1. Trang chủ
  2. » Luận Văn - Báo Cáo

Báo cáo khoa học: Human ATP-dependent RNA ⁄ DNA helicase hSuv3p interacts with the cofactor of survivin HBXIP ppt

12 471 0
Tài liệu đã được kiểm tra trùng lặp

Đang tải... (xem toàn văn)

Tài liệu hạn chế xem trước, để xem đầy đủ mời bạn chọn Tải xuống

THÔNG TIN TÀI LIỆU

Thông tin cơ bản

Định dạng
Số trang 12
Dung lượng 474,57 KB

Các công cụ chuyển đổi và chỉnh sửa cho tài liệu này

Nội dung

The C-terminal fragment of hSuv3p that interacts with HBXIP is important for hSuv3p mitochondrial import In order to verify whether deletion of the C-terminal fragment of hSuv3p could ha

Trang 1

interacts with the cofactor of survivin HBXIP

Michal Minczuk1, Seweryn Mroczek1, Sebastian D Pawlak1,* and Piotr P Stepien1,2

1 Department of Genetics, University of Warsaw, Pawinskiego 5A, Warsaw, Poland

2 Institute of Biochemistry and Biophysics, Polish Academy of Sciences, Pawinskiego 5A, Warsaw, Poland

The NTP-dependent RNA⁄ DNA helicase Suv3p

belongs to the Ski2 class of DExH-box RNA helicases

and its orthologues have been found in bacteria, yeast,

plants and animals [1] The product of the SUV3 gene

was described for the first time in Saccharomyces

cere-visiae [2], where it functions in mitochondrial RNA

surveillance Yeast Suv3p is one of the two subunits of

a protein complex called mitochondrial degradosome

or MtEXO, which displays an NTP-dependent

exo-ribonucleolytic activity [3–5] The second component

of the degradosome is Dss1p, a single-strand specific

exoribonuclease with motifs similar to bacterial

RN-ase II [6] The RNA-degrading activity of the

degrado-some complex is necessary for maintaining proper

mitochondrial RNA metabolism in yeast Mutations in either of the two degradosome subunits result in over-accumulation of excised group I introns [7], distur-bances in processing at 5¢ and 3¢ ends of mtRNA precursors, lack of mitochondrial translation and in changes in steady-state levels of mature mt mRNAs; yeast strains bearing deletions of SUV3 or DSS1 genes are respiratory incompetent but are viable on ferment-able carbon sources [4,8,9]

In contrast to yeast, much less is known about the human SUV3 and its physiological functions Our recent report indicated that the hSUV3 exhibits typical characteristics for a nuclear-encoded mitochondrial gene, which is constitutively expressed [10] The human

Keywords

apoptosis; helicase; intracellular localization;

mitochondria; mitochondrial import

Correspondence

M Minczuk, Department of Genetics,

University of Warsaw, Pawinskiego 5A,

02-106 Warsaw, Poland

Fax: +48 22 5922244

Tel: +48 22 5922240

E-mail: mminczuk@ibb.waw.pl

*Present address

Laboratory of Bioinformatics and Protein

Engineering, International Institute of

Molecular and Cell Biology, Ks Trojdena 4a,

02-109 Warsaw, Poland

(Received 4 July 2005, revised 5 August

2005, accepted 10 August 2005)

doi:10.1111/j.1742-4658.2005.04910.x

The human SUV3 gene encodes an NTP-dependent DNA⁄ RNA DExH box helicase predominantly localized in mitochondria Its orthologue in yeast is a component of the mitochondrial degradosome complex involved

in the mtRNA decay pathway In contrast to this, the physiological func-tion of human SUV3 remains to be elucidated In this report we demon-strate that the hSuv3 protein interacts with HBXIP, previously identified as

a cofactor of survivin in suppression of apoptosis and as a protein that binds the HBx protein encoded by the hepatitis B virus Using deletion analysis we identified the region within the hSuv3 protein, which is respon-sible for binding to HBXIP The HBXIP binding domain was found to be important for mitochondrial import and stability of the Suv3 protein

in vivo We discuss the possible involvement of the hSuv3p–HBXIP inter-action in the survivin-dependent antiapoptotic pathway

Abbreviations

aa, amino acids; IAP, inhibitor of apoptosis; BIR, baculovirus IAP repeat; FITC, fluorescein isothiocyanate; GFP, green fluorescence protein;

HA, hemagglutinin; hSUV3, human SUV3; IVT, in vitro translation; TAP, tandem affinity purification.

Trang 2

Suv3p enzyme expressed in Escherichia coli had a

strong preference for a double stranded DNA, while

also displaying an NTP-dependent RNA helicase

activ-ity [11] Recently, Shu et al [12] confirmed the multiple

substrate unwinding activity of the human Suv3

pro-tein, being able to unwind DNA, RNA and

hetero-duplex substrates The unwinding reaction was found

to depend on conformational change of the protein

induced by pH The human Suv3p has a bona fide

mitochondrial leader sequence and using

immunofluo-rescence analysis, an in vitro mitochondrial uptake

assay and subfractionation of human mitochondria we

showed that hSuv3p is a soluble protein localized in

the mitrochondrial matrix [11] However, intracellular

localization studies using polyclonal antibodies, raised

against the heterologously expressed protein, revealed

that in HeLa cells endogenous hSuv3p also exhibits a

faint nuclear localization signal in addition to a strong

mitochondrial signal [11]

Interestingly, recent data have shown that a fraction

of the Suv3p helicase may indeed be localized in the

nucleus Such a suggestion was made by Bader [13],

who employed in silico analysis of the network of

yeast protein–protein interactions Employing this

method he identified yeast Suv3p as a potential

mem-ber of proliferating cell nuclear antigen-like complex

In addition, high-throughput analysis of yeast

pro-tein–protein interactions has revealed several nuclear

protein partners of the yeast Suv3p, most of them

being involved in DNA replication, repair and

recom-bination [14] Among the identified Suv3p interactors

the following proteins have been reported: (a) the

SGS1 helicase, involved in maintaining genome

stabil-ity, homologous to E coli RecQ and human WRN

helicase (defective WRN helicase leads to premature

aging disorder Werner syndrome); (b) the RFC4

pro-tein, a DNA binding ATPase that acts as a processivity

factor for DNA polymerase delta and epsilon and

loads proliferating cell nuclear antigen; (c) MEC3

pro-tein, involved in checkpoint control and DNA repair;

and (d) DDC1 protein, involved in the DNA damage

checkpoint In agreement with above observations Shu

et al [12] have recently suggested the nuclear

localiza-tion of a fraclocaliza-tion of cellular human Suv3p, but no

data were presented The authors proposed that

hSuv3p has multiple physiological roles in the cell,

including telomere maintenance, DNA repair and cell

cycle checkpoint control

In this paper we show the results of the yeast

two-hybrid system in screening for interactors of the

human SUV3 gene product We demonstrate that

hSuv3p interacts with HBXIP, which was previously

identified as a cofactor of survivin in apoptosis

suppression and as a protein binding to the hepatitis B viral protein X

Results

Identification and characterization of the HBXIP– hSuv3p interaction in the two-hybrid system

In order to screen the cDNA library derived from HeLa cells in two-hybrid system as described by Finley

& Brent [15] we constructed baits by linking N-ter-minal or C-terN-ter-minal part of hSuv3p (residues 1–479 and 380–786, respectively) to the LexA DNA binding domain LexA-hSuv3p 380–786 fusion was chosen for further two-hybrid experiments after our initial tests have shown the lack of its self-activation ability, proper nuclear import and operator binding ability in yeast cells (supplementary Appendix S1, Fig S1) The

6· 106 library clones were screened and 57 positive yeast colonies were identified and subjected to sequen-cing Among 57 positive colonies 18 appeared to be independent clones and HBXIP (HBx interacting pro-tein) proved to be the most frequently occurring cDNA among the isolates (seven out of 18 of the hSuv3p interacting clones; sequence characterization of the clones is shown in supplementary Appendix S1, Fig S2) In order to rule out the possibility of nonspe-cific interaction of HBXIP, different nonrelevant baits including bicoid, CD4 and IC-LexA fusion proteins [15] were tested with all positive interacting cDNA clones In order to identify and partially characterize the hSuv3p domain interacting with HBXIP prey clones, several deletion mutants of the C-terminal hSuv3p bait were used in the two-hybrid test As depicted in Fig 1A, the hSuv3p fragment necessary for interaction with HBXIP is contained within the

136 C-terminal amino acids of hSuv3p (amino acids 650–786)

HBXIP interacts with hSuv3p in vitro

We employed an in vitro binding test in order to exclude the possibility that the interaction between the HBXIP and hSuv3p proteins occurred through a yeast-derived bridging protein(s) and to provide evi-dence of direct binding of the proteins in a different system We constructed HBXIP fusion protein contain-ing TAP tag [16] at the C-terminus We purified HBXIP-TAP fusion on IgG-agarose resin after hetero-logous expression in E coli and studied the interaction with an in vitro translated (IVT) [35S]methionine labe-led hSuv3p In this assay two versions of hSuv3p were used: full-length protein (hSuv3p 1–786) and a protein

Trang 3

lacking the 136 C-terminal amino acids (hSuv3p 1–

650); both of them contained a c-myc epitope-tag at

the C-terminus As illustrated in Fig 1B only IVT of

full-length hSuv3p interacted with HBXIP-TAP

immo-bilized on IgG-agarose No interaction was detected

in the case of heterologously purified TAP-tag or the

IgG-agarose resin only (Fig 1B) This result is consis-tent with the finding that the 136 amino acid-long C-terminal part of hSuv3p is responsible for forming a complex with the HBXIP protein

HBXIP shows nucleo-cytosolic localization

in human cells The exact subcellular localization of HBXIP has not been studied up to date To address this issue we tried

to develop the anti-HBXIP polyvalent antibodies We purified HBXIP-TAP fusion in the two step procedure

as described by Rigaut et al [16] after heterologous expression in E coli but we failed to obtain high-affin-ity antibodies against the small hydrophobic HBXIP protein in rabbit Therefore, to determine the subcellu-lar distribution of HBXIP, the protein was C-termin-ally tagged with c-myc or HA epitope and its subcellular localization was studied in transiently transfected HeLa cells The cells were stained with the primary anti-myc (or anti-HA) monoclonal antibodies and visualized with fluorescein isothiocyanate (FITC)-conjugated secondary antibodies In addition, the cells were stained with nuclear marker (DAPI) and mitoch-ondrial marker (MitoTracker CMXRos) As presented

in Fig 2A,B for HeLa cells the HBXIP-myc fusion showed double nucleo-cytosolic localization and practi-cally no colocalization with mitochondria was observed The same result was obtained for the HBXIP-HA fusion and in the case of simian COS-1 cells (data not shown)

Next, in order to provide further evidence on intra-cellular localization of HBXIP, a subintra-cellular fraction-ation experiment was performed As depicted in Fig 2C the HBXIP protein was confirmed to reside in the cytosolic fraction of HeLa cells and it was not found in the mitochondrial fraction

The C-terminal fragment of hSuv3p that interacts with HBXIP is important for hSuv3p mitochondrial import

In order to verify whether deletion of the C-terminal fragment of hSuv3p could have an effect on protein function in vivo we transiently expressed C-terminally truncated (136 amino acids) hSuv3p (hSuv3p-myc 1– 650) in COS-1 cells First, we analyzed the subcellular distribution of the truncated mutant protein using myc monoclonal antibodies visualized with anti-mouse secondary antibodies conjugated with FITC In the case of hSuv3p-myc 1–650 fusion colocalization with the mitochondrial marker MitoTracker was substantially reduced, as compared to the wildtype

A

B

Fig 1 Interaction of hSuv3p with HBXIP (A) Schematic

representa-tion of the hSuv3p baits used in this study with the relative binding

affinities to HBXIP prey clone In order to screen the HeLa

cell-derived cDNA library by the yeast two-hybrid screening, stable bait

was generated by fusing the LexA DNA binding domain with the

C-terminal part of hSuv3p (hSuv3p 380–786) The yeast bait vector

carrying the full-length hSuv3p was also tested under the same

experimental conditions Several hSuv3p deletion mutants were

used to specifically identify and partially characterize the hSuv3p

domain that is necessary for interaction with HBXIP (B) In vitro

interaction of hSuv3p with HBXIP The SDS ⁄ PAGE analysis of the

binding of wildtype hSuv3p or its mutant devoid of the C-terminal

136 amino acids (hSuv3p 1–650) to IgG-agarose-HBXIP is shown.

Both variants of hSuv3p have been synthesized using IVT in the

presence of [ 35 S]methionine Lanes 1 and 2 show hSuv3p (WT)

and the hSuv3p 1–650 mutant, respectively, bound to HBXIP-TAP

immobilized on IgG-agarose Lanes 3 and 4 show the binding of

hSuv3p (WT) and the hSuv3p 1–650 mutant, respectively, bound to

IgG-agarose resin alone Lanes 5 and 6 show the binding of hSuv3p

(WT) and the hSuv3p 1–650 mutant, respectively, to heterologically

expressed TAP-tag immobilized on IgG-agarose Lanes 7 and 8

show 20% of the input of the IVT [ 35 S]methionine labeled hSuv3p

(WT) and the hSuv3p 1–650 mutant, respectively.

Trang 4

hSuv3-myc protein expressed under exactly the same conditions, and a significant fraction of the protein was retained in the cytosol (Fig 3) It therefore appears that, in addition to the N-terminal mitochond-rial leader peptide, the 136 amino acid-long C-terminal part of hSuv3p is involved in targeting of the protein

to mitochondria It is worth mentioning, however, that the 136 amino acid-long C-terminal part of hSuv3p alone cannot serve as a bona fide mitochondrial target-ing signal because hSuv3p 650–786-TAP protein con-struct (see below) is not localized in mitochondria (data not shown)

The C-terminal fragment of hSuv3p that binds HBXIP is important for hSuv3p stability The experiment described above also showed that the truncated form of hSuv3p, lacking the domain respon-sible for HBXIP binding, was expressed in a signifi-cantly lower number of cells as compared to the wildtype form of the protein The transfection effi-ciency was  21% and 1.5% for full-length and trun-cated hSuv3p, respectively (although the transfection conditions and plasmid DNA preparations were the same in both cases) Such a difference could be the result of mRNA instability, accelerated protein degra-dation or cell toxicity of the truncated hSuv3 protein

In order to discriminate between those possibilities we measured the mRNA steady-state levels by Northern hybridization As illustrated in Fig 4A there was no significant difference in mRNA level for hSuv3p wild-type and truncated construct In order to exclude that lowered expression frequency observed for the trun-cated hSuv3p results from cellular toxicity, the wild-type and 1–650 forms were coexpressed with green fluorescence protein (GFP) as a internal marker (GFP was encoded within the vector backbone, therefore all cells expressing either form of hSuv3p expressed GFP

as well) First, we measured the transfection efficiency for the constructs encoding either version of hSuv3p

by counting GFP-positive cells; in both cases the effi-ciency was  17% Then, after transient expression, wildtype and truncated forms of hSuv3p, the proteins were visualized by anti-myc monoclonal antibodies and anti-mouse Texas Red conjugated secondary antibod-ies Furthermore, we studied the correlation between red (hSuv3p-derived) and green (GFP-derived) fluores-cence for full-length and truncated forms of hSuv3p as described in Experimental procedures As presented in Fig 4B,C the correlation observed was 84 ± 12% and

9 ± 6% for wildtype hSuv3p and hSuv3p 1–650 trun-cated protein, respectively This result inditrun-cated that the hSuv3p version lacking HBXIP binding domain

A

B

C

Fig 2 Intracellular localization of HBXIP in mammalian cells (A)

HeLa cells were grown on coverslips and transiently transfected

with cDNA encoding c-myc tagged HBXIP (HBXIPmyc) After

incu-bation with MitoTracker red, fixation and permeabilization as

des-cribed in Experimental procedures the cells were immunostained

with anti-(c-myc) monoclonal antibody (9E10), which was then

visu-alized with fluorescein isothiocyanate-conjugated antibody At the

final stage, nuclei were stained with DAPI present in the mounting

medium The figure shows representative fluorescent image of

cells stained with DAPI (blue), c-myc-tagged HBXIP (green) and

MitoTracker (red) taken by a confocal microscope Similar results

were obtained for cDNA encoding HA tagged HBXIP expressed in

HeLa cells as wells as for the HBXIPmyc and HBXIP-HA expressed

in COS-1 cells (B) Magnified confocal microscope images prepared

as in (A) for HeLa cells transfected with cDNA encoding HBXIPmyc

are shown (C) HeLa cells were transiently transfected with cDNA

encoding c-myc tagged HBXIP (HBXIPmyc) Cell lysates from

unfractionated HeLa cells (T), the cytoplasmic (C) and the

mitoch-ondrial fraction (M) were immunoblotted with anti-myc monoclonal

antibodies The fractions were verified using anti-hSuv3p serum

used here as a marker for the mitochondrial fraction.

Trang 5

might be less stable in comparison to wildtype hSuv3p

protein

In order to address the question of whether the

hSuv3p 1–650 truncated protein is less stable than the

wildtype form in vivo, protein synthesis was inhibited in

transfected cells by cycloheximide and protein decay

rates were measured at various time points thereafter

In these studies, COS-1 cells were transiently transfected

with expression vectors encoding either TAP-tagged

wildtype hSuv3p or the 1–650 truncated form of

hSuv3p At 24 h following the transfection, the protein

synthesis was inhibited with cyclohexamide treatment

and total cell extracts were prepared at 0, 2, 4, 6 or 8 h

after cycloheximide addition The protein levels at

var-ious time points were analyzed by western blot using

the PAP antibody (i.e antibody against protein A

which is encompassed within the TAP tag) to measure

protein decay rates As presented in Figure 5A the

TAP-taged hSuv3p 1–650 form lacking the protein

frag-ment responsible for hSuv3p–HBXIP interaction was

significantly less stable as compared to the wildtype

hSuv3TAP The absence of HBXIP binding domain

resulted in 50% decline in the truncated protein levels

within 4 h following inhibition of protein synthesis by

cycloheximide We next wanted to examine whether the presence of HBXIP binding protein fragment of the hSuv3 protein by itself can increase the protein stability For these studies the protein decay rates were measured for the COS-1 cells transfected with the expression vec-tors encoding the HBXIP binding domain of hSuv3p fused to TAP (construct hSuv3TAP 650–786; Fig 5B)

or the TAP protein alone Figure 5B illustrates that the protein stability of TAP fused to the HBXIP binding domain of the hSuv3p protein is much higher in com-parison to TAP alone These results suggest that the hSuv3p C-terminal fragment (residues 650–786), which was also found to bind HBXIP, plays an important role

in regulation of the hSuv3 protein stability

Discussion

The data presented in this paper indicate that human hSuv3p helicase interacts with HBXIP protein The human HBXIP is a small protein of 91 amino acids that was discovered by Melegari et al [17], as the result of yeast two-hybrid screening for the interactors of hepati-tis B virus-encoded protein HBx The viral HBx protein seems to be the major cause of hepatocarcinogenesis

Fig 3 Localization of the truncated form of hSuv3p lacking the HBXIP interacting C-terminal domainCOS-1 were grown on coverslips and transiently transfected with cDNA encoding c-myc tagged hSuv3p (hSUV3myc WT) or the hSuv3p mutant lacking the C-terminal 136 amino acids (hSuv3myc 1–650) After incubation with MitoTracker red, fixation and permeabilization the cells were immunostained with anti-(c-myc) monoclonal antibody, which was then visualized with fluorescein isothiocyanate-conjugated antibody (9E10) Fluorescent images of mito-chondria stained with MitoTracker (red) and c-myc-tagged variants of hSuv3p (green) were taken by a confocal microscope Colocalization of either forms of hSuv3p with mitochondria appears in yellow ⁄ orange in digitally overlaid images.

Trang 6

[18] and is a multifuctional regulator of transcription,

cell responses to genotoxic stress, protein degradation

and signaling pathways [19] Recent data indicate that

HBx localizes in mitochondria, and its overexpression

induces a perinuclear mitochondrial distribution and

loss of a mitochondrial membrane potential [20,21]

Studies with the mutant HBx proteins revealed that its

mitochondrial targeting sequences are important for

mitochondrial localization, mitochondrial membrane

potential disruption and cell death [21,22]

Further-more, it has been reported that HBx interacts with at

least two mitochondrial proteins: (a) VDAC3, which is

confined to the outer mitochondrial membrane [22];

and (b) the heat shock protein 60 predominately

locali-zed in the mitochondrial matrix [23] However, the

exact physiological significance of the

intramitochond-rial localization of HBx and of the HBx–HBXIP

inter-action remains unknown

Recently it has been shown that HBXIP is a neces-sary cofactor of survivin in the process of suppression

of apoptosis in cancer cells Survivin is a small protein (16.5 kDa) that contains N-terminal zinc binding bacu-lovirus inhibitor of apoptosis repeat (BIR) domain linked to a C-terminal amphipathic helix [24] Under normal physiological conditions survivin is involved in coordinating the chromosomal and cytoskeletal events

of mitosis [25] In most cancer cells survivin is strongly up-regulated, forms a complex with HBXIP and inhibits apoptosis The mechanism of the inhibition is mediated

by binding of the survivin-HBXIP to Apaf1 and pre-venting the activation of procaspase 9 [26] Thus, HBXIP has an important function in apoptosis suppres-sion The siRNA inhibition of either survivin or HBXIP results in restoration of apoptotic ability of cancer cells

To the best of our knowledge no other interactors

of the HBXIP have been reported, nor has its exact

A

B

C

Fig 4 Expression of hSuv3p lacking the HBXIP interacting domain A Northern blot analysis of the steady-state levels of the hSuv3myc and hSuv3myc 1–650 mRNA COS-1 were transiently transfected with the pcDNA3.1(–) vector (lane 1), cDNA encoding c-myc tagged, wildtype form of hSuv3p (lane 2, expected transcript length 2690 nt) or the hSuv3p mutant lacking the C-terminal 136 amino acids (lane 3, expected transcript length 2282 nucleotides) Total RNA from the cells was isolated and subjected to the northern blot analysis as described in the Experimental procedures In the conditions applied the hybridization signal corresponding to the endogenous hSUV3 mRNA is not visible (lane 1) (B) Coexpression of GFP with either the c-myc tagged wildtype hSuv3p or the hSuv3p mutant lacking the C-terminal 136 amino acids COS-1 were grown on coverslips and transiently transfected with cDNA encoding c-myc tagged hSuv3p (hSUV3myc WT) or the hSuv3p mutant (hSuv3myc 1–650) After fixation and permeabilization the cells were immunostained with anti-(c-myc) monoclonal antibody, which was then visualized with TexasRed-conjugated antibody The panel shows representative fluorescent images of c-myc-tagged variants

of hSuv3p (red) coexpressed with GFP (green) (C) Quantitative analysis of the correlation between the expression of the hSuv3p variants and GFP Black columns represent the correlation between hSuv3p-derived (red column) and GFP-derived (green column) fluorescence in the cells coexpressing GFP with either the wildtype form of hSuv3p (hSuv3myc WT) or the truncated version lacking of the C-terminal 136 amino acids (hSuv3p 1–650) calculated as described in Experimental procedures from the three independent experiments.

Trang 7

subcellular localization been analyzed In this work the

fragment of the hSuv3 protein encompassing 380–786

of its total 786 amino acids has been used as the bait

in a yeast two-hybrid system Out of 18 bona fide

clones representing interacting human proteins, seven

clones were found to encode the HBXIP protein The

results of our two-hybrid screen have been confirmed

by pull-down assays

We constructed a series of deletions of the hSUV3

cDNA and demonstrated that only the 136 amino acid

long C-terminal domain of the hSuv3 protein is

responsible for the observed interaction with HBXIP

This domain has no obvious homology to the domain

within the hepatitis B virus protein X, which is known

to bind HBXIP as well [17] Nevertheless, the 136

amino acid domain seems to be of importance for

functioning of the hSuv3 protein Its deletion not only

abolishes interactions with HBXIP, but leads to

delo-calization of the hSuv3 protein: a significant portion of

it has been found in the cytosol In addition, the

trun-cated hSuv3 protein is less stable

Initially we assumed that because hSuv3p was

shown to be a mitochondrial protein, the site of the

discovered hSuv3p–HBXIP interaction should be a

mitochondrion In contrast to this, our data on

subcellular distribution of HBXIP indicated mainly cytosolic or nuclear localization Therefore, the inter-action of hSuv3p with HBXIP in vivo might occur out-side mitochondria, for instance: (a) in the cytosol before Suv3p translocates through mitochondrial mem-branes or (b) in the nucleus, where a fraction of hSuv3p has been recently suggested to reside [12] This result is in agreement with studies of Marusawa et al [26], which suggested that the HBXIP–survivin interac-tion is not localized in mitochondria On the other hand, owing to limited detection limits of the methods used by both us and others, it cannot be excluded that

a vary small amount of HBXIP is localized in mito-chondria Another possibility is that HBXIP may change its cellular localization and, for example, could

be translocated to mitochondria in certain physiologi-cal conditions

What could be the physiological significance of the observed hSuv3p–HBXIP interaction? Two hypotheses can be put forward First, HBXIP can serve as a chap-erone for Suv3p, necessary for its proper import into mitochondria after being translated on cytosolic ribo-somes Because a fraction of truncated Suv3p, i.e lack-ing the C-terminal HBXIP bindlack-ing domain, can be found in the cytosol, the binding of HBXIP may

Fig 5 The role of the C-terminal, HBXIP interacting, domain of hSuv3p in protein stability (A) The protein stability of the wildtype hSuv3TAP and the hSuv3TAP variant lacking the C-terminal 136 amino acids (hSuv3TAP 1–650) in mammalian cells COS-1 cells were transfected with pcSUVTAP or pcSUVTAP-1–650 and after 24 h the protein synthesis was inhibited by addition of cycloheximide to the culture medium The protein levels were examined by Western blot in the indicated time points (B) The protein stability of the fusion protein containing the C-ter-minal 136 amino acids of hSuv3p fused to TAP (hSuv3TAP-650–786) and TAP alone in mammalian cells COS-1 cells were transfected with pcSUVTAP-650–786 or pcTAP and treated as described in (A) The graphs illustrate the densitometric quantification of the amounts of pro-tein for a representative experiment presented as a percentage of propro-tein in the time point 0 h ‘P’ and ‘M’ indicates the precursor and mature mitochondrial form of hSuv3p, respectively The asterisk indicates an unidentified degradation product.

Trang 8

promote mitochondrial localization of hSuv3p

Analy-sis of recent reports on mitochondrial localization⁄

function of HBx [27,28] suggests that the interaction of

HBx and HBXIP might also be necessary for the

import of HBx into mitochondria The region within

HBx responsible for the binding of HBXIP [17]

coin-cides with the domain necessary for the mitochondrial

localization of HBx (supplementary Appendix S1,

Fig S3) Therefore, this hypothesis would assume

similarity with the functions of HBXIP in transport of

viral hepatitis B protein X into mitochondria It

should be stressed, however, that the HBx terminal

domain is only one of the determinants of

mitochond-rial localization: similarly as for hSuv3p, the

N-term-inal leader sequence is required for the import as well

The chaperone hypothesis also seems to be in

agree-ment with our data that the hSuv3 protein devoid of

the 136 amino acid C-terminal domain is significantly

less stable This in turn is consistent with the report by

Zhao et al [29], which shows that mutations of critical

amino acid residues within the BIR domain of

survi-vin, which is responsible for interaction with HBXIP

[26], sensitize survivin to degradation

Our second hypothesis assumes that the interaction

of hSuv3p and HBXIP plays a role in the suppression

of apoptosis by the survivin–HBXIP complex

Accord-ingly, hSuv3p would interact with this complex, which

prevents binding of Apaf1 to procaspase 9 [26]

Because HBXIP was shown to be a necessary cofactor

in this process, by binding to survivin, the ability of

hSuv3p to interact with HBXIP would constitute an

important regulatory mechanism in apoptosis

suppres-sion in cancer cells Our preliminary data indicate that

this possibility cannot be excluded, as siRNA

inhibi-tion of hSUV3 in HeLa cells resulted in apoptosis (A

Dmochowska, unpublished data, Warsaw, Poland)

Interestingly, recent reports have shown that survivin

also localizes in mitochondria, and in response to cell

death stimulation, the mitochondrial pool of survivin

is displaced into cytosol, where it prevents casapase

activation [30,31] Such trafficking of the proteins in

and out of mitochondria might constitute an important

element in apoptosis control It is clear that more

research is needed to test the involvement of hSUV3 in

this pathway and our experiments are in progress

Experimental procedures

Plasmid construction

The bait plasmids, used in the two-hybrid screen, encoded

LexA DNA binding domain fused to various fragments

of hSuv3p and were constructed using pEG202 [15] as

described below The schematic representation of all the bait fusion proteins is shown in Fig 1A Numbers in the LexA fusion names correspond to amino acid positions in the hSuv3 protein fragments All enzymes used for cloning were purchased from Fermentas (Vilnius, Lithuania) The pEGhSUV3-1–479 plasmid encoding the LexA-hSuv3p 1–479 fusion was constructed by PCR amplification

of the appropriate hSuv3p cDNA fragment using the fol-lowing primers: CCGGAATTCTCGATGTCCTTCTCCC GTGC (forward; incorporating EcoRI site, underlined) and GCGGGATCCGAAACCGTGAGCTGAATCTGCC (reverse, incorporating BamHI site, underlined) The result-ing fragment was cloned into pEG202 usresult-ing EcoRI and BamHI

The pEGhSUV3-380–786 plasmid encoding the LexA-hSuv3p 380–786 fusion was constructed by PCR amplifica-tion of the appropriate hSuv3p cDNA fragment using the following primers: GCGGAATTCTCTGTGAGTCGGCA GATTGAA (forward; incorporating EcoRI site, under-lined) and CATGCCATGGCTAGTCCGAATCAGGTTC

CT (reverse, incorporating NcoI site, underlined) The resulting fragment was cloned into pEG202 using EcoRI and NcoI

The pEGhSUV3-1–786 plasmid encoding the LexA-hSuv3p 1–786 fusion was constructed by PCR amplification

of the appropriate hSuv3p cDNA fragment using the for-ward primer as in case of pEGhSUV3-1–479 and the reverse primer as in case of pEGhSUV3-380–786 The resulting fragment was cloned into pEG202 using EcoRI and NcoI

The pEGhSUV3-380–786D393–506 plasmid encoding LexA-hSuv3p 380–786D393–506 fusion was constructed by excision of the PvuII-PvuII form pEGhSUV3-380–786 and religation

The pEGhSUV3-380–735, pEGhSUV3-380–650 and pEGhSUV3-380–580 plasmids encoding the LexA-hSuv3p 380–735, 380–650 and 380–580 fusions, respectively, were constructed by PCR amplification of the appropriate hSuv3p cDNA fragments using the forward primer as in case of pEGhSUV3-380–786 and the following reverse primers: CCATCCATGGCTAGGAAGCAAGGGACAGC TCTCC, GGATCCATGGTCATGGAAACATATCCATA AATCGG and CCATCCATGGTCAGTTGATAGGAGC TGTGAAGAAAAC, respectively (all incorporating NcoI site, underlined) The resulting fragments were cloned into pEG202 using EcoRI and NcoI

The pEGhSUV3-650–786 and pEGhSUV3-650–735 plas-mids encoding LexA-hSuv3p 650–786 and 650–735 fusions, respectively, were constructed by PCR amplification of the appropriate hSuv3p cDNA fragments using the following forward primer CCTGAATTCGATGCCAGCCTTATTCG AGATCTCC (EcoRI site underlined) and the reverse prim-ers as in the case of 380–786 and pEGhSUV3-380–735, respectively The resulting fragments were cloned into pEG202 using EcoRI and NcoI

Trang 9

The pcHBXIPmyc and pcHBXIP-HA constructs used for

immunoflourescence analysis that encode HBXIP fused to

C-terminal epitope tags c-myc and HA, respectively, were

constructed as follows: the cDNA fragment encoding

HBXIP of full-length was PCR amplified using the

follow-ing reverse primers: for pcHBXIPmyc CCATAAGCTTCA

CAGGTCCTCCTCGGAGATCAGCTTCTGCTCAGAGGC

CATTTTGTGCACTGCC introducing c-myc epitope

cod-ing sequence (italic) and HindIII site (underlined); for

pcHBXIP-HA CCATAAGCTTCAGAGGCTAGCGTAATC

CGGAACATCGTATGGGTAAGAGGCCATTTTGTGCAC

TGCC introducing HA epitope coding sequence (italic) and

HindIII site (underlined) In both cases the forward BCO1

primer (CCAGCCTCTTGCTGAGTGGAGATG) was used,

which binds upstream of the multiple cloning site (MCS) in

the cDNA library pJG4-5 plasmid [15] The 3–54 clone

(supplementary Appendix S1, Fig S2) selected from the

cDNA library in the yeast two-hybrid system was used as a

template for both constructs The resulting fragment was

cloned into EcoRI and HindIII sites of pcDNA3.1(–) vector

(Invitrogen, Carlsbad, CA, USA)

The pET15HBXIP-TAP construct used for

overexpres-sion of the HBXIP-TAP fuoverexpres-sion in E coli was constructed

as follows: the BamHI and NcoI fragment encoding

TAP-tag was excised from pBS1539 [16] and inserted into the

pET15b bacterial expression vector (Novagen, Madison,

WI, USA) The resulting plasmid was named pET15TAP

and expressed TAP-tag only Then the fragment encoding

HBXIP was PCR-amplified using the 3–54 clone template

(supplementary Appendix S1, Fig S2) and the following

primers: CGATCCATGGAGGCGACCTTGGAGCAG

(forward) and GACTCCATGGAGGCCATTTTGTGC

ACTG (reverse), both incorporating NcoI site The

obtained PCR fragment was cloned into pET15TAP using

NcoI site

The pchSUV3myc plasmid used for expression of

wild-type hSuv3p in a c-myc-tagged form in mammalian cells

was as described previously [11] The pchSUV3-1–650myc

construct encoding hSuv3p lacking the 136 C-terminal

amino acids with a c-myc epitope (named hSuv3p 1–650)

was constructed as follows: a DNA fragment encoding the

first 650 amino acids of hSuv3p was PCR amplified using

the following primers: GCATCTAGACACGATGGCCTT

CTCCCGTGCCCTATTGTGG (forward) introducing XbaI

site (underlined) and CGTGAATTCACAGGTCCTCCTCG

GAGATCAGCTTCTGCTCTGGAAACATATCCATAAAT

CGGTAGC (reverse) introducing c-myc epitope coding

sequence (italic) and EcoRI site (underlined) The

pchSUV3myc (see above) plasmid served as a template The

resulting fragment was cloned into XbaI and EcoRI sites of

the pcDNA3.1(–) vector (Invitrogen)

The pTRhSUV3myc and pTRhSUV3-1–650myc

con-structs used for coexpression of the wildtype form of

hSuv3p or the hSuv3p 1–650 mutant with GFP were

con-structed by subcloning of the NheI-EcoRI fragments from

pchSUV3myc and pchSUV3-1-650myc, respectively, into pTRACER CMV⁄ Bsd (Invitrogen)

In order to obtain the pchSUV3TAP plasmid, encoding the full-length hSuv3p as a C-terminal fusion with TAP-tag, the hSUV3 cDNA was amplified using the following primers: TACCCATGGGCATCTGCTCTGCCCTTCG – forward and CATGCCATGGCTAGTCCGAATCAGGT TCCT – reverse, both incorporating NcoI site (underlined) The resulting fragment was cloned into NcoI site of the pET15TAP plasmid (see above) Then the BamHI-BamHI fragment was subcloned into the pchSUV3myc vector The pchSUV3TAP-1–650 plasmid encoding the truncated form of hSuv3p (lacking 136 C-terminal amino acids) fused

to TAP-tag was constructed as described for pchSUV3TAP with the exception of using the following reverse pri-mer: GGATCCATGGTCATGGAAACATATCCATAAA TCGG

The pchSUV3TAP-650–786 plasmid encoding the C-ter-minal part of hSuv3p as a fusion with TAP-tag was obtained by the PCR amplification of the appropriate cDNA fragment from the pchSUV3TAP plasmid with the following primers: CCTCTCGAGATGGATGCCAGCCTT ATTCGAGATCTCC – forward (incorporating XhoI site, underlined) and GCTGAATTCTCAGGTTGACTTCCCC GCGGAGTTCG – reverse (incorporating EcoRI site, underlined) The resulting fragment was cloned into the pcDNA3.1(–) vector Please note: letter in boldtype in the reverse primer represents the AfiG mutation introduced in order to disrupt the EcoRI site present in the original TAP sequence

The pcTAP plasmid encoding TAP-tag only was gener-ated similarly to pchSUV3-650–786-TAP with the exception that the forward primer had the following sequence: CGTCTCGAGATGGAAAAGAGAAGATGGAAAAAG AATTTC (XhoI site is underlined)

Two-hybrid screening Before conducting the two-hybrid screening, the LexA-hSuv3p 1–479 and LexA-LexA-hSuv3p 380–735 baits were tested

in order to verify whether the fusion proteins are able to enter the nucleus, bind LexA operators, and not activate transcription of the reporter genes by themselves The test was performed as described previously [15] Additionally, it was verified by immunoblotting with the anti-hSuv3p serum described in [11] whether the full-length bait proteins were made by the yeast cells transformed with the pEGhSUV3-1–479 or pEGhSUV3-380–786 bait plasmids

The yeast two-hybrid screening was performed according

to the sequential method described previously [15] with a HeLa-derived cDNA library cloned into the yeast pJG4-5 shuttle vector [32] Briefly, the EGY48 yeast strain (contain-ing LEU2 reporter gene) was transformed with the HIS pSH18-34 LacZ reporter plasmid [15] and the URA pEG-hSUV3-380–786 bait plasmid (this work) Then the TRP

Trang 10

library plasmids were transformed into this strain using

high-efficiency transformation as described previously [15]

and the transformants ( 3 · 107

) were counted, harvested from the 24· 24 cm plates (Nunc, Wiesbaden, Germany)

and frozen for storage at )70 C Aliquots of the

library-transformed pellets were thawed and plated onto selective

medium (containing galactose and lacking leucine –Gal⁄ Raf

ura-his-trp-leu-) following 4 h of amplification In the next

step, single yeast resistant colonies were replicated onto the

following media: Gal⁄ Raf trp-leu-, Gal ⁄ Raf

ura-his-trp- X-gal, Glu ura-his-ura-his-trp-leu- and Glu ura-his-ura-his-trp-X-gal

Galactose dependent LEU+ blue colonies were identified

and subjected to a plasmid DNA isolation procedure as

described previously [15] Library cDNA inserts were then

PCR amplified with the BCO1 and BCO2 primers [15] and

subjected to restriction analysis in order to identify

repetit-ive clones, and then sequenced Human proteins encoded

by the library inserts were identified by blastx [33]

Plas-mids of independent clones encoding different clones of

HBXIP (supplementary Appendix S1, Fig S2) were rescued

using the E coli KC8 strain [34] Next, the HBXIP prey

plasmids were retransformed into the yeast EGY48 strain

carrying plasmids encoding nonspecific baits, i.e fusions of

LexA with either bicoid, CD4, CD4D85 or IC [15]

Addi-tionally, one of the HBXIP clones (supplementary

Appen-dix S1, Fig S2, 3–54) was transformed into the EGY48

strain harbouring several deletion mutants of the

hSUV3-380–786 C-terminal bait (Fig 1A) The resulting strains

were tested on the selective media as described above

Protein purification

In order to overexpress the HBXIP-TAP fusion or the

TAP-tag control the E coli BL21-CodonPlus(DE3)-RP

strain (Stratagene, Kirkland, WA, USA) transformed with

pET15HBXIP-TAP or pET15TAP, respectively, was grown

to D600¼ 0.6 and induced for 20 h in 16 C with 1 mm

iso-propyl thio-b-d-galactoside Bacterial pellets were incubated

in the IPP150 buffer without NP40 (10 mm Tris⁄ HCl

pH 8.0, 150 mm NaCl, 1 mm EDTA) supplemented with

proteinase inhibitor cocktail (Roche, Mannheim, Germany),

1 mm phenylmethanesulfonyl fluoride and 100 mm

lyso-syme After the incubation, NP40 was added to 0.1% (w⁄ v)

and the samples were sonicated Insoluble material was

pel-leted (26 000 g), the supernatant was loaded on the

IgG-agarose column equilibrated with IPP150 and the sample

was rotated for 2 h at 4C Following the incubation,

unbound proteins were eluted with IPP150 and a portion of

the resin with bound HB-XIP or TAP-tag was mixed with

the SDS loading buffer and boiled for 5 min The

IgG-agarose immobilized proteins were then resolved using

SDS⁄ PAGE, stained with Coomassie and subjected to

den-sitometry in order to assay the purity of the sample In

addition, the purified proteins were immunobloted and

probed with PAP antibodies (Sigma, Steinheim, Germany)

In vitro protein–protein interaction The in vitro interaction between HBXIP and hSuv3p was studied as follows: the wildtype form of hSuv3p or the hSuv3p 1–650 mutant lacking the C-terminal 136 amino acids were in vitro translated (IVT) in the presence of [35S]Met using the TNT Quick coupled transcription⁄ trans-lation system (Promega, Madison, WI, USA) and the pchSUV3myc or pchSUV3-1–650myc plasmid as a tem-plate The [35S]Met labeled proteins were incubated with purified and IgG-agarose immobilized HBXIP-TAP (or TAP-tag) in 0.1 m phosphate buffer (pH¼ 8.1) for 1 h at

4C After the incubation, the IgG-agarose resin was inten-sively washed with 0.1 m phosphate buffer, mixed with the SDS loading buffer, boiled for 5 min and resolved in the SDS⁄ PAGE gel After electrophoresis the gel was dried and subjected to autoradiography

Immunofluorescence experiments and cell fractionation

For the immunofluorescence studies of HBXIP, hSuv3myc and its hSuv3myc 1–650 mutant form in HeLa or COS-1 the cells were plated in 6-well cluster dishes with a cover slip placed at the bottom of the well and grown overnight

in DMEM (Sigma, St Louis, MO, USA) supplemented with 10% FCS and 4 mm glutamine The cells were then trans-fected using FuGene6 reagent (Roche, Indianapolis, IN, USA) At 24 h after the transfection staining of the mito-chondria and the immunodetection of the tagged proteins was carried out as described previously [11] The primary antibodies against c-myc and HA as well as the secondary anti-mouse antibodies conjugated with fluorescein isothio-cyanate or TexasRed were purchased form Santa Cruz Bio-technology (Santa Cruz, CA, USA) In some experiments, where indicated, cell nuclei were stained with DAPI present

in the Vectashield mounting medium (Vector Laboratories, Burlingame, CA, USA)

In order to study the correlation between expression levels

of the hSuv3p variants and GFP cells were transfected with pTRhSUV3myc or pTRhSUV3-1–650myc and prepared for immunofluorescence analysis as described above Then, for

100 randomly selected cells intensity of red fluorescence, derived form TexasRed conjugated secondary antibody bound to either form of hSuv3p, and green fluorescence, derived from GFP, were calculated using imagej software (W Rosband http://rsb.info.nih.gov/ij) and expressed as rel-ative units The ‘correlation TexasRed vs GFP’ as presen-ted on Fig 4C was obtained by dividing the sum of the red fluorescence by the sum of green fluorescence

For cellular fractionation experiments three to four 6-well cluster dishes of HeLa cells were transfected with pcHBXIP-myc as described above At 24 h after the transfection cell fractionation and isolation of mitochondria on sucrose gra-dient were performed as described previously [11] The

Ngày đăng: 23/03/2014, 15:21

TỪ KHÓA LIÊN QUAN

TÀI LIỆU CÙNG NGƯỜI DÙNG

TÀI LIỆU LIÊN QUAN

🧩 Sản phẩm bạn có thể quan tâm