Báo cáo khoa học: Intrabodies against the EVH1 domain of Wiskott–Aldrich syndrome protein inhibit T cell receptor signaling in transgenic mice T cells potx

Constructs con-taining the VH signal peptide sequence, 18SHL⁄ SHL-CL and 21SHL⁄ SHL-CL, were able to bind to GST-WASP15 see Experimental procedures for defini-tion of WASP15, whereas no b

syndrome protein inhibit T cell receptor signaling in

transgenic mice T cells

Mitsuru Sato1, Ryo Iwaya1,2, Kazumasa Ogihara1,3, Ryoko Sawahata1,3, Hiroshi Kitani1, Joe Chiba2, Yoshikazu Kurosawa4and Kenji Sekikawa1,5

1 Department of Molecular Biology and Immunology, National Institute of Agrobiological Sciences, Ibaraki, Japan

2 Department of Biological Science and Technology, Tokyo University of Science, Chiba, Japan

3 Institute for Antibodies Co., Ltd, National Institute of Agrobiological Sciences, Ibaraki, Japan

4 Institute for Comprehensive Medical Science, Fujita Health University, Toyoake, Aichi, Japan

5 Kitasato University School of Veterinary Medicine and Animal Sciences, Aomori, Japan

Intracellular antibodies (intrabodies) may be useful

tools for not only clinical applications such as viral

neutralization and cancer therapy but also functional

analysis of proteins inside cells A variety of intrabody

formats have been used Single-chain variable frag-ments (scFvs) consist of one heavy chain variable region (VH) linked through a flexible peptide spacer, usually a repeated motif of 3· GGGGS, to one light

Keywords

cytosolic protein; functional knockdown;

intrabody; T-cell receptor signaling; Wiskott–

Aldrich syndrome protein (WASP)

Correspondence

K Sekikawa, Department of Molecular

Biology and Immunology, National Institute

of Agrobiological Sciences, 3-1-5,

Kannondai, Tsukuba, Ibaraki 305-0856,

Japan

Tel ⁄ Fax: +81 29 8386039

E-mail: sekiken@nias.affrc.go.jp

(Received 2 August 2005, revised 4 October

2005, accepted 10 October 2005)

doi:10.1111/j.1742-4658.2005.05011.x

Intracellularly expressed antibodies (intrabodies) have been used to inhibit the function of various kinds of protein inside cells However, problems with stability and functional expression of intrabodies in the cytosol remain unsolved In this study, we show that single-chain variable fragment (scFv) intrabodies constructed with a heavy chain variable (VH) leader signal sequence at the N-terminus were translocated from the endoplasmic reti-culum into the cytosol of T lymphocytes and inhibited the function of the target molecule, Wiskott–Aldrich syndrome protein (WASP) WASP resides in the cytosol as a multifunctional adaptor molecule and mediates actin polymerization and interleukin (IL)-2 synthesis in the T-cell receptor (TCR) signaling pathway It has been suggested that an EVH1 domain in the N-terminal region of WASP may participate in IL-2 synthesis In trans-genic mice expressing anti-EVH1 scFvs derived from hybridoma cells pro-ducing WASP-EVH1 mAbs, a large number of scFvs in the cytosol and binding between anti-EVH1 scFvs and native WASP in T cells were detec-ted by immunoprecipitation analysis Furthermore, impairment of the pro-liferative response and IL-2 production induced by TCR stimulation which did not affect TCR capping was demonstrated in the scFv transgenic

T cells We previously described the same T-cell defects in WASP trans-genic mice overexpressing the EVH1 domain These results indicate that the EVH1 intrabodies inhibit only the EVH1 domain function that regu-lates IL-2 synthesis signaling without affecting the overall domain structure

of WASP The novel procedure presented here is a valuable tool for in vivo functional analysis of cytosolic proteins

Abbreviations

BrdU, 5-bromo-2¢-deoxyuridine ER, endoplasmic reticulum; intrabody, intracellular expressed antibody; EVH1, enabled ⁄ vasodilator-stimulated phosphoprotein (Ena ⁄ VASP) homology 1; FITC, fluorescein isothiocyanate; GST, glutathione S-transferase; IL, interleukin; scFv, single-chain variable fragment; TCR, T cell receptor; VH, heavy chain variable; VL, light chain variable; WASP, Wiskott–Aldrich syndrome protein; WIP, WASP-interacting protein.

chain variable (VL) They are able to fold and retain

the antigen-binding specificity and affinity of the

paren-tal antibody [1,2] scFvs are expressed more easily than

whole antibodies assembled with heavy and light chains

by disulfide bonds In general, the antibody fragments

for assembling the scFvs are isolated from either

anti-body phage display libraries [3] or well-characterized

hybridoma producing mAbs Although screening of

phage libraries allows the selection of antibody

frag-ments directed against a variety of antigens, the

screened antibody fragments often show low or

inter-mediate affinity for the antigen Therefore, large-scale

libraries and extensive screening are required for the

selection of the antigen-specific antibody fragments On

the other hand, the antibody fragments isolated from

hybridomas have high affinity and specificity for the

target molecules However, the cloning of heavy and

light chain variable regions by RT-PCR can be difficult

because of the presence of nonspecific variable region

transcripts produced by myeloma cells that are fused to

the antibody-producing cells

In functional proteomics, comprehensive protein

analyses have been demonstrated [4] However, the

development of a new procedure for domain analysis

of protein is necessary Gene knock-out technologies

that rely on developing a phenotype from null

muta-tion of the gene in embryonic stem cells are powerful

tools for understanding gene function Recently, RNA

interference (RNAi) which can eliminate specific

mRNA and lead to gene silencing has been developed

[5] However, these gene knock-out and silencing

tech-niques cannot be used to analyze domain structures

and functions and post-translationally modified protein

functions Dominant negative gene knock-out

proce-dures succeed in inhibiting the targeted domain

func-tions of proteins, but not in all cases

Antibodies have been used for various purposes for a

long time For example, they have been used as reagents

for Western blotting, immunostaining,

immunoprecipi-tation and blocking of protein function Therefore, if

intrabodies retain their specificity and high-affinity

bind-ing properties, they may be useful tools for inhibition of

protein function inside the cell In fact, much attention

has been paid to intrabodies for clinical applications

The functional knockdown of target proteins, such as

HIV gp120, chemokine receptor, growth factor

recep-tors, MHC class I, Ras oncogene, p53 tumor

suppres-sor, and protein kinases has been demonstrated [6–12]

If the target proteins are synthesized and processed in

the endoplasmic reticulum (ER), scFvs are expressed

with the signal peptide at the N-terminus of VHand VL

with the ER retention signal KDEL (Lys-Asp-Glu-Leu)

at the C-terminus Folded scFvs can bind to the target

proteins on the lumen side and inhibit transport of tar-get proteins in the process of functional maturation [6,8] If the targets are cytosolic proteins, scFvs without the signal peptide are used for expression in the cytosol However, expression levels of scFvs are low in the cyto-sol, and binding of scFv intrabodies to target molecules

is difficult to detect [13] A small quantity of intrabodies

in the cytosol may explain the low translational effi-ciency and low stability of intrabodies in the cytosol Wiskott–Aldrich syndrome protein (WASP), the causal gene product of the X-linked immunodeficiency (WAS) [14,15], participates in TCR signaling as a cyto-solic adaptor molecule [16–18] It is well known that TCR stimulation activates various signaling cascades accompanied by recruitment of adaptor molecules, protein kinases and regulatory molecules into the membrane-receptor complexes, resulting in the correct initiation and amplification of the signaling reaction WASP is an adaptor molecule containing multiple domains: for example, a GTPase-binding domain, which is thought to interact with Cdc42, and a pro-line-rich region, which interacts with the Src homol-ogy 3 domain of the adaptor Nck, Grb2 and several kinases [19–22] Furthermore, WASP is also associated with the actin-related protein (Arp2⁄ 3) complex through its C-terminal region The association of WASP and the Arp2⁄ 3 complex activates the actin nucleation activity of the Arp2⁄ 3 complex [23]

To investigate further the function of the WASP-EVH1 domain in the TCR signaling pathway, we developed transgenic (Tg) mice that express intrabodies that specifically bind to the WASP-EVH1 domain The cDNA fragments that encode variable regions of heavy and light chains were isolated from two established hybridomas producing WASP-EVH1-specific mAbs We constructed several scFvs consisting

of VH and VL regions with⁄ without the VH leader sequence at the N-terminal and with⁄ without the

CL(j) region behind the VL region None of the con-structs contained the KDEL sequence at the C-termi-nus We compared the quantity of scFv intrabodies and assessed their binding activity to the WASP-EVH1 domain in the scFv gene-transfected T cells Finally,

we succeeded in expressing the functional scFv intra-bodies in the cytosol and precisely knocking down the targeted protein domain in scFv transgenic mice

Results

Construction of anti-WASP-EVH1 scFvs

To assess the binding activity to native WASP in

T cells, mAb clones (17, 18 and 21) were confirmed

by immunoprecipitaion Clones 18 and 21 were able

to bind to the native form of WASP expressed in

T cells, but clone 17 was not able to

immunoprecipi-tate native WASP (Fig 1A) On the basis of this

result, clones 18 and 21 were selected for

construc-tion of scFv intrabodies For the design of primers

for PCR amplification of cDNA that encodes

sub-type-specific VH and VL regions, mAbs were checked

by an isotyping test Clones 18 and 21 were

classi-fied as IgG3⁄ j and IgG2b ⁄ j, respectively The

appropriate cDNA fragments of the VH and VL

regions were then generated by RT-PCR A

compar-ison of the VH and VL amino-acid sequences of

clones 18 and 21 is shown in Fig 1B All of the VH

regions and the complementarity-determining region

3 of the VL regions differed strongly between the

two clones

Generation of scFv from hybridomas was achieved

by well-established molecular engineering methods

The four-step PCR using appropriate primers allowed

amplification and assembly of the VH and VL regions

(Fig 2A) To investigate the stability of scFvs, we

designed several scFv constructs with and without the

N-terminal leader signal sequence of the VH region

and with and without the CL(j) region following the

VL region, which are described as HL, SHL, HL-CL

and SHL-CL in Fig 2B

Expression of scFv intrabodies and binding

to WASP

In all scFv gene-transfected T cells, expression of scFv intrabodies was detected by Western blot analysis However, scFvs containing the VH signal peptide sequence and CLregion (SHL or SHL-CL) were highly expressed in T cells (Fig 3A) These results strongly suggest that the addition of the VH signal peptide sequence and CL(j) region to scFvs increases the sta-bility of the scFv intrabodies in T cells

An in vitro binding assay was performed using gluta-thione S-transferase (GST) pull-down to detect the binding activity of anti-WASP scFvs Constructs con-taining the VH signal peptide sequence, 18SHL⁄ SHL-CL and 21SHL⁄ SHL-CL, were able to bind to GST-WASP15 (see Experimental procedures for defini-tion of WASP15), whereas no binding activity of scFvs that did not contain the signal peptide sequence

of the VH region was detected (Fig 3B) Although the expression levels of 18SHL⁄ SHL-CL and 21SHL⁄ SHL-CL were almost the same in the scFv gene-transfected T cells, 21SHL⁄ SHL-CL bound more strongly to GST-WASP15 than 18SHL⁄ SHL-CL (Fig 3A,B) Furthermore, to examine the interaction

in vivo between scFv intrabodies and the target mole-cule, WASP, scFv gene-transfected T cells were lysed

Fig 1 Selection of WASP EVH1 mAbs for

assembling scFvs and aligned amino-acid

sequences of the VHand VLregions (A)

Immunoprecipitation of T cell lysates with

WASP EVH1 mAbs produced by established

hybridomas T cell lysates were

immunopre-cipitated with 5 lgÆmL)1control mouse IgG

(lane 1), clone 17 (lane 2), clone 18 (lane 3),

clone 21 (lane 4) or commercially available

WASP mAb (lane 5) and analyzed by

West-ern blotting with WASP polyclonal antibody.

Control T cell lysates were loaded in lane 6.

The 30-kDa bands (arrowhead) indicated

secondary antibody cross-reactive

nonspe-cific proteins (B) Comparison of deduced

amino-acid sequences of the VHand VL

frag-ments derived from WASP EVH1 mAbs 18

and 21 Shared amino acids are indicated by

bars Leader signal sequences and three

complementarity-determining regions are

shown in gray boxes Four framework

regions (FR) are marked above the

sequence.

and immunoprecipitated with WASP mAb A strong

interaction between WASP and 21SHL⁄ SHL-CL scFvs

was detected by Western blot analysis with Myc tag

antibody, whereas 18SHL⁄ SHL-CL scFvs and other

scFvs were not able to associate with native WASP

(Fig 3C) The binding specificity for the WASP EVH1

domain was demonstrated by in vivo interaction

between T7-tagged WASP15 and 21SHL⁄ SHL-CL

scFvs (Fig 3D) These results suggest that 21SHL and

21SHL-CL are stably expressed as intrabodies with

domain-specific binding capabilities, and are able to associate with native WASP in T cells

To detect cleavage of the VHsignal peptide sequence from the N-terminal scFv-VH region, the N-terminal amino-acid sequence of scFv 21SHL-CL expressed in

T cells was determined Unfortunately, we could not detect the N-terminal sequence by the well-established Edman method because the N-terminal amino-acid residue was blocked Moreover, we could not detect a

V H (G 4 S) 3 V L

V H (G 4 S) 3 V L

Myc

Myc

C L

C L

Myc

Myc

SHL

HL

SHL-CL

HL-CL

11, 12

13, 14 A

B

Fig 2 Constructions of anti-WASP EVH1 scFvs (A) Cloning of

vari-able region of immunoglobulin heavy and light chains from

hybri-doma cells producing WASP EVH1 mAb The arrows represent

the following primers used to amplify the antibody fragments:

pri-mer 1, 5¢-CACCCAAGCTTGCCACCATGGGCAGACTTACTTCTTCATTC-3¢;

primer 2, 5¢-CAGAACCACCACCCCCTGAGGAGACGGTGACTGAGG

ATCC-3¢; primer 3, 5¢-CACCCAAGCTTGCCACCATGCAGGTTACTCT

GAAAGAGTC-3¢; primer 4, 5¢-CACCCAAGCTTGCCACCATGAAATG

CAGCTGGGTTATCTTC-3¢; primer 5, 5¢-CAGAACCACCACCCCCTG

AGGAGACGGTGACTGAGGTTCC-3¢; primer 6, 5¢-CACCCAAGCTT

GCCACCATGGAGGTTCAGCTGCAGCAGTCTG-3¢; primer 7, 5¢-GGT

GGAGGAGGTTCTGATGTTTTGATGACCCAAACTCCAC-3¢; primer 8,

5¢-CGAATGCGGCCGCCCGTTTGATTTCCAGCTTGGTGC-3¢; primer 9,

5¢-GGTGGAGGAGGTTCTGATGTTGTTCTGACCCAAACTCCACTC-3¢;

primer 10, 5¢-CGAATGCGGCCGCCCGTTTCAGCTCCAGCTTGGTCC-3¢;

primer 11, 5¢-TCAAAACATCAGAACCTCCTCCACCGGATCCTCCAC

CTCCAGAACCACCACCCCC-3¢; primer 12, 5¢-GAACAACATCAGAA

CCTCCTCACCGGATCCTCCACCTCCAGAACCACCACCCCC-3¢;

pri-mer 13, 5¢-CGTCTCCTCAGGGGGTGGTGGTTCTGGAGGTGGAG

GATCCGGTGGAGGAGGTTCT-3¢; primer 14, 5¢-CGTCTCCTCA

GGGGGTGGTGGTTCTGGAGGTGGAGGATCCGGTGGAGGAGG

TTCT-3¢ In all primers, underlined sequences indicate restriction site

of HindIII and NotI, and bold letters indicate full or part of the (Gly 4

-Ser)3linker sequence (B) Schematic representation of the four scFv

formats (SHL, HL, SHL-CL, and HL-CL) Shown are the leader signal

sequence, V H region, polypeptide linker (G 4 S) 3 , V L region, light chain

constant [CL(j)] region and Myc tag sequence.

G W G W G W G W G W G W G W G W G W

49.9

32.3

49.9

49.9 32.3

32.3

(kD)

(kD)

(kD)

WASP

A

B

C

G: GST W: GST-WASP15

87

49.9

T7-WASP15

D

21SHL-CL-Myc 21SHL-Myc 49.9

32.3 (kD)

28.8

22

vector 18HL 18SHL 18HL

-CL 18SHL -CL 21HL 21SHL 21HL -CL 21SHL -CL

vector 18HL 18SHL 18HL

-CL 18SHL -CL 21HL 21SHL 21HL

-CL 21SHL -CL

vector 18HL 18SHL 18HL

-CL 18SHL -CL 21HL 21SHL 21HL -CL 21SHL -CL

T7 -WASP15 T7 -WASP15 + 21HL

T7 -WASP15 + 21SHL

T7 -WASP15 + 21HL -CL

T7 -WASP15 + 21SHL -CL

Fig 3 Expression of anti-WASP scFvs and detection of their bind-ing activity to WASP in T cells (A) Western blot analysis of protein extracts of anti-WASP scFv DNA-transfected T cells The immuno-blot was probed with Myc tag mAb (B) In vitro binding assay using GST pull-down All anti-WASP scFv DNA-transfected T cells were lysed and incubated with GST (G) or GST-WASP15 (W) fusion pro-tein noncovalently bound to glutathione–Sepharose beads Bound proteins were analyzed by Western blotting with Myc tag mAb (C)

In vivo association between scFvs and WASP All scFv DNA-trans-fected cell lysates were immunoprecipitated with WASP mAb and analyzed by Western blotting with Myc tag mAb (top panel) or WASP mAb (bottom panel) (D) EVH1 domain-specific binding of scFv T7-WASP15 and scFv DNA cotransfected cell lysates were immunoprecipitated with biotinylated T7 tag mAb Immunocom-plexes were recovered by on streptavidin–agarose and analyzed by Western blotting with Myc tag mAb (top panel) or T7 tag mAb (bot-tom panel) Arrowheads indicate secondary antibody cross-reactive nonspecific proteins.

VHsignal peptide sequence by MS analysis of

21SHL-CL digested with lysyl endopeptidase (data not

shown) These results suggest that the VH signal

pep-tide sequence was cleaved from the N-terminal VH

region The culture supernatant of 21SHL-CL

scFv-expressed T cells was examined but the scFv could not

be detected (data not shown) These results suggest

that, even if scFvs are expressed with a signal

sequence, they do not enter the secretory pathway

Generation of anti-WASP scFv transgenic mice

scFv 21SHL and 21SHL-CL vector DNAs were chosen

as the transgenes for development of transgenic mice

with the functional knockdown WASP-EVH1 domain

High expression of 21SHL and 21SHL-CL was

detected in T and B cells from the spleens of the

21SHL⁄ 21SHL-CL scFv transgenic mice (Fig 4A)

Eight 21SHL transgenic founders and 10 21SHL-CL

transgenic founders carrying the scFv intrabody

expression vectors were obtained In four of eight

21SHL lines and five of 10 21SHL-CL lines, the same

levels of expression of 21SHL and 21SHL-CL were

detected (data not shown) Furthermore, T and B cells from the spleens of both scFv transgenic mice were solubilized with 1% digitonin buffer and immuno-precipitated with WASP mAb and Myc tag mAb to examine the in vivo interaction between scFvs and endogenous WASP Binding of intracellular scFvs and WASP was detected in both T and B cells from scFv transgenic spleens by immunoprecipitation (Fig 4B–D)

Impaired antigen receptor-induced proliferation

in anti-WASP scFv transgenic T cells, but not

B cells

To assess the effects of the anti-WASP scFvs 21SHL and 21SHL-CL on T-cell function, the proliferative response to stimulation with CD3e antibody (2c11) was examined Compared with the wild-type, T cells from 21SHL transgenic mice and 21SHL-CL trans-genic mice were impaired in their proliferative response

to CD3e antibody stimulation to the same extent as in WASP15 transgenic T cells [24] (Fig 5A) These find-ings indicate that the function of the WASP N-ter-minal EVH1 domain is blocked by scFv 21SHL and 21SHL-CL intrabodies in the T cells In contrast with

T cells, proliferative responses to antigen receptor sti-mulation with anti-IgM Ab F(ab¢)2 or CD40 antibody were normal in the scFv transgenic B cells (Fig 5B) Therefore, the EVH1 domain of WASP is not func-tional, at least in the Ag receptor-induced proliferative response of B cells

T cells from the other three 21SHL transgenic lines and the other four 21SHL-CL transgenic lines were also impaired in their proliferative response to stimula-tion with CD3e antibody (data not shown), confirming that there were no problems in the integration site of the transgene

Lymphoid development in anti-WASP scFv transgenic mice

T-cell development in the spleen can be followed by examining the expression patterns of the CD4 and CD8 surface antigens The population of mature single-positive thymocytes (either CD4+CD8– or CD4–CD8+) was almost the same in wild-type, 21SHL transgenic, and 21SHL-CL transgenic mice (Fig 5C) Likewise the expression pattern of CD3 was nearly the same Furthermore, the percentages of splenic T and B lineage cell populations were normal (Fig 5C) In addition, T lineage cell populations in the thymus and

B lineage cell populations in the bone marrow were almost the same for wild-type, 21SHL transgenic and 21SHL-CL transgenic mice (Fig 5D,E) These results

T cell B cell

21SHL-CL-Myc

21SHL-Myc

21SHL-CL-Myc

21SHL-Myc

WASP

WASP

49.9

32.3

49.9

32.3

(kD) 21SHL 21SHL

-CL 21SHL 21SHL

-CL

WB: anti-Myc tag

IP: anti-WASP

WB: anti-Myc tag

IP: anti-Myc tag

WB: anti-WASP

A

B

C

D

Fig 4 Expression of anti-WASP scFvs and in vivo interaction

between scFvs and WASP in scFv transgenic mice T and B cells.

(A) Western blot analysis of protein extracts of T and B cells from

the spleens of the 21SHL and 21SHL-CL scFv transgenic mice The

immunoblot was probed with Myc tag mAb (B, C) In vivo

associ-ation between scFvs and WASP The scFv 21SHL and 21SHL-CL

transgenic T and B cell lysates were immunoprecipitated with

WASP mAb and Myc tag mAb and analyzed by Western blotting

with Myc tag mAb and WASP mAb Arrowheads indicated

secon-dary antibody cross-reactive nonspecific proteins (D) Both scFv

transgenic mice T and B cell lysates were analyzed by Western

blotting with WASP antibody.

indicate that anti-WASP scFvs do not have a marked

effect on lymphocyte development

Impaired interleukin (IL)-2 production induced

by TCR stimulation, but not antigen receptor

capping

To assess whether the 21SHL and 21SHL-CL scFvs

affect IL-2 production induced by TCR stimulation,

purified T cells from spleens of wild-type, WASP15 transgenic, 21SHL transgenic and 21SHL-CL trans-genic mice were stimulated with immobilized CD3e antibody and IL-2 in the culture supernatant and deter-mined by ELISA T cells expressing 21SHL and 21SHL-CL scFvs were impaired in IL-2 production induced by TCR stimulation, whereas the defect in IL-2 production of scFv transgenic T cells was slight com-pared with the WASP15 transgenic T cells (Fig 6A)

In addtion, purified T cells were incubated in vitro with fluorescein isothiocyanate (FITC)-conjugated CD3e antibody at either 37
C or 4
C (stimulation or nonstimulation) to assess whether the 21SHL and 21SHL-CL scFvs affect TCR-induced capping The rate of antigen-receptor capping of T cells was the same in all the mice (Fig 6B) These results indicate that the anti-WASP scFvs 21SHL and 21SHL-CL inhi-bit the signaling cascade of IL-2 production via TCR stimulation without affecting the regulation of the cytoskeleton, including antigen-receptor capping These findings strongly indicate that IL-2 synthesis is medi-ated directly by the WASP EVH1 domain and not by secondary events resulting from WASP-mediated actin cytoskeletal rearrangements induced by TCR signaling

Subcellular localization of anti-WASP scFvs

To examine the subcellular localization of anti-WASP scFvs 21SHL and 21SHL-CL in T cells, cell extracts

of their scFv-transgenic T cells were fractionated into the subcellular compartments, cytosolic proteins and

CD4

CD3

CD4

IgM

10.7

22.1

36.3

31.5

3.6

5.5

7.1

0.4

11.7

24.5

37.9

33.4

4.1

7.0

9.7

0.1

13.4

25.8

35.4

34.5

3.7

8.0

9.1

0.1

wild 21SHL Tg 21SHL-CL Tg

A

B

C

D

E

Fig 5 Antigen receptor-induced proliferation in anti-WASP scFv transgenic T and B cells, and lymphoid development in anti-WASP scFv transgenic mice (A) T-cell proliferation Splenic T cells from anti-WASP scFv 21SHL transgenic, 21SHL-CL transgenic, WASP15 transgenic and wild-type mice were cultured in medium alone or in the presence of CD3e antibody (B) B-cell proliferation Splenic

B cells from anti-WASP scFv 21SHL transgenic, 21SHL-CL trans-genic, WASP15 transgenic and wild-type mice were cultured in medium alone or in the presence of IgM antibody F(ab¢) 2 or CD40 antibody Each stimulation was performed in the presence of exo-genous IL-4 In each experiment, cells were cultured for 48 h, then

10 l M BrdU was added to the T and B-cell cultures The cells were reincubated for an additional 16 h, and BrdU incorporation was quantified by ELISA Values represent means ± SE of triplicate cultures and are representative of three independent experi-ments Statistical significance is indicated by *(P < 0.05) and

**(P < 0.005) (C)–(E) FACS analyses of lymphocytes from wild-type, anti-WASP scFv 21SHL transgenic and 21SHL-CL transgenic mice Two-color flow cytometric analyses were performed on spleen (C), thymus (D) and bone marrow (E) Percentages of repre-sentative lymphoid populations are noted The results shown are representative of at least three male mice for each analysis at the age of 8 weeks.

membrane⁄ membrane organelles In general, the scFv

intrabodies (VH–linker–VL format) with heavy chain

signal peptide sequences cross the rough ER

mem-brane and enter the secretory pathway through the

trans-Golgi network However, equivalent amounts of

scFv 21SHL were detected in both the cytosol and

membrane fractions, and most of the scFv 21SHL-CL

was detected in the cytosol fraction in anti-WASP scFv

transgenic T cells (Fig 7A) To confirm the presence

of cross-contamination in both fractions of scFv

21SHL-CL transgenic T cells, each fraction was

exam-ined by Western blotting with WASP antibody and

Ribophorin I antibody specifically expressed in the

cytosolic and membrane fractions, respectively These

results show that neither fraction was

cross-contamin-ated (Fig 7A)

When scFv intrabodies were expressed in NIH-3T3

fibroblastic cells, the scFvs were localized in the

subcel-lular compartments NIH-3T3 cells were transfected

with scFv 21SHL-CL (with leader signal sequence) or

21HL-CL (without leader signal sequence) genes and

then their subcellular fractions were subjected to

West-ern blotting with Myc tag antibody The majority of

the intrabodies expressed without signal sequence were

detected in the cytosol, whereas most of the intrabodies

expressed with the signal sequence were detected in the membrane fraction (Fig 7B) These results indicate that the post-translational processing of ER-coupled protein synthesis must be different among cell types such as lymphocytes and fibroblasts

On immunostaining, colocalization of 21SHL-CL scFv and endogenous WASP was observed in the cyto-sol of the scFv DNA transfected T cells (Fig 7C) Again these results indicate that scFv intrabodies expressed with the VH signal peptide sequence are localized in the cytosol of T cells Taken together, the results strongly suggest that scFv intrabodies synthes-ized in the ER are released from the ER membrane into the cytosol by retro-translocation in lymphocytes including T cells [25]

In general, when proteins synthesized in the ER are misfolded or incompletely assembled into oligomeric forms, they are transferred from the lumen of the ER into the cytosol, so-called retro-translocation In the cytosol, the retro-translocated proteins are polyubiqui-tinated and degraded by proteasomal proteolysis [26–29] Our results suggest that the WASP scFv intra-bodies expressed with the VH signal sequence are translocated across the ER membrane into the cytosol without degradation The cell lysates or

unstimulate

stimulate

B

A

Fig 6 IL-2 production was impaired, but

not antigen receptor capping induced by

TCR stimulation (A) Splenic T cells from

anti-WASP scFv 21SHL transgenic,

21SHL-CL transgenic, WASP15 transgenic and

wild-type mice were cultured in medium

alone or in the presence of anti-CD3e Ab.

Each cell culture supernatant was collected

at 24 h IL-2 in the supernatant was

quanti-fied by ELISA Values are mean ± SE from

triplicate cultures and are representative of

three independent experiments Statistical

significance is indicated by *(P < 0.005) and

**(P < 0.001) (B) Splenic T cells from

anti-WASP scFv 21SHL transgenic, 21SHL-CL

transgenic, WASP15 transgenic and

wild-type mice were incubated with

FITC-conju-gated CD3e antibody at either 4
C or 37
C

for 30 min The treated cells were placed on

polyethylenimine coated eight-well tissue

culture glass slides, fixed, analyzed and

photographed at · 100 using confocal

micro-scopy The rate of capping of unstimulated

and stimulated T cells was determined by

counting the number of caps in  200 cells ⁄

experiment The wild-type and transgenic

mice used for these experiments were

8 weeks old.

cipitates with Myc tag antibody were immunoblotted with ubiquitin antibody to determine if polyubiquitina-tion of the anti-WASP scFv 21SHL and 21SHL-CL occurs in the T cells 21SHL and 21SHL-CL were not polyubiquitinated in the scFv-transgenic T cells How-ever, the polyubiquitination of nonspecific proteins was observed in the scFv-transgenic T cell lysate (Fig 7D) These results indicate that the scFv genes with signal sequence are translated in the ER, and, after cleavage of the signal peptide sequence, are trans-located from the ER into the cytosol without poly-ubiquitination and degradation

Discussion

In this study, we show that the scFv intrabodies con-structed with a leader signal sequence at the N-termi-nus inhibited the domain function of a cytosolic protein, and preserved the strong binding activity for target molecules under the reducing conditions of the cytosol in scFv-transgenic lymphocytes

This study also demonstrates that the successful expression of intrabodies in the cytosol is related to translational efficiency, post-translational processing, and modification of scFvs We constructed several scFvs with or without the N-terminal leader signal sequence of the VH region and with or without the

CL(j) region following the VLregion (Fig 2B) Fusion

of scFvs with the CL(j) region has already been shown

to increase intracellular stability and target protein inactivation in some cases, but not all [30,31] scFvs containing the VH signal sequence and CL(j) region (SHL or SHL-CL formats) were highly expressed in

T cells compared with scFvs not containing the VH signal sequence (HL or HL-CL formats) (Fig 3A) Binding activity of 18SHL⁄ SHL-CL and 21SHL ⁄ SHL-CL scFvs was detected, but not in the HL⁄ HL-CL formats, by in vitro binding assay using GST pull-down (Fig 3B) These results strongly suggest that scFvs with the native VH signal sequence and CL(j) region increase the binding capabilities of scFv intra-bodies in T cells

In this study, we established two hybridoma cell lines (clones 18 and 21) producing WASP EVH1 mAbs which were able to equivalently immunoprecipitate with native WASP in T cells Then we isolated cDNA fragments for assembling anti-WASP scFvs from them Although the expression levels of 18SHL⁄ SHL-CL and 21SHL⁄ SHL-CL were almost the same in the scFv gene-transfected T cells, 21SHL⁄ SHL-CL bound more strongly to GST-WASP15 than 18SHL⁄ SHL-CL (Fig 3A,B) Furthermore, a strong interaction between native WASP and 21SHL⁄ SHL-CL scFvs was detected

1 2 3 4 1 2 3 4

anti-Myc tag anti-Ub

21SHL

21SHL-CL

WASP

Ribophorin I

C: Cytosolic M: Membrane/

Organelle

21HL-CL 21SHL-CL

anti-WASP scFv-Tg T cell

anti-WASP scFv transfected NIH-3T3

anti-Myc tag anti-WASP merged image

(kD)

199

133

87

40.1

31.6

40.1

32.3

49.9

40.1

87

50.7

87

50.7

49.9

40.1

(kD)

A

B

C

D

Fig 7 Subcellular localization of anti-WASP scFvs Cell extracts of

(A) anti-WASP scFvs transgenic T cells and (B) anti-WASP scFv

DNA-transfected NIH-3T3 cells were fractionated into the

subcellu-lar compartments, cytosolic proteins and membrane ⁄ membrane

organelles The fractionated cell extracts were analyzed by Western

blotting with Myc tag, WASP or Ribophorin I antibodies (C)

Co-localization of anti-WASP scFv and endogenous WASP in the

cytosol of T cells Anti-WASP scFv 21SHL-CL DNA electroporated

T cells were fixed and incubated with Myc tag antibody or WASP

mAb After being washed, the cells were stained with

FITC-conju-gated anti-rabbit IgG or Alexa Fluor 546-conjuFITC-conju-gated anti-mouse IgG.

The treated cells were analyzed and photographed at · 100 using

immunofluorescence microscopy (D) Anti-WASP scFvs were not

polyubiquitinated in the scFv transgenic mice T cells

Immunopre-cipitates with Myc tag antibody (lanes 1 and 2) and cell lysates

(lanes 3 and 4) from anti-WASP scFv 21SHL transgenic (lanes 1

and 3) or 21SHL-CL transgenic (lanes 2 and 4) mice T cells were

analyzed by Western blotting with Myc tag or ubiquitin antibody.

The smear bands (arrow) indicate polyubiquitination of nonspecific

proteins in the T cells The arrowhead indicates secondary antibody

cross-reactive nonspecific proteins.

by immunoprecipitaion analysis, whereas 18SHL⁄

SHL-CL scFvs and other scFvs were not able to

asso-ciate with the native WASP in vivo (Fig 3C) Also, the

EVH1 domain-specific binding of 21SHL⁄ SHL-CL

was demonstrated (Fig 3D) These results indicate that

the differences in in vivo binding activity between

18SHL⁄ SHL-CL and 21SHL ⁄ SHL-CL may be due to

folding This structural property is necessary for

anti-gen binding, when antibody fragments are converted

into the scFv format and expressed in the reducing

environment of the cytosol

The primary mechanism for protein degradation of

misfolded proteins is the ubiquitin–proteasome system

which governs the quality control of proteins

Mis-folded proteins synthesized by cotranslational or

post-translational events in the ER are retro-translocated

from the ER into the cytosol and rapidly degraded

after polyubiquitination [26–29] Immunoglobulin

heavy and light chains are cotranslated and assembled

with disulfide bonds in the ER lumen An ER

resi-dent chaperone, Bip, binds to the constant region of

immunoglobulin and stabilizes and maintains the

integrity of the immunoglobulin form [32,33] Bip

contains the ER retention signal sequence KDEL

(Lys-Asp-Glu-Leu) at the C-terminus and elaborates

the tertiary structure of immunoglobulin during

trans-location from the ER to the secretory pathway In

general, it was thought that the scFv intrabodies

con-structed with leader signal sequences cross the rough

ER membrane and enter the secretory pathway

through the trans-Golgi network However, in this

study the majority of scFvs were detected in the

cyto-solic fraction (Fig 7A), and not in the T-cell culture

supernatant (data not shown) These results indicate

that the scFvs constructed with the signal sequence

can be transferred from the ER to the cytosol by

retro-translocation without polyubiquitination and

proteasome degradation We propose two possible

interpretations of these results One is that

ubiquitina-tion is not coupled with retro-translocaubiquitina-tion as has

been shown for cholera toxin release from the ER

into the cytosol accompanied by rapid folding [34]

The other is that the scFv modifications of the

N-ter-minal residues occur after cleavage of the signal

sequence in the ER In MyoD, which is a

tissue-spe-cific transcriptional activator that acts as a master

switch for muscle development, modification of the

N-terminal residue protects it from ubiquitination and

protein degradation irrespective of the presence of

internal lysine residues [35] In T cells, scFvs

con-structed with the VH signal sequence seem to be

modified at the N-terminal residue after cleavage of

the signal peptide sequence in the ER However, we

have not yet confirmed the N-terminal amino-acid sequence of scFvs by the Edman method

Interestingly, when NIH-3T3 cells were transfected with the 21SHL-CL scFv (containing the VH signal sequence) or the 21HL-CL scFv (not containing the

VHsignal sequence), most of the 21SHL-CL scFv was detected in the membrane fraction, whereas the

21HL-CL scFv accumulated in the cytosol (Fig 7B) These results indicate that the mechanisms of retro-transloca-tion differ among different cell types Although we do not know the mechanisms leading to retro-transloca-tion without proteasome degradaretro-transloca-tion, the scFv intra-bodies constructed with signal sequences may be designed for practical use in functional knockdown of cytosolic proteins in T cells

T cells from WASP-deficient mice showed a marked reduction in antigen receptor capping accompanied by actin polymerization and IL-2 production induced by TCR stimulation It has been hypothesized that defects

in IL-2 production in WASP-deficient T cells may be a secondary phenomenon resulting from defects in actin remodeling and immune synapse formation induced by TCR stimulation [17,18,36,37] However, we previously demonstrated that T cells from WASP15 transgenic mice that overexpress WASP-EVH1 domain were impaired with respect to proliferation and IL-2 pro-duction induced by TCR stimulation, but antigen receptor capping and actin polymerization were nor-mal [24] This suggest the direct involvement of the EVH1 domain in the IL-2 synthesis pathway In the present study, purified anti-WASP scFv 21SHL and 21SHL-CL transgenic T cells were impaired with respect to proliferation and IL-2 production induced

by CD3e antibody stimulation (Figs 5A and 6A) In terms of cytoskeletal rearrangement, normal antigen receptor capping induced by CD3e antibody stimula-tion was observed similar to the wild-type and WASP15 transgenic mice T cells (Fig 6B) These results indicate that the role of WASP in regulating IL-2 production is independent of its role in immune synapse formation The following experimental data support this hypothesis WASP-deficient T cells form conjugates with antigen-specific B cells normally and can form immune synapses accompanied by polariza-tion of cytoskeleton-regulating proteins, but defects in IL-2 production are observed [38] Furthermore, analy-sis of a series of WASP-deletion mutants shows that the WASP homology-1 (WH1)⁄ EVH1 domain is res-ponsible for NF-AT transcriptional activation [39] These findings indicate that the functions of WASP may be more complex than previously believed The inability of WASP-deficient, WASP15 trans-genic, and anti-WASP scFv 21SHL⁄ SHL-CL transgenic

T cells to proliferate in response to TCR stimulation is

similar to the defects observed in T cells from

Vav-deficient mice [40,41] It has been previously shown

that Vav is a potent regulator of the IL-2 promoter, in

particular NF-AT⁄ AP-1-mediated gene transcription

[42] Furthermore, the WASP-interacting protein

(WIP) and WASP interaction is important for

Vav-mediated activation of NF-AT⁄ AP-1 gene

transcrip-tion induced by TCR stimulatranscrip-tion [43] WIP-deficient

T cells were impaired in proliferation and immune

syn-apse formation induced by TCR stimulation [44] It is

possible that the overexpressed WASP15 and

anti-WASP 21SHL⁄ SHL-CL scFvs inhibit WIP and

endog-enous WASP interactions, because the WIP-binding

site in endogenous WASP is included in WASP15 and

may overlap the target region of our anti-WASP

scFvs The molecules that interact with the EVH1

domain which overlaps our scFv intrabody-binding

epitope need to be identified

We examined the effects of anti-WASP 21 SHL⁄

SHL-CL scFvs on lymphocyte development and B-cell

function The anti-WASP scFvs did not have a marked

effect on lymphocyte development (Fig 5C–E)

Fur-thermore, B cells from anti-WASP 21 SHL⁄ SHL-CL

scFv transgenic mice proliferated normally in response

to stimulation by IgM and CD40 antibodies (Fig 5B)

B cells from WASP-deficient and WASP15 transgenic

mice also proliferated normally after IgM antibody

stimulation [17,18,24] We have not yet clarified the

significance of WASP expression in B lymphocytes

Finally, we demonstrate here that scFv intrabodies

bind to the EVH1 domain of WASP and inhibit IL-2

synthesis in T cells Therefore, scFv intrabodies should

be valuable tools for identifying novel protein

func-tions, and transgenic mice that express scFv

intrabod-ies may be useful in functional knockdown models

Experimental procedures

Construction of GST fusion protein and mAb

preparation

A cDNA fragment for mouse WASP exon 1–5 (amino acids

1–171) which includes the EVH1 domain (designated

WASP15) was generated by PCR (sense primer, 5¢-CGA

ATGCGGCCGCAATGAATAGTGGCCCTG-3¢; reverse

primer, 5¢-CGAATGCGGCCGCTCACTCCTCATTGATT

GG-3¢) [24], digested with NotI, and subcloned into the

pGEX-4T-2 expression vector (Amersham Biosciences,

Pis-cataway, NJ, USA) The GST-WASP15 fusion protein was

produced in BL21 Escherichia coli cells and purified on a

glutathione–Sepharose 4B affinity chromatography column

(Amersham Biosciences) according to the manufacturer’s

instructions mAbs were prepared from mice immunized with GST-WASP15 fusion protein by the conventional procedure

Cloning and construction of WASP-EVH1 scFv intrabodies

We identified subtype mAbs (18, 21) using a mouse mAb isotyping kit IsoStrip (Roche Diagnostics, Mannheim, Ger-many) We performed a four-step PCR to generate appro-priate cDNA fragments that encoded the VH and VL

region Total RNA from hybridoma cells was reverse-tran-scribed using the SMARTTM RACE cDNA Amplification Kit (Clontech, Palo Alto, CA, USA) The cDNA fragments for the VH and VL regions containing the leader signal sequence and CH1 or CL constant region sequences were generated by PCR using subtype-specific primers (heavy chain, clone 18, IgG3: sense primer 5¢-CTAATACGACTC ACTATAGGGCAAGCAGTGGTATCAACGCAGAGT-3¢ and reverse primer 5¢-GTACTGGGCTTGGGTATTCT AGGCTC-3¢; clone 21, IgG2b: sense primer 5¢-AAGCAG TGGTATCAACGCAGAGTACGCG-3¢ and reverse pri-mer 5¢-GGACAGGGGTTGATTGTTGAAATGGG-3¢; light chain, clone 18 and 21, j: sense primer 5¢-CTAATAC GACTCACTATAGGGCAAGCAGTGGTATCAACGCA GAGT-3¢ and reverse primer 5¢-CCTGTTGAAGCTCTTG ACAATGGGTG-3¢) The second PCR products for VH

region were classified into two forms containing the native

VHleader signal sequence at the 5¢ end of the VHfragment (SVH form) or no VH leader signal sequence (VH form) The second PCR amplification was performed with the fol-lowing primers: 18SVH, sense primer 1 and reverse primer 2; 18VH, sense primer 3 and reverse primer 2; 21SVH, sense primer 4 and reverse primer 5; 21VH, sense primer 6 and reverse primer 5; 18VL, sense primer 7 and reverse primer 8; 21VL, sense primer 9 and reverse primer 10 Primer sequences are shown in the legend to Fig 2 The third PCR products were amplified using the following primers: 18SVH–linker, sense primer 1 and reverse primer 11; 18VH– linker, sense primer 3 and reverse primer 11; 21SVH–linker, sense primer 4 and reverse primer 12; 21VH–linker, sense primer 6 and reverse primer 12; linker)18VL, sense primer

13 and reverse primer 8; linker)21VL, sense primer 14 and reverse primer 10 The third PCR products were mixed in the following combinations: 18SVH–linker and linker) 18VL, 18VH–linker and linker)18VL, 21SVH–linker and linker)21VL, 21VH–linker and linker)21VL and single-chain antibodies, scFvs, designated 18SHL, 18HL, 21SHL, and 21HL assembled by the fourth PCR amplification using the following primers: 18SHL, sense primer 1 and reverse primer 8; 18HL, sense primer 3 and reverse primer 8; 21SHL, sense primer 4 and reverse primer 10; 21HL, sense primer 6 and reverse primer 10 The fourth PCR products were digested with HindIII–NotI and cloned into the pCAGGS-MCS expression vector [45,46] The Myc tag