Báo cáo khoa học: Proteomic and biochemical analysis of 14-3-3-binding proteins during C2-ceramide-induced apoptosis pot

C2-ceramide promotes changes in 14-3-3-binding patterns in HeLa cells during C2-ceramide-induced apoptosis With the aim of further analyzing the role of 14-3-3 proteins in apoptosis, an

proteins during C2-ceramide-induced apoptosis

l A list of the large number of protein–protein interactions described in this article is available via the MINT article ID MINT-7899808

Abbreviations

CAN, acetonitrile; ASK1, apoptosis signal-regulating kinase 1; B23, nucleophosmin; BAD, Bcl-xL ⁄ Bcl-2-associated death promoter; BAK, Bcl2-antagonist ⁄ killer; BAX, Bcl2-associated X protein; BMH1 ⁄ 2, yeast 14-3-3 homolog; CaM, calmodulin; COX IV, cytochrome c oxidase subunit IV; DIG, digoxigenin; DNA-PK, DNA-dependent protein kinase; FADD, Fas-associated death domain; FOXO, forkhead box protein; G418, geneticin; GADPH, glyceraldehyde-3-phosphate dehydrogenase; GFP, green fluorescent protein; HIP-55, hematopoietic progenitor kinase 1-interacting protein of 55 kDa; JC-1, 5,5¢,6,6¢-tetrachloro-1,1¢,3,3¢-tetraethylbenzimidazolylcarbocyanine iodide; LC-MS ⁄ MS, liquid chromatography-tandem MS; MAPK, mitogen-activated protein kinase; NF-jB, nuclear factor-jB; RIP1, receptor-interacting protein 1; RIP3, receptor-interacting protein 3; siRNA, small interfering RNA; Smac, second mitochondrial-derived activator of caspase; STAT3, signal transducer and activator of transcription 3; TAP, tandem affinity purification; TNF-a, tumor necrosis factor-a; TSC2, tuberous sclerosis protein 2; VASP, vasodilator-stimulated phosphoprotein.

The term 14-3-3 denotes a large family of acidic

pro-teins that exist primarily as homodimers and

heterodi-mers within all eukaryotic cells [1,2] In mammals,

there are seven 14-3-3 isoforms, designated by Greek

letters (a⁄ b, g, e, c, s ⁄ h, f ⁄ d, and r) and encoded by

seven different genes [3,4] 14-3-3 proteins play central

regulatory roles in eukaryotic cells by binding to

diverse target proteins, thereby modulating the function

of the associated partners [5] In most cases, 14-3-3

proteins regulate cellular processes by binding to

spe-cific phosphoserine and phosphothreonine motifs

within target proteins [6] Two optimal 14-3-3

phos-phopeptide ligands with the consensus sequences

RSX(pS⁄ T)XP and RX(Y ⁄ F)X(pS ⁄ T)XP (where pS ⁄ T

represents phosphoserine or phosphothreonine, and X

is any amino acid) have been defined [7] Alternatively,

some 14-3-3 proteins bind to phosphorylated motifs

that are completely different to the consensus sites

described above [8], or even bind to unphosphorylated

motifs [9]

14-3-3 binding can alter the enzymatic activity,

sub-cellular localization, protein–protein interactions,

dephosphorylation and proteolysis of individual target

proteins [10] Many 14-3-3 target proteins have been

shown to be involved in cancers, diabetes, Parkinson’s

disease, and other neurological diseases [11]

More-over, 14-3-3 proteins have been shown to be key

regu-lators of a large number of processes, such as control

of cell proliferation, the cell cycle, regulation of human

metabolism, and apoptosis in mammalian cells [12–20]

In a number of cases, interaction of 14-3-3 proteins

with their target proteins promotes events that support

cell survival, mediating an essential antiapoptotic

signal in cells [21]

Apoptosis is an active process of cell death that

plays a critical role in normal development,

mainte-nance of tissue homoeostasis and elimination of

dam-aged or unwanted cells through a balance of

antiapoptotic and proapoptotic factors, which may be

shifted by extracellular signals [22] It has been

reported that 14-3-3 binds members of the Bcl-2

fam-ily, named Bcl-xL⁄ Bcl-2-associated death promoter

(BAD) and Bcl-2-associated X protein (BAX),

inhibit-ing their proapoptotic activities [23,24] 14-3-3 inhibits

cell death caused by other death promoters, such as

apoptosis signal-regulating kinase 1 (ASK1) [25]

Fur-thermore, 14-3-3 protein binds to a member of the

family of forkhead transcription factors named

fork-head box protein (FOXO), blocking its translocation

to the nucleus and later activation of death genes [26]

These functions of 14-3-3 proteins have been reported

to be dependent on their dimeric structure Thedimeric status of 14-3-3 proteins is regulated by site-specific serine (Ser58) phosphorylation by sphingosine-dependent kinase 1 This serine is located within thedimer interface of 14-3-3 proteins, and its phosphoryla-tion promotes the formation of a monomeric form of14-3-3 Thus, phosphorylation of Ser58 on 14-3-3fcontrols its ability to modulate target protein activity,and this may have significant implications for the regu-lation of many cellular processes, including apoptosis,

by preventing dimer-dependent inactivation of poptotic BAD or BAX [27] Ceramide, a bioactivelipid mediator, was found to be an apoptosis inducerthat activates sphingosine-dependent kinase 1, regu-lates Bcl-2 expression, blocks survival signals, and acti-vates phosphatases (protein phosphatase 1 and proteinphosphatase 2A) [28–31] Several studies have pro-posed that ceramide and its metabolic derivatives betherapeutically applied in cancer-suppressing strategies[32–36]

proa-Inhibition of apoptosis by 14-3-3, through knownprocesses such as association with BAD, FOXO, andASK1, and other unknown processes that involvemitogen-activated protein kinase (MAPK) and phos-phoinositide 3-kinase cascades, suggests that 14-3-3has an important antiapoptotic function Expression of

a polypeptide that prevents 14-3-3 proteins from ing to targets in mammalian cells triggers apoptosisand decreases viability in prostate, lung and cervixcancer cell lines [37,38] Furthermore, treatment with2-methoxyestradiol resulted in decreased 14-3-3 expres-sion that, in parallel with apoptosis induction,decreased cell growth [39], and the use of 14-3-3f anti-sense in cancer cell lines increased the sensitivity of thecells to stress-induced apoptosis, such as that induced

bind-by UV light, IR light, and doxorubicin [40–42] On theother hand, several studies found increased expression

of 14-3-3f in lung, stomach and breast cancers [42–47].These data suggest that 14-3-3 proteins have a role inregulating cancer cell proliferation and, as such, could

be targeted by cancer therapies

Several proteomics studies have been performed tofind new 14-3-3-interactor proteins under physiologicalconditions or even during mitosis [12–16,18–20] Never-theless, the work reported here is the first study

to include a comprehensive proteomics analysis of14-3-3-binding proteins under physiological conditions

as compared with apoptosis stimulation, with the aim ofincreasing our knowledge of the role of 14-3-3 proteins

in the apoptotic pathway Because antineoplastic pies ultimately eliminate tumor cells by the induction of

thera-apoptosis, a comprehensive understanding of how

14-3-3-mediated survival pathways inhibit apoptosis

may allow the use of 14-3-3 antagonists to sensitize

tumor cells for effective therapy

Thus, to identify novel cellular survival functions of

14-3-3 proteins, global proteomics and biochemical

analyses were carried out to identify proteins that

bind 14-3-3 proteins during apoptotic and survival

conditions These 14-3-3-interacting proteins were

purified from extracts of both control and

C2-cera-mide-stimulated HeLa cells, using tandem affinity

purification (TAP) methodology The proteins,

identi-fied by liquid chromatography–tandem MS

(LC-MS⁄ MS) analysis, were involved in multiple cellular

biological processes, but a pool of these proteins

had important functions in apoptosis through

regula-tion of intermediate filament integrity, cell blebbing,

formation of apoptotic bodies, DNA repair, and

regulation of oncogenic or death promoters

dur-ing apoptosis Usdur-ing the small interferdur-ing RNA

(siRNA) technique, the survival role of 14-3-3f during

C2-ceramide-induced apoptosis was characterized

The involvement of identified C2-ceramide-regulated

14-3-3-binding proteins with several processes that

control apoptosis suggests possible survival roles of

14-3-3 proteins in addition to others that have been

previously characterized

Results

Identification of 14-3-3-binding proteins related

to apoptosis

A few years ago, in a proteomics study of

14-3-3-affin-ity purification of over 200 human phosphoproteins,

new links of 14-3-3 proteins with the regulation of

cellular metabolism, proliferation and trafficking were

shown [12] Related to the functions of 14-3-3 proteins

as regulators of cell survival with central roles in

inhib-iting apoptosis, several apoptotic-related

14-3-3-bind-ing proteins were identified in our study Thus, we

found further 14-3-3-interactor proteins that are

regu-lators of apoptosis, such as receptor-interacting protein

kinase 1 (RIP1), programmed cell death protein

6⁄ ALG2 (apoptosis-linked gene 2), second

mitochon-drial-derived activator of caspase (Smac), signal

trans-ducer and activator of transcription 3 (STAT3), and

hematopoietic progenitor kinase 1-interacting protein

of 55 kDa (HIP-55) Using both MS and

MALDI-TOF⁄ TOF MS tryptic mass fingerprinting, those

proteins were identified as 14-3-3-interactor proteins;

however, studies of their presence in the eluted fraction

from the 14-3-3-affinity chromatography column to

confirm these data were not performed at the time.Here, western blotting analysis showed the presence ofthe corresponding protein with the appropriate molec-ular mass in the ARAApSAPA elution pool from the14-3-3-affinity chromatography column (Fig 1) Thesedata show that proteins such as RIP1, Smac, STAT3and HIP-55 were eluted from the affinity column, con-firming these proteins as 14-3-3-interactor proteinsunder physiological conditions Note that none ofthese proteins was eluted from the column by eitherextensive washing under high-salt conditions or mockelution with control phosphopeptides that do not bind

to 14-3-3 proteins These results indicate that isolatedproteins bind to the phosphopeptide-binding sites onthe 14-3-3 proteins, either directly or as components ofprotein complexes

As mentioned above, 14-3-3 interacts with sis-related proteins such as BAD, FOXO or ASK1 toperform its apoptosis-suppressing role in cells Here,

BAXBIDCaspase-8Caspase-9FADD

HIP-55

BAKBAD

RIP1SmacSTAT3

RIP3

Fig 1 14-3-3-affinity chromatography of human HeLa cell extracts Clarified HeLa cell extract was subjected to chromatography on 14-3-3–Sepharose, as described in Experimental procedures Column fractions were subjected to SDS ⁄ PAGE, using 10% Tris ⁄ glycine gels, and transferred to nitrocellulose membranes The amounts of protein subjected to SDS ⁄ PAGE were as follows: extract, flow through and beginning of salt wash (1st Wash), 40 lg of each; middle and end of salt wash (2nd Wash and 3rd Wash, respectively), protein undetectable; control (phospho)peptide pool, < 1 lg; and ARAApSAPA elution pool, 2 lg Western blots were probed with antibodies against the indicated proteins related to apoptosis.

14-3-3 interaction with other apoptosis-related proteins

was analyzed by its presence in the 14-3-3-affinity

chromatography elution pool Thus, proapoptotic

pro-teins such as receptor-interacting protein kinase 3

(RIP3) and Bcl-2-antagonist⁄ killer (BAK) were eluted

from the 14-3-3-affinity chromatography column,

sug-gesting a broad role of 14-3-3 proteins in apoptosis

regulation Note that the well-known apoptosis-related

14-3-3-binding protein BAD [23] was eluted from the

affinity chromatography column, giving confidence in

this technique Additionally, the proapoptotic protein

BAX [24], which is known to be a 14-3-3-interactor

protein, did not appear to be eluted from the column,

probably because its defined interaction with 14-3-3

proteins is independent of phosphorylation (which is a

requirement for elution from the column) On the

other hand, members of extrinsic apoptosis pathways,

such as caspase-8, Fas-associated death domain

(FADD), and Bcl-2-interacting domain, did not bind

to 14-3-3 proteins under the conditions tested

C2-ceramide promotes changes in 14-3-3-binding

patterns in HeLa cells during

C2-ceramide-induced apoptosis

With the aim of further analyzing the role of 14-3-3

proteins in apoptosis, an evaluation of the ability of

proteins to bind and to be regulated by 14-3-3 proteins

during C2-ceramide-induced apoptotis was carried out

Previous results have established C2-ceramide as an

inducer of programmed cell death [28] Thus,

C2-cera-mide-induced cell death in HeLa cells was analyzed,

and the time when this death occurred was established

HeLa cells were left untreated or exposed to

C2-cera-mide (50 lm) for the indicated times (Fig 2A) Sample

extracts were processed, and cell death was determined

as a percentage of the sub-G1 population The results

in Fig 2A show that 50 lm C2-ceramide promoted cell

death in HeLa cells in a time-dependent manner

In order to evaluate the 14-3-3-binding status of

proteins from HeLa cell extracts during

C2-ceramide-induced cell death, cells were treated in the presence or

absence of C2-ceramide (50 lm), and clarified extracts

were run into a gel and electrotransferred to a

nitrocel-lulose membrane Ponceau dyes showed differential

protein expression, probably because ceramide is

linked to nuclear factor-jB (NFjB) and SAPK⁄ JNK

cascades, which control protein expression in cells

[48,49], or perhaps because death initiation requires

caspase-dependent cleavage of specific targets [50–54]

Nevertheless, a digoxigenin (DIG)–14-3-3 overlay assay

showed protein bands with a significantly decreased

14-3-3-binding signal during C2-ceramide-induced cell

death (Fig 2B) These data are intriguing, and maysuggest deregulation of the association of 14-3-3 pro-teins with their targets during C2-ceramide treatment

To further investigate the role of 14-3-3 proteins ing C2-ceramide treatment, downregulation of 14-3-3proteins was performed and its effects on C2-ceramidecell death were analyzed in HeLa cells

dur-First, levels of expression of seven human 14-3-3 forms were analyzed in a cervical cancer cell line(HeLa) and in several breast cancer cell lines (Fig 3).The data showed that four different 14-3-3 isoformswere expressed in HeLa cells, 14-3-3f and 14-3-3hbeing the best expressed Note that similar results were

iso-A

0 10 20 30 40 50

bind-found in different breast cancer cell lines, where

14-3-3f and 14-3-3h were expressed well and

uni-formly Meanwhile, other human 14-3-3 isoforms

showed low expression levels in HeLa cells, and were

also differently expressed in several types of breast

cancer cell line Previous reports suggested that 14-3-3f

overexpression occurs in a high percentage of breast

tumors in the early stage of the disease, contributing

to the transformation of cells and also to the further

progression of breast cancer [42] On the other hand,

downregulation of 14-3-3f reduced

anchorage-indepen-dent growth and sensitized cells to stress-induced

apoptosis [42] These data suggest an important role of

14-3-3f overexpression in cancer; it is considered to be

a molecular marker for disease recurrence in breast

cancer patients, and may serve as an effective

thera-peutic target in patients whose tumors overexpress

14-3-3f On the other hand, many reports suggest

important regulatory functions of this isoform in the

apoptotic pathway, through interactions with specific

components of the apoptotic process [55,56]

Downregulation of 14-3-3f with siRNA

oligonucleotide enhances C2-ceramide-induced

apoptosis in HeLa cells

To investigate the role of 14-3-3f downregulation

during C2-ceramide-induced apoptosis, sensitization

effects on cell death were analyzed in Hela cells inwhich 14-3-3 binding was blocked by decreasing thelevels of 14-3-3f expression, using 14-3-3f siRNA.Clarified extracts from HeLa cells, transfected with14-3-3f siRNA or scrambled siRNA, were immunob-lotted with antibodies against all human 14-3-3 iso-forms (note that all mammal isoforms were tested, butonly four of them were visible in HeLa cells) Fig-

ure 4A shows specific downregulation of 14-3-3f forms by 14-3-3f siRNA oligonucleotide, but nodifference was observed in other human isoforms.Cell death was determined as the percentage of thesub-G1 population, in order to evaluate the effects of14-3-3f downregulation on C2-ceramide inducedapoptosis in transfected HeLa cells with 14-3-3f orscrambled siRNA The results in Fig 4B show that14-3-3f siRNA did not promote cell death on its ownafter 48 h of transfection (or after an additional 24 h;data not shown) Otherwise, downregulation ofendogenous 14-3-3f sensitized HeLa cells to cell deathpromoted by C2-ceramide at 50 lm Previously, itwas reported that downregulation of 14-3-3 proteinssensitized cells to stress-induced apoptosis, such asthat induced by UV light and doxorubicin [41,42] To

iso-my knowledge, this is the first study to analyze

in detail the effects of 14-3-3 downregulation onC2-ceramide-induced apoptosis These results suggest

an important role of 14-3-3f in C2-ceramide-inducedcell death, probably by binding to and regulation ofspecific targets that play important roles in C2-cera-mide-induced cell death

Knockdown of 14-3-3f promotes induced activation of caspase-8 and regulation

C2-ceramide-of the mitochondrial apoptotic pathwayMitochondrial dysfunction appears to be important

in C2-ceramide signaling of apoptosis In vitro studieshave shown that C2-ceramide itself is not an efficientinducer of nuclear apoptosis, unless mitochondria arepresent [57] It is still a matter of debate whetherC2-ceramide acts directly or indirectly on mitochon-dria, but some data suggest that C2-ceramide couldsignal mitochondrial apoptosis by inhibiting the pro-tein kinase Akt, which is responsible for BAD phos-phorylation, hence leading to inhibition of theantiapoptotic protein Bcl-2 by BAD [58–60] More-over, C2-ceramide induces cytochrome c release frommitochondria in a caspase-independent fashion,leading to the activation of executioner caspases andalso activation of the initiator caspase-8 [61], effectsthat are completely abolished by Bcl-2 and Bcl-xL[62,63]

Fig 3 Analysis of expression levels of several 14-3-3 isoforms in

cervical and breast cancer cell lines Extracts from cervical cancer

cells (HeLa) and several breast cancer cell lines (EvsaT,

MDA-MB-435, MDA-MB-231, MCF-7 ⁄ E6, MCF-7 ⁄ C4, BT-474, and SKBR3)

(30 lg), grown under physiological conditions, were subjected to

SDS ⁄ PAGE, using 10% Tris ⁄ glycine gels, and transferred to a

nitro-cellulose membrane Western blots were probed with antibodies

against several isoforms of 14-3-3 proteins.

As downregulation of 14-3-3f has been seen to

enhance C2-ceramide-induced cell death, the aim was

to obtain further insights into the mechanism of

sensi-tization to C2-ceramide with 14-3-3f siRNA by

investi-gating the C2-ceramide-induced mitochondrial

apoptotic pathway Therefore, western blot analysis

was performed to examine the presence of

cyto-chrome c in cytosolic and membrane fractions from

extracts of HeLa cells transfected with 14-3-3f siRNA

and treated with C2-ceramide The results in Fig 4C

show lowered cytochrome c levels in the

mitochondria-containing membrane fraction and the release of

cyto-chrome c to the cytosolic fraction on C2-ceramide

treatment when 14-3-3f was downregulated

To confirm that the apoptosis cascade was fully

active in 14-3-3f siRNA-transfected HeLa cells treated

with C2-ceramide, the proteolytic degradation of thenuclear protein poly(ADP-ribose) polymerase (PARP),

a substrate of effector caspases, and of the effector pase-8 were analyzed As shown in Fig 4D, PARPcleavage was clearly induced in C2-ceramide-treatedHeLa cells previously transfected with 14-3-3f siRNA,but no PARP cleavage was observed in untreatedHeLa cells Cell extracts of indicated samples wereanalyzed by western blot to determine caspase-8 acti-vation Procaspase-8 is first cleaved to the p43⁄ p41intermediate fragments, releasing the small subunitp12, and then subsequently processed to generate thelarge, catalytically active p18 subunit [64] On theother hand, procaspase-8 has been reported to becleaved in the presence of C2-ceramide, both nativeand exogenous, releasing active caspase-8, showingthat caspase-8 plays a role downstream of C2-ceramide

cas-in the cell death process [65,66] As shown cas-in Fig 4D,neither the downregulation of 14-3-3f nor C2-ceramidetreatment alone promoted caspase-8 activation at theindicated times, but a combination of both led tothe processing of procaspase-8 to its 43 and 41 kDa

Fig 4 Downregulation of endogenous 14-3-3f sensitizes cells to C2-ceramide-dependent apoptosis (A) HeLa cells were transfected either with siRNA oligonucleotide targeting 14-3-3f or with a scram- bled RNA oligonucleotide, as described in Experimental procedures After 48 h, extracts from untransfected cells (C) or cells transfected either with siRNA 14-3-3f (14-3-3) or with scrambled siRNA (SC) were harvested for immunoblot analysis to verify knockdown of endogenous 14-3-3f but not other isoforms (14-3-3r, 14-3-3e, and 14-3-3h) Tubulin was used as a protein loading control (B) HeLa cells transfected either with 14-3-3f or scrambled siRNA oligonu- cleotide, or without siRNA (control), were treated with 50 l M C2-ceramide for the indicated times Apoptosis was measured as percentage of cells with sub-G1 DNA content, as described in Experimental procedures Columns represent the average of three different experiments (C) HeLa cells were transfected as in (A) and treated with 50 l M C2-ceramide for an additional 4 or 8 h Follow- ing treatment, cells were lysed, and cytosolic proteins were sepa- rated from mitochondria as described in Experimental procedures Levels of cytochrome c in cytosolic and membrane fractions were determined by western blot COX IV was used as a mitochondrial loading control, and tubulin was used as a cytosolic protein loading control (D) HeLa cells untransfected (C) or transfected either with siRNA oligonucleotide targeting 14-3-3f (14-3-3) or with a scram- bled RNA oligonucleotide (SC) were treated in the presence or absence of 50 l M C2-ceramide for an additional 4 h HeLa cells were harvested for immunoblotting to analyze caspase-8 process- ing with mouse monoclonal antibody against human caspase-8 Both the 55 ⁄ 53 kDa native forms and the 43 ⁄ 41 kDa intermediate cleavage products are indicated by arrows PARP cleavage was detected by immunoblotting with antibody against PARP; intermedi- ate cleavage products are indicated by arrows 14-3-3f antibodies were used to verify knockdown of this isoform, and tubulin was used as a protein loading control.

intermediate fragments In conclusion, downregulation

of endogenous 14-3-3f sensitizes HeLa cells to the

C2-ceramide-induced mitochondrial apoptotic pathway

and activation of caspase-8 and PARP cleavage These

data suggest an extensive and important role of 14-3-3

proteins in C2-ceramide-induced apoptosis, probably

through regulation of already known apoptosis-related

14-3-3-binding proteins, some of them most likely still

to be identified

Purification of 14-3-3-binding proteins from HeLa

cells stably expressing green fluorescent protein

(GFP)–TAP–14-3-3f by TAP method

The data shown above suggest an interesting role of

14-3-3 proteins in C2-ceramide-induced apoptosis,

taking into consideration that 14-3-3 downregulation

sensitizes cells to C2-ceramide-induced apoptosis

Thus, it was considered that 14-3-3 proteins modulated

C2-ceramide-induced apoptosis by binding to

well-known apoptosis-related proteins, but possibly also by

association with other targets with central roles in the

apoptotic process that remain to be identified

There-fore, the aim was to identify new targets of 14-3-3

proteins involved in C2-ceramide-induced apoptosis

To identify proteins associated with 14-3-3 in vivo,

a TAP tag approach was used, which allows the

isola-tion of native protein complexes from cells ectopically

expressing the tagged protein of interest [67] The TAP

tag was fused to 14-3-3f as previously described [68]

This construct, generously provided by D Alessi

(MRC, Dundee, UK), was successfully used to analyze

LKB1 phosphorylation-dependent 14-3-3 binding of

protein kinases closely related to AMP-activated

pro-tein kinase, such as QSK and SIK, in 293 cells [68]

Here, HeLa cells stably expressing

GFP–TAP–14-3-3f were generated and analyzed to determine the size,

level of expression and distribution of stably

transfect-ed fusion protein (Fig 5A,B) Western blot analysis

with polyclonal antibody against 14-3-3f showed GFP–

TAP–14-3-3f of the expected size with a similar level of

expression to that of endogenous protein (Fig 5A)

Moreover, the fusion protein showed a cytoplasmic

localization identical to the previously described

locali-zation for endogenous 14-3-3f [4,69] (Fig 5B)

With regard to the goal of purifying and identifying

new 14-3-3-binding proteins involved in

C2-ceramide-induced apoptosis, HeLa cells stably expressing

GFP–TAP–14-3-3f were used for subsequent protein

purification and identification by the TAP method

Thus, stably transfected HeLa cells were either

exponen-tially proliferating (untreated) or treated with

C2-cera-mide to induce apoptosis (see Experimental procedures)

Eluted pools from control and C2-ceramide-treatedGFP–TAP–14-3-3f-expressing HeLa cells, purified byTAP, were further analyzed by LC-MS⁄ MS

Identification of 14-3-3-affinity purified proteins

by LC-MS⁄ MS analysisAnalysis by LC-MS⁄ MS of purified 14-3-3-bindingproteins from cells undergoing control and C2-cera-mide-induced apoptosis showed different potentialligands of 14-3-3f in both conditions The 14-3-3 inter-actors were grouped according to the processes inwhich they had previously been involved (Tables 1 andS1) The identified 14-3-3-binding proteins includedproteins involved in cell signaling, metabolic pathways,

19 kDa

HeLa HeLa 14-3-3ζ

cytoskeletal dynamics, RNA binding, DNA bindingand chromatin structure, cellular trafficking, and pro-tein folding Some of them were previously shown to

be associated with 14-3-3 isoforms (indicated inTable S1) Detection of those 14-3-3 ligands already

Table 1 Comparative analysis of 14-3-3-binding proteins identified

by TAP–MS from control or C2-ceramide-treated

GFP–TAP–14-3-3f-expressing HeLa cells This is an abbreviated version of Table S1;

proteins identified by TAP and LC-MS ⁄ MS analysis were grouped

into functional classes, and data were searched against the

Euro-pean Bioinformatics Institute ⁄ International Protein Index human

database, using the MASCOT search algorithm (see Experimental

procedures) The data were obtained by LC-MS ⁄ MS analysis of

tandem affinity-purified 14-3-3f-associated proteins from GFP–TAP–

14-3-3f-expressing HeLa cells left untreated (control) or stimulated

with C2-ceramide to induce apoptosis Each protein identification

was manually confirmed to ensure that no other human proteins

matched the peptide sequences obtained Interactions validated by

biochemical methods are indicated in bold.

Histone H2A type 1

B23

Ttransforming growth factor-b

-induced transcription factor

40S Ribosomal protein S3 Elongation factor 1 a1

Protein folding and processing

hydrolase 42

ATP synthase subunit b Carbamoyl-phosphate synthase, mitochondrial precursor

CoA synthase U6 snRNA-specific terminal uridylyltransferase 1 Cellular signaling

Hydroxymethylglutaryl-Histone H1.2

Myosin regulatory light chain 2 Myosin light chain

kinase 2 Titin

Table 1 (Continued).

CaM Centrosomal Nek2-associated protein 1

TSC2 Myosin light chain kinase 2

14-3-3r 14-3-3b ⁄ a DNA-PK catalytic subunit Serine ⁄ threonine protein kinase WNK4 Cellular organization

Vimentin Lamin-A ⁄ C a-Actinin-2 a-Actinin-3 Desmin VASP Myosin-2 Myosin-3 Myosin-7 (myosin heavy chain 7)

Ankyrin repeat domain-containing protein 18A

Ankyrin repeat domain-containing protein 18A Heat-shock protein b1 a-Actin-2

Unclassified Keratin, type II cytoskeletal 8 Keratin, type II

cytoskeletal 8 Keratin, type I cytoskeletal 17 Keratin, type I

cytoskeletal 17 Keratin, type I cytoskeletal 18

Tropomyosin-1 a chain

known implies that the conditions used here for the

TAP tag purification allowed the identification of

genuine 14-3-3 ligands Detailed analysis of the

14-3-3-asociated proteins found showed that 46 of them

were exclusively present in one of the conditions

ana-lyzed and 15 of them were involved, to a greater or

les-ser extent, in the apoptotic process, according to

previous reports (detailed inTable 2) It is interesting to

note that 14-3-3f copurified with other 14-3-3 isoforms,

which is in accordance with previous reports showing

heterodimerization among different 14-3-3 isoforms [1]

Detection of 14-3-3-binding motifs on purified

and identified 14-3-3-binding proteins

The TAP tag approach allows the isolation of native

protein complexes from cells ectopically expressing the

tagged protein of interest, so proteins associated with

14-3-3 proteins were purified and identified in this

study (Tables 1 and 2) Frequently, 14-3-3 proteins

regulate cellular processes by binding to

phosphory-lated motifs (phosphoserine and phosphothreonine)

within target proteins [6], but, because of the

methodo-logical characteristics of the TAP tag approach,

this phosphorylation-dependent binding of identified

proteins is not evident

Two optimal 14-3-3 phosphopeptide ligands with the

consensus sequences [RSX(pS⁄ T)XP and RX(Y ⁄ F)X

(pS⁄ T)XP] have been defined [7], although some 14-3-3

proteins bind to phosphorylated motifs that are

com-pletely different to the consensus sites, or even bind to

unphosphorylated motifs [9] To investigate the

phos-phorylation-dependent binding to 14-3-3 proteins of

the identified proteins, the presence of putative 14-3-3

consensus binding sites was determined for identified

14-3-3f-associated proteins, using the software

scan-site [70] (Table 2) (detailed in Table S3)

Low-strin-gency settings of the scansite algorithm were applied

to analyze 14-3-3-binding consensus motif mode I

[RSX(pS⁄ T)XP] on identified proteins Note that most

proteins studied were identified in normal cell growth

conditions, and lost association with 14-3-3 in the

treat-ments with ceramide To determine whether this

associ-ation was phosphorylassoci-ation-dependent, extracts from

GFP–TAP–14-3-3f HeLa cells were loaded onto an

IgG–agarose chromatography column

Phosphoryla-tion-dependent 14-3-3-binding proteins were eluted

using a phosphopeptide (ARAApSAPA) that competes

with proteins for 14-3-3 binding in a

phosphorylation-dependent manner The data in Fig 6A show desmin

to be a protein eluted from the affinity column To my

knowledge, desmin, a protein that has been shown to

actively participate in the execution of apoptosis [51],

was clearly identified here for the first time as aphosphorylation-dependent 14-3-3-associated proteinunder normal growth conditions, using LC-MS⁄ MS(Table 2) and biochemical validation (Fig 6A) Fur-thermore, the data shown here confirm vasodilator-stimulated phosphoprotein (VASP), nucleophosmin(B23) and calmodulin (CaM), whose 14-3-3 bindingwas suggested in previous studies, as phosphorylation-dependent 14-3-3-associated proteins (Fig 6A)

The data in Fig 6A show vimentin to be a phorylation-dependent 14-3-3-binding protein in con-trol conditions Analysis using the highest-stringencysettings in the scansite algorithm showed Ser39 invimentin to be the most probable 14-3-3-binding site(Table S3) These data support previous findings sug-gesting that 14-3-3 binding of vimentin is a phosphory-lation-dependent mechanism [71] Tuberin [tuberoussclerosis protein 2 (TSC2)], a tumor suppressor proteinthat antagonizes the mTOR signaling pathway,was also found to be a phosphorylation-dependent14-3-3-binding protein These data support previousresults showing that Akt phosphorylation of Ser939 inTSC2 is required for its association with 14-3-3 [72].Both results gave confidence in this technique

On the other hand, the TAP tag approach and phopeptide-specific elution from IgG–agarose chroma-tography columns allows the isolation of native proteinsfrom cells either directly or as components of proteincomplexes To determine whether isolated proteinsundergo direct interactions with 14-3-3, immunoprecipi-tation assays for several isolated apoptotis-related14-3-3-binding proteins were performed Figure 6Bshows VASP and B23 to be phosphorylation-dependent14-3-3-associated proteins that undergo direct interac-tions with 14-3-3 proteins TSC2 also showed a directinteraction with 14-3-3 proteins, supporting previousresults [72], and giving confidence in this technique

phos-Biochemical validation of identified14-3-3-associated proteins related to apoptosisThe combination of TAP and LC-MS⁄ MS allowed iden-tification of 14-3-3-binding proteins from both controlcells and those subjected to C2-ceramide treatments.These data showed a pool of 14-3-3-interactor proteinsinvolved in apoptosis, the 14-3-3-binding pattern beingregulated during C2-ceramide-induced apoptosis(Table 2) Silver staining of a gel loaded with the elutedfractions from TAP purification showed different bands

of 14-3-3-binding proteins between control and mide-induced apoptosis conditions (Fig 7) These dataalso support the idea that C2-ceramide-induced apop-tosis promoted changes in the 14-3-3-binding pattern

C2-cera-Table 2 Selected apoptosis-related associated proteins identified by TAP–MS and immunoblotting analysis Apoptosis-related binding proteins identified in this study by LC-MS ⁄ MS and ⁄ or western blot analysis are listed and grouped by their functions The role of every protein in the apoptotic process is reported References are cited for consensus binding sites (CBSs) for every protein; for the details

14-3-3-of every site found, see Table S3 Underlining indicates proteins that undergo 14-3-3-binding under control conditions but lose this tion after C2-ceramide treatment Nonunderlined proteins bind to 14-3-3 proteins under conditions of C2-ceramide-induced apoptosis.

Chromatin

structure, DNA

binding

121992 Histone H2A.x a [2] Histone H2AX induction occurs only in apoptotic nuclei in cells, and

is implicated in the restoration of genomic integrity in response to DNA double-strand breaks [80]

114762 B23 b,c [4] B23 negatively regulates p53 and antagonizes stress-induced

apoptosis in human normal and malignant hematopoietic cells [76] RNA binding 108935845 Heterogeneous nuclear

ribonucleoproteins C1 ⁄ C2 a

[4] Upregulation of hnRNP C1⁄ C2 during ischemia or staurosporine-induced apoptosis in mice may foster the synthesis of XIAP as a protective pathway against apoptotic effects [95]

Metabolism 4033707 Carbamoyl-phosphate

synthase, mitochondrial precursora

[2] Carbamoyl-phosphate synthase (CPS) is part of a multienzymatic protein (CAD) required for the de novo synthesis of pyrimidine nucleotides and cell growth CAD is a target for caspase-dependent regulation during apoptosis, in this case a fast inactivation of CPS occurs [89]

Cellular

signaling

417101 Histone H1.2 a [1] Histone H1.2 is translocated to mitochondria and associates with

BAK in cells undergoing bleomycin-induced apoptosis Upon DNA damage, histone H1.2 acts as a positive regulator of apoptosome formation, triggering activation of caspase-3 and caspase-7 via APAF-1 and caspase-9 [96–98]

127169 Myosin regulatory

light chain 2 a

[1] Myosin regulatory light chain phosphorylation is critical for apoptotic membrane blebbing and the active morphological changes during apoptosis [90]

108861911 Titin a [9] Titin expression is induced by cyclosporin A via activation of MAPK

pathways, and this may promote proliferation, promote invasion and inhibit apoptosis of human first trimester trophoblasts [91]

49037474 CaMa,b [0] CaM has been shown to regulate apoptosis in tumor models.

CaM-specific inhibitor increased apoptotic cell death with morphological changes characterized by cell shrinkage and nuclear condensation [92]

1717799 TSC2 b,c [17] TSC2 is a tumor suppressor that antagonizes the mTOR signaling

pathway, thus regulating cell growth and proliferation TSC2 activates BAD to promote apoptosis and negatively regulate Bcl-2’s antiapoptotic effects on low serum deprivation-induced apoptosis [99–101]

4506539 RIP1 b [4] RIP1 is a specific mediator of the p38 MAPK response to TNF-a

[94]

205371831 RIP3 b [4] Overexpression studies revealed RIP3 to be a potent inducer of

apoptosis, being capable of selectively binding to large prodomain initiator caspases and attenuating both RIP1 and TNF-a

of proteins from cell extracts Additionally, to validate

the LC-MS⁄ MS data, antibodies against some of the

identified proteins were used to analyze their 14-3-3

interaction in survival or apoptotic conditions (Fig 7)

Thus, a combination of LC-MS⁄ MS and western blot

analysis showed that proteins related to the apoptotic

process, such as RIP1, VASP, and RIP3, were

14-3-3-binding proteins whose association was lost during

C2-ceramide-induced apoptosis Meanwhile, the

cata-lytic unit of DNA-dependent protein kinase

(DNA-PK), a protein with an essential role in DNA

double-strand break repair in the early stages of apoptosis,

raised their 14-3-3-binding status after treatment with

C2-ceramide (Fig 7) Thus, LC-MS⁄ MS and

biochem-ical validation analysis confirms a pool of

apoptosis-related proteins whose 14-3-3-binding status changes

during apoptosis, suggesting an extensive role for

14-3-3 proteins during apoptosis initiation

Stable expression of GFP–TAP–14-3-3f in HeLa

cells delays C2-ceramide-induced cell death

Previous studies have clearly shown that 14-3-3

pro-teins are survival propro-teins with antiapoptotic effects in

cells, by binding to well-known antiapoptotic proteins

and probably also to the apoptosis-related proteins

reported in the present work [37] According to the

effect of 14-3-3f knockdown in sensitizing cells to

C2-ceramide-induced apoptosis, overexpression of

14-3-3f is able to delay cell death promoted by

cera-mide at longest time analysis (36 h) (Fig 8) Thus,

evaluation of mitochondrial membrane potential

changes, using

5,5¢,6,6¢-tetrachloro-1,1¢,3,3¢-tetraethyl-benzimidazolylcarbocyanine iodide (JC-1), clearly

shows a delay in apoptosis induction in those cells that

overexpress 14-3-3f These data support the important

role that 14-3-3 proteins have in C2-ceramide-induced

apoptosis, by binding to well-known apoptosis-relatedproteins, and probably also to other apoptosis-relatedproteins that have been suggested here

DiscussionThe aim of this study was to gain further understand-ing of the role of 14-3-3 proteins in cellular fate,promoting cell survival or inhibiting proapoptotic pro-cesses in cells New aspects are evident from this work:(a) apoptosis-related proteins such as RIP3, BAK anddesmin were identified as new phosphorylation-depen-dent 14-3-3-binding proteins under normal growthconditions; (b) apoptosis-related proteins previouslyidentified by others, using MS⁄ MS analysis, such asRIP1, Smac, STAT3, B23, and CaM, were confirmedhere by immunoblot analysis to be phosphorylation-dependent 14-3-3-associated proteins; (c) C2-ceramide-induced apoptosis promoted decay of the 14-3-3-bindingsignal of proteins in cell extracts; (d) depletion of14-3-3f sensitized cells to C2-ceramide-induced celldeath, whereas overexpression of this isoform delayedcell death; (e) a combination of TAP purification andLC-MS⁄ MS identified 15 proteins involved in cellsurvival processes, their 14-3-3-binding status beingchanged when apoptosis was promoted; and (f) immu-noblot analysis showed that the 14-3-3-binding status

of VASP, RIP1 and RIP3 decayed during inducedapoptosis, whereas the association of DNA-PK with14-3-3 increased during cell death

Several regulators of apoptosis, such as RIP1, Smac,STAT3, and HIP-55, were previously identified as14-3-3-interactor proteins by a combination of 14-3-3-affinity chromatography purification and MALDI-TOF

MS⁄ MS techniques [12] In this study, we identifiedthese proteins as 14-3-3-interactor proteins basically by

MS However, a proper study confirming these data

Table 2 (Continued).

Cellular

organization

55977767 Vimentin a,b [3] Mutations in vimentin disrupt the cytoskeleton in fibroblasts and

delay the execution of apoptosis Cleavage of the p53–vimentin complex enhances TNF-a-related apoptosis-inducing

ligand-mediated apoptosis in fibroblasts [50,52,53]

6686280 Desmin a,b [4] Caspase proteolysis of desmin at Asp263 produces a

dominant-negative inhibitor of intermediate filaments, and actively participates in the execution of apoptosis [51]

1718079 VASP b,c [1] VASP binds to aII-spectrin and this association attenuates aII-spectrin

cleavage during apoptotic cells Cleavage of the plasma membrane-associated spectrins leads to cell shrinkage, membrane blebbing, the

formation of apoptotic bodies, and irreversible cell death [79]

a Identification by LC-MS ⁄ MS b Identification by biochemical approaches c Proteins with low scores identified by LC-MS ⁄ MS.