formation in
gene, siFLuc), GFP-RCP displayed a typical transfected cells expressing dominant-negative
WT
punctate pattern throughout the cell. However, in Rab14 (Rab14S25N) or two RCP mutants (RCPI621E
cells in which Rab11 had been depleted GFP- and RCP385-649) in comparison with cells
RCP was cytosolic (Figure 11A). In contrast, expressing GFP alone, or the wild-type versions of
WT
depletion of Rab14 with two independent siRNA each protein (Figure 12 A, B). Interestingly, both
duplexes did not affect its membrane localization. RCPWT and Rab14WT displayed a striking
9
Structure of Rab14:RCP
concentration at the neurite tips whereas Rab11 , acids S580 and S582 contribute towards Rab14
WT
Rab11 Rab14 , and RCP were binding, we performed kinetic analyses of
S25N, S25N 385-649
predominantly localized to the cell body (Figure Rab:RCP interactions by fluorescence techniques.
12A). RCP displayed a diffuse cytosolic We found that the RCP mutant (S580N+S582F)
I621E NF
pattern. These results suggest that in N2a cells was not significantly impaired in binding to
RCP functions in conjunction with Rab14, but Rab14. Similarly, the phospho-mimetic RCP
EE
independently of Rab11, to regulate neurite mutant (S580E+S582E) bound to Rab14 at least as
formation. strongly as wild-type RCP. We cannot, however,
definitively rule out a functional role for regions
DISCUSSION upstream of the conventional RBD, perhaps in
Structure of Rab14:RCP modulating the strength of protein interactions.
The structure of Rab14:RCP reveals Our in vitro measurements were limited to the
several unusual features when compared to region 559-649 of RCP, since longer fragments
conventional FIP complexes with Rab11 and were unstable and either degraded or mis-folded.
Rab25. An intramolecular disulfide bond is Nevertheless, our work demonstrates that the
observed between Cys26 and Cys40 - Rab14 primary mode of the Rab14 interaction with RCP
uniquely contains these two residues among the is via the C-terminal RBD.
D
Rab family of GTPases. Upon superposition of o
w
n
several Rab:effector complexes [Rab8:OCRL, RCP membrane association and role in lo
a
Rab3:Rabenosyn; (11,37)], the C atom of the neuritogenesis de
d
equivalent loci are within 6Å disβtance. Thus, Although Rab14 clearly binds to RCP and fro
m
these two segments of Rabs, which comprise helix the two proteins colocalize in HeLa cells in an h
RBD-dependent manner, the mechanism for the ttp
αsw1i tcahn dI ),t haer e sculbosseeq uine nst paloceo pi n( wthhei cahc tipvree cGedTePs subcellular association of RCP with Rab14 ://ww
w
state. However, the significance of a disulfide remains unclear. RNA interference experiments .jb
bond and whether it forms in vivo requires further suggest that, at least in some cell types, Rab11 is c.o
structural and functional studies. It is noteworthy the principal Rab that recruits RCP to vesicular brg/
y
that Cys26 of Rab1 forms an intramolecular membranes. Depletion of Rab11 results in the g
u
disulfide bridge with Cys126 in the Rab1:DrrA majority of RCP becoming cytosolic, whereas est o
complex (38), suggesting that it is indeed reactive Rab14 depletion did not result in a similar effect. n D
in vitro. Despite the disulfide bridge in Rab14, the However, our results in N2a cells demonstrating ec
e
switch I conformation closely resembles the that RCP and Rab14, but not Rab11, are important mb
e
structures of Rab11/25 in complex with FIPs. In for neuritogenesis suggest that RCP also has r 1
3
contrast, switch II is significantly perturbed, Rab11-independent functions. Indeed, Qi et al , 2
0
mainly due to the presence of Met20 which reported that the RCP-dependent transport of the 18
HIV-1 Env protein does not involve Rab11 (23).
sterically interferes with the region facing the γ-
There are many examples of Rabs that
phosphate of GTP. These conformational
share/compete for effectors. For example, Rab6
differences along with reduced electrostatic
and Rab11 bind to Rab6IP1/DENND5B (10),
complementarity may account for the observed
Rab4 and Rab14 interact with RUFY1 (39), and
weaker affinity between Rab14 and class I FIPs.
both Rabenosyn-5 (11) and myosin Va (40) bind
A recent study has suggested that Rab14
to multiple Rabs. Further complexity arises from
binding to RCP is mediated by a region upstream
Rab cascades in which one Rab mediates the
of the conventional Rab-binding domain. Residues
transport of a guanine-nucleotide exchange factor
Ser580/Ser582 were reported to be critical for
(GEF) that activates a downstream Rab. Such
RCP binding. However, our structural,
cascades regulate secretory vesicle transport (41),
biochemical and cellular assays are consistent with
primary ciliogenesis (42), and intra-Golgi
binding of Rab14 to the classical RBD of the FIPs
transport (43). Further work is required to
(residues 595-640 of RCP), as we have previously
determine if there is an interplay between Rab11
reported (17). In order to investigate if amino
and Rab14 in the regulation of RCP function and
10