[CANCER RESEARCH 58. 4766-4770. November I. 1998]

Advances in Brief

Hiimartin, the Product of the Tuberous Sclerosis l (TSC1} Gene, Interacts withTuberin and Appears to Be Localized to Cytoplasmic Vesicles1

Tracey L. Plank, Raymond S. Yeung, and Elizabeth Petri Henske2

Dcfiat'tinenl of Medical Oncology, Fo.\ Chase Cancer Center. Philadelphia Pennsylvania Will ¡T.L P., E. P. H.j. and Department cf Surgery, University of Washington. Seattle.

Wuxhinxtiin IK. S. Y.¡

Abstract

Tuberous sclerosis is an inherited syndrome associated with mutationsin two tumor suppressor genes: TSCI and TSC2. Tuberin, the product ofTSC2, appears to be localized to the Golgi apparatus and may have afunction in vesicular transport. The function of hamartin, the product ofTSCI, is not known. In this report, we demonstrate an interaction betweenhamartin and tuberin, which is detectable at endogenous protein levels.Hamartin is present in a cell line derived from the Eker rat that lacksfunctional tuberin, indicating that the stability of hamartin is not dependent on its interaction with tuberin. Hamartin is localized to the membrane/particulale (PI(H)) fraction of cultured cells. The I"I(10 localization is

unchanged in the Eker cells. Finally, we show that at endogenous expression levels, hamartin has a punctate pattern of immunofluorescence in thecytoplasm. Taken together, the presence of hamartin in the membrane/participate fraction and its pattern of cytoplasmic staining suggest that itis localized to cytoplasmic vesicles. If altered vesicular trafficking leads totumorigenesis in tuberous sclerosis, TSCI and TSC2 may have a novelmechanism of tumor suppression.

Introduction

TSC3 is an autosomal dominant tumor suppressor gene syndrome

characterized by seizures, mental retardation, autism, and tumors ofmultiple organ systems (1, 2). The tumors, which are primarily benign, arise in the brain, heart, kidney, skin, and retina. Two genes areassociated with TSC: (a) TSCÃŒon chromosome 9q34 (3); and (b)TSC2 on chromosome 16pl3 (4). Germ-line TSCÃŒand TSC2 muta

tions appear to be inactivating, and loss of heterozygosity at either theTSCI or the TSC2 region occurs in TSC tumors (5-8), indicating that

TSCI and TSC2 are tumor suppressor genes. The protein product ofthe TSC2 gene, tuberin, is expressed in a variety of different cell types(9, 10). Sequence analysis has revealed a region near the COOHterminus of tuberin with homology to the catalytic domain of the GAPRaplGAP. Tuberin has GAP activity for Rapi (11) and Rab5 (12).Tuberin has been localized to the perinuclear region of cultured cellsand has been shown at endogenous expression levels to colocalizewith Rapi in the stacks of the Golgi apparatus (13).

The TSCI gene was identified in 1997 (3). Its protein product,hamartin, is primarily hydrophilic. with a putative coiled-coil domain

near the COOH terminus. Hamartin has no homology to tuberin orother known vertebrate proteins. It does have a possible homologue inSchizoxaccltanmiyces pombe with no identified function. Recently, acoiled-coil domain near the COOH terminus of hamartin was shown

Received 8/14/98; accepted 9/17/98.The costs of publication of this article were defrayed ¡npart by the payment of page

charges. This article must therefore he hereby marked advertisement in accordance with18 U.S.C. Section 1734 solely to indicale this fact.

1Supported in part by grants from the National Tuberous Sclerosis Association. The

Lymphangiomyomatosis Foundation, and the W. W. Smith Charitable Trust.2 To whom requests for reprints should be addressed, at Department of Medical

Oncology, Fox Chase Cancer Center. 7701 Burholme Avenue. Philadelphia. PA 19111.Phone: (2151 728-2428: Fax: (215) 728 2741: E-mail: EP_Henske<s>fccc.edu.

1The abbreviations used are: TSC. tuberous sclerosis complex; GAP. GTPase-acti-

valing protein: GST. glutathione 5-transferasc.

to bind specifically to a coiled-coil domain near the NH2 terminus of

tuberin (14). suggesting that tuberin and hamartin are partners in acommon cellular pathway.

To gain insight into the function of hamartin, we have raised arabbit polyclonal antibody against a COOH-terminal peptide from

hamartin that specifically recognizes a Mr 130,000 protein in bothhuman and rat cells. In this report, we independently demonstrate theinteraction of hamartin with tuberin in vivo at endogenous proteinlevels. We also show that hamartin is localized to the membrane/paniculate (100.000 X g) fraction in cells containing tuberin and alsoin Eker rat tumor-derived cells lacking tuberin. Hamartin has a punc

tate pattern of cytoplasmic immunofluorescence, which is consistentwith vesicular localization. These findings suggest that hamartin islocalized to cytoplasmic vesicles. This pattern of localization, togetherwith data suggesting that tuberin plays a role in vesicular transport,may indicate that tumorigenesis in TSC is the result of aberrantvesicular trafficking.

Materials and Methods

Antibodies. A polyclonal antibody to hamartin referred to as Ham-C was

generated against a peptide from the COOH terminus of the TSCI polypeptide(residues 1146-1164). The peptide was conjugated to keyhole limpet hemo-

cyanin and used as an immunogen. Two rabbits were immuni/.ed with 0.1 mgof the peptide in an emulsion with Freund's adjuvant. Boosts with O.I mg of

peptide were given during weeks 2. 6, and 8. Antiserum was prepared afterammonium sulfate precipitation (50% saturation) by immunoat'finity purifica

tion using an AminoLink column with the COOH-terminal peptide as the

immobili/.ed ligand (Pierce Chemical Co.). Fractions from the column containing antibodies to hamartin were identified by dot-blot analysis with thepeptide immobili/.ed on nitrocellulose. The Ham-C antibody was used at a

1:100 dilution for Western blotting. The C20 tuberin antibody (Santa CruzBiotechnology. Inc.) was generated against a peptide mapping at the COOHterminus of tuberin. The C20 antibody was used at a 1:1500 dilution forWestern blotting. For immunoprecipitation. 3 ;nl of the C20 antibody wereadded to 500 ¡JL\of cell lysate. The L3 tuberin antibody was generated againsta GST fusion protein containing the COOH-terminal of tuberin ( 12). A 1:2500

dilution of the L3 antibody was used for Western blotting. For immunoprecipitation, 1 ¡i\of the L3 antibody was added to 500 ¿¿1of cell lysate. Theanti-/3-tubulin antibody (Boehringer Mannheim) was used at a 1:400 dilution

for immunoblotting.Cell Lines and Culture Conditions. The human embryonic kidney cell

line 293. the rat tlbroblast cell line Rat I-R12. and the rat normal kidneyepithelial cell line NRK-52E were obtained from the American Type CultureCollection. Ratl-R12 cells, which are transf'ected with a plasmid containing a

tetracycline transactivator gene and the neomycin resistance gene, were cultured in DMEM containing 10% fetal bovine serum and 0.4 mg/ml G418.NRK-52E cells were maintained in DMEM containing 0.1 mM nonessential

amino acids and 5% calf serum. The 293 cell line was cultured in DMEM with10% horse serum. The Eker tumor cell line ERC18M (12). which was the giftof Alfred Knudson (Fox Chase Cancer Center. Philadelphia, PA), was derivedfrom an Eker rat renal cell carcinoma with the subsequent passage of the cellline through a SCID mouse. Passage of ERC18M cells was performed asdescribed previously (12).

4766

Research. on February 27, 2021. © 1998 American Association for Cancercancerres.aacrjournals.org Downloaded from

SI HCELLULAR LOCALIZATION OF HAMARTIN

cDNA Clones and GST-Fusion Protein. A 1304-bp fragment of TSC1was amplified from Marathon-Ready Human Brain cDNA (Clontech) using

the Advantage cDNA PCR kit (Clontech). The fragment encompassing aminoacid residues 730-1164 was amplified using the primer pair TSCÌ-Ì(5'-CGGGATCCGAGGAACATAATGCTGCCATG-3') and TSC1-4 (.V-ATAA-GAATGCGGCCGCGCTGTGTTCATGATGAGTCT-3'). The cDNA was se-

quenced using Taq cycle sequencing to confirm its identity as TSC I. Theresulting PCR product was digested with BiimHl and Null and cloned in-frameinto the corresponding sites in pGEX4T-3 (Pharmacia Biotech) to produce afusion between GST and the portion of hamartin encoding the putative coiled-coil region to the stop codon. The Mr 72.000 GST-TSC1 fusion protein wasexpressed in BL21 and purified over glutathione-Sepharose 4B (Pharmacia

Biotech).Immunoblotting. The expression of hamartin and tuberin was analy/ed by

Western blotting of the lysates prepared from the cell lines. Equivalentamounts of protein were mixed with sample buffer [62.5 mM Tris-HCl (pH6.8). 2% SDS. 10% glycerol. 0.1% bromphenol blue, and 0.7 M ß-mercapto-ethanolj. boiled, and resolved on 10% SDS-polyacrylamide gels. Extracts of

normal brain and kidney were purchased from Clontech. The proteins wereelectrotransferred to Hybond enhanced chemiluminescence nitrocellulosemembranes (Amersham). and the membranes were blocked in PBS containing5% nonfat dried milk and 0.1% Tween 20. Primary antibodies (C20 at adilution of 1:1500. Ham-C at a dilution of 1:100. and L3 at a dilution of

1:2500) were incubated with the membranes for l h in PBS containing 5%nonfat dry milk and 0.1 % Tween 20. After washing in PBS with 0.25% Tween20. the membranes were blocked again and then incubated with an antirabbitperoxidase-linked secondary antibody (1:750) from Amersham. Proteins were

detected by enhanced chemiluminescence (Amersham).Immunoprecipitation. For preparation of whole cell extracts for immuno-

precipitation. the cells from a confluent T-125 flask were washed and collectedby scraping in lysis buffer [ 10 HIMTris-HCl (pH 7.6). 50 HIMNaCI, 30 mM Na

PP,. 50 mM sodium fluoride, and 2% NP40) supplemented at the time of usewith 1 mM sodium orthovanadate. 2.5 mM EDTA. 1 mM phenylmethylsulfonylfluoride, 10 /ng/ml leupeptin. 1 JUMpepstatin. and 0.23 unit/ml aprotinin. Celllysates for immunoprecipitation were also prepared in radioimmunoprecipita-

tion assay buffer (PBS with 1% NP40. 0.5% sodium deoxycholate. and 0.1%SDS) supplemented as described above. The cells were allowed to lyse on icefor 30 min and disrupted further by passage through a 21-gauge needle, and thelysate was cleared by microcentrifugation at 15.IXX)x ¡>for 20 min at 4°C.For

immunoprecipitation, aliquots of the cell lysates equali/ed for protein contentwere then incubated with primary antibody and a 1:1 mixture of proteinA-agarose and protein G-agarose (Life Technologies. Inc.) overnight at 4°C.

The immune complexes were collected by microcentrifugation at 15.(XX)x gfor 2 min at 4°Cand washed three times with lysis buffer. The immunopre-

cipitates were resuspended in SDS sample buffer, boiled for 5 min. andsubjected to 10% SDS-PAGE analysis.

Cell Fractionation. Cells from a confluent T-125 flask were removed by

scraping in 5.0 ml of PBS. and the cell pellet was resuspended in 1.0 ml ofhypotonie buffer [10.0 mM Tris-HCl (pH 7.4). 1.0 mM DTT, 0.2 mM MgCU

5 mM KC1, and 16 jig/ml aprotinin]. The cells were allowed to swell on ice for5 min and lysed by 20 strokes in a Dounce homogenizer. The nuclei wereremoved from the homogenate by centrifugation at l.(XX) X g for 15 min at4°C.Sucrose was added to the supernatant to a final concentration of 0.25 M.

and EDTA was added to a final concentration of 1 mM. followed by centrifugation at 100.000 x K for l h at 4°C.The pellet from the 100.000 x g spin,

which contained plasma membranes as well as mitochondria. Golgi, lyso-

somes, and vesicles, was resuspended in 2.0 ml of lysis buffer. The volume ofthe supernatant from the 100.000 X g spin, which contained cytoplasmicproteins, was brought to 2.0 ml with lysis buffer (final concentration. 1x lysisbuffer). The protein concentration of each fraction was normalized, and 50 figof each fraction were subjected to 10% SDS-PAGE. After electrophoresis, the

proteins were transferred to nitrocellulose and immunoblotted as describedabove. To confirm that cellular fractionation was effective, the ß-tubulinlevels

in lysates from each fraction were analyzed.Immunofluorescence. The 293 cells were grown on glass coverslips and

fixed with methanol at —¿�20°Cfor 5 min. Fixed cells were blocked with 10%

calf serum in PBS for 20 min at room temperature. After washing, the cellswere stained with a 1:25 dilution of anti-hamartin antibody for I h at 37°C.

Control coverslips were incubated in PBS without antibody. After three washes

with PBS with 1% BSA. the cells were incubated with Texas Red-X-conju-

gated AffmiPure goat antirabbit IgG (Jackson ImmunoResearch Labs. Inc.),washed four times in 1% BSA in PBS. and mounted. Confocal microscopy wasperformed using a Nikon Optiphot II confocal microscope.

Results

The Ham-C Antibody Recognizes a MT 130,000 Protein. Wehave generated an antibody to hamartin, Ham-C. that recognizes a Mr

130,000 protein, the predicted molecular weight of hamartin, in human kidney (Fig. \A) and brain (Fig. Iß)protein extracts. This proteinwas not detected when the blot was incubated with preimmune serum(data not shown). Hamartin and tuberin are not homologous to eachother, and there was no evidence of cross-reactivity between hamartinand tuberin antibodies. The specificity of Ham-C was established byits recognition of a A/r 72,000 TSC 1-GST fusion protein composed ofthe COOH-terminal region of TSC1 (amino acid residues 730-1164)fused to GST (Fig. 1/4). Ham-C does not recognize GST alone. The

intensity of the Mr 130,000 band detected in the brain and the Mr72,000 TSC 1-GST fusion protein was greatly reduced after preincu-

bation of the antibody with the peptide against which it was generated(data not shown).

Hamartin Is Expressed in Human and Rat Cell Lines, Includingthe Eker Tumor-derived Cell Line. Several human and rat cell lineswere tested for hamartin expression. The A/r 130.000 hamartin proteinwas detected in the human embryonic kidney cell line 293, the ratkidney epithelial cell line NRK-52E, and the rat fibroblast cell lineRatl-R12 (Fig. \B). The detection of hamartin by Ham-C in the rat

cell lines demonstrates that the antibody recognizes the homologousrat protein. Hamartin expression was also seen in the Eker rat tumor-

GST-K GST TSC1

180 kD -

130 kD -

72 kD-

B

130 kD •¿�

C180 kD -

29 kD-

Brain Eker NRK-S2E RAT1-R12

Hamartin

Bruin Eker NRK-52E RAT1-RI2

Tuberin

Fig. 1. Characterization of humartin antibod) Ham-C and the expression of hamartinand tuhcrin in human and ral cell lines. A. the Ham-C antibody recognizes a M, I30.000protein in normal human kidney IA'). The antibody also recognizes a Mr 72.000 GST-

fusion protein containing the COOH-terminus of TSCi iGST-TSCl). hul not GST alone,which has a molecular weight of 29.000. 0. hamartin is expressed in human hrain and incell lines derived from the human kidney (293). rat kidney (NRK-52E). and ral lìhrohlasis(Ratl-Rl2). Hamartin is also expressed in the Eker rat Junior-derived cell line ERCI8M,which lacks tuberin. C, the M, 180.000 tuberin protein is detected in human hrain and inthe 293. NRK-52E. and RAT I-R12 cell lines. Tuberin is not detected in the Eker cells.

4767

Research. on February 27, 2021. © 1998 American Association for Cancercancerres.aacrjournals.org Downloaded from

SI W hLLULAR LOCALIZATION OF HAMARTIN

Cell Line: 293 Eker NRK-52E

IP: L3 C20 L3 C20 L3 C20

IKIlkl)Tuberin

IP: L3 C20 L3 C20 L3 C20

Fig. 2. Coimmunoprecipitaiion of hamartin wilh lubcrin. Immunoprecipitations wereperformed on Ihe human (293) and rat (NRK-52E) kidney cell lines and the Eker(umor-derived cell line (ERC18M) using two tuherin antibodies (L3 and C20). A. the L3antibody recognizes a M, 180.000 protein corresponding to the molecular weight ofluherin after immunoprecipitation with the L3 and C20 antibodies in both the 293 and theNRK-52E cell lines, hut not in the Eker cell line. lì.the immunoprecipitates from the 293and NRK-52E cells, but not those from the Eker cells, contained a M, I30.000 proteinrecogni/.ed by the Ham-C antibody.

derived cell line ERC18M (Fig. la). The Eker cell line lacks tuberinexpression (Fig. 1C), suggesting that hamartin is stable in the absenceof luberin.

Hamartin Coimmunoprecipitates with Tuberin. The clinicalmanifestations of TSC1- and TSC2-linked disease are nearly identical.

suggesting that hamartin and tuberin may function in the same cellularpathway. To determine whether hamartin and tuberin associate witheach other, we performed Coimmunoprecipitation experiments usingboth the L3 and the C20 tuberin antibodies. To verify that theantibodies worked efficiently for immunoprecipitation. blots of theimmunoprecipitates from the human (293) and rat (NRK-52E) kidney

cell lines were incubated with the tuberin antibodies L3 (Fig. 2/t) andC20 (data not shown). A M, 180.000 protein, corresponding to themolecular weight of tuberin, was precipitated by both the L3 and theC20 (Fig. 2A) antibodies from the 293 and NRK-52E cells. As

expected, tuberin was not present in the immune complexes recoveredfrom the Eker cell line.

A Mr 13(),(XK)protein, corresponding to the molecular weight ofhamartin, was detected by Ham-C in the tuberin immunoprecipitatesfrom both the human (293) and rat (NRK-52E) kidney cell lines (Fig.

20). Hamartin was not detected in immunoprecipitates of lysates fromthe Eker cell line ERC18M (Fig. 2B). Neither tuberin or hamartin wasdetected in the immunoprecipitates when the C20 antibody wasblocked by preincubation with the peptide against which it wasgenerated (data not shown). The Ham-C antibody did not immuno-

precipitatc hamartin (data not shown).Hamartin Is in the Membrane/Particulate Fraction of Cultured

Cells. The postnuclear supernatant from 293 cells was fractionated bya 1(K),(MX)x f<ccntrifugation into membrane/particulate (P100) and

cytoplasmic (SIOO) components. Western blot analysis demonstratedthat both hamartin and tuberin were recovered in the P100 fraction(Fig. 3iA). ß-Tubulinwas present primarily in the SIOO fraction, as

expected (Fig. 3/4). The P100 fraction contains the plasma membraneand intraccllular organellcs including the Golgi apparatus, mitochondria, lysosomes, and vesicles.

To determine whether hamartin is present in the P 100 fractionbecause of its interaction with tuberin. we performed cell fractionationon lysates from the Eker tumor-derived cell line ERC18M. Hamartin

remained in the P 100 fraction of the Eker cells (Fig. 3ß),which lackfunctional tuberin.

Subcellular Localization of Hamartin by Confocal Microscopy.The cellular fractionation results indicated that both hamartin andtuberin are present in the membrane/particulate fraction. Using con-

focal microscopy, we studied the subcellular localization of hamartin.Subconfluent 293 cells were stained with the Ham-C antibody and

analyzed by indirect immunofluorescence. The staining was cytoplasmic (Fig. 4, A and B), without significant perinuclear clustering thatwould have suggested Golgi localization. At higher magnifications,the staining had a punctate appearance (Fig. 4, C and D), which isconsistent with localization to cytoplasmic vesicles.

Discussion

The TSC1 gene was recently identified (3) after a positional cloningeffort that spanned 10 years. The sequence of TSCl, which contains acoiled-coil domain and a short putative transmembrane domain, has

provided few clues to its function. The first evidence pointing to apotential functional role for hamartin was its interaction with tuberin(14), suggesting that hamartin and tuberin are protein partners in acommon cellular pathway.

Although the function of tuberin is not completely understood, anumber of findings suggest that tuberin plays a role in vesiculartrafficking. These findings include its localization in the Golgi apparatus (13) and its interaction via a region near its COOH terminus withrabaptin-5, a small cytosolic GTPase-binding protein (12). Rabaptin-5

displays specific binding for the active form of the GTPase Rab5 thatis associated with early endosomes (15). Consistent with the interac-

Total SIOO P100

130 kD Hamartin

SIOO P100 Total

180 kDTuberin

Total SIOO P100

47 kD -fc- "/3-Tubulin

B

130 kD

Total SIOO P100

Hamartin

Fig. 3. Localization of hamartin and tuberin to the membrane/particulate fraction inhuman and ral cell lines. A. after a 100.000 x Äccntrifugation of the postnuclearsupernatant of 293 cells, both hamartin and tuherin were observed in the membrane/paniculate (PIOO) fraction. ß-Tubulinwas localized primarily to the cyloplasmic (SIOO)fraction, as expected. B. hamartin was localized to the P100 fraction of the Eker rat cellline ERC18M. suggesting that the localization of hamartin to the P100 fraction is notdependent on its interaction with tuberin.

4768

Research. on February 27, 2021. © 1998 American Association for Cancercancerres.aacrjournals.org Downloaded from

SUBCELLULAR LOCALIZATION OF HAMARTIN

60x 60x

:•

•¿�:

Fig. 4. Localization of hamartin to the cytoplasm of cultured cells. Hamartin ¡mmunofluorescence was detected in the cytoplasm of cultured 2')3 cells using unifocal microscopy(«and b. X60). At higher magnification, the staining had a punctate appearance, suggesting locali/ation to the cytoplasinic vesicles (<•and ¡I.>'.M)0|.

tion of tuberin with rabaptin-5, tuberin has been shown to have

functional GAP activity for Rab5 (12), and cells without tuberindisplay an increased rate of fluid-phase endocytosis (12). These data

suggest that tuberin may negatively regulate endocytic fusion eventsby converting Rab5 from its active GTP-bound state to an inactiveGDP-bound state. Whether the association of hamartin with tuberin

modulates or contributes to endosóme fusion is not known. Tuberinhas other functions that may be distinct from its potential role invesicular trafficking, including the regulation of cell growth arrest atthe G0-G, boundary (16) and the regulation of transcription mediated

by steroid hormone receptors (17).Recently, hamartin and tuberin were shown to associate at endog

enous protein levels in HeLa cells and in a human fibroblast cell line(14). The interaction domain was mapped using the yeast two-hybridsystem to the coiled-coil domains present in both proteins (14). In this

study, we have independently demonstrated the interaction of hamartin and tuberin at endogenous expression levels in both human and ratkidney-derived cell lines using tuberin and hamartin antibodies and

cell lines that are different from those in the previous report. Thisinteraction strongly supports a model in which hamartin and tuberinfunction in a common cellular pathway, underlying the nearly identical clinical phenotypes of TSC1- and TSC2-linked disease.

We also show here, for the first time, that hamartin is expressed inan Eker rat tumor cell line. ERC18M (12), which lacks functionaltuberin. Eker rats, which are a model of autosomal dominant inheritance of renal carcinoma, carry a 6.3-kb insertion into Tsc2 that

eliminates approximately 30% of the gene, including the GAP domain(18-20). The presence of hamartin in the Eker tumor cell line suggests

that the stability of hamartin is not dependent on its interaction withtuberin.

We found that hamartin is localized to the membrane/paniculate(100,000 X g) fraction of cultured cells. This traction contains theplasma membrane, mitochondria. Golgi, lysosomes, and vesicles. Thelocalization of hamartin to the P100 fraction was unexpected, given itsoverall hydrophilicity and single, short. 18-amino acid, putative trans-

membrane domain. In contrast, tuberin. which is also locali/.cd to themembrane/paniculate fraction (Ref. 11 and this report), has fourputative transmembrane domains. Because hamartin and tuberin interact with each other, the localization of hamartin in the P100 fractioncould result from its association with tuberin. To address this issue,we performed cellular fractionation on Eker cells, which lack functional tuberin. Hamartin remained in the P KM)fraction in the Ekercells, suggesting that the subcellular localization of hamartin may notrequire an interaction with tuberin.

The localization of hamartin to the membrane/paniculate fractionsuggested the possibility of colocalization with tuberin in the Golgiapparatus. However, by confocal imaging of cultured human kidneycells, we found that at endogenous expression levels, hamartin islocalized to the cytoplasm, with a punctate pattern of immunofluo-

rescence suggestive of vesicular localization. This is the first report ofthe subcellular distribution of hamartin at endogenous expressionlevels. A similar pattern of cytoplasmic immunofluoresccncc has beenreported when hamartin is overexpressed in Cos cells (14).

Taken together, the membrane/paniculate fraction localization andconfocal microscopy data suggest that hamartin may be localized tocytosolic vesicles. This localization could be either on the vesicular

4769

Research. on February 27, 2021. © 1998 American Association for Cancercancerres.aacrjournals.org Downloaded from

SUBCELLULAR LOCALIZATION OF HAMARTIN

membrane or within the vesicles. If hamurtin is localized to themembrane of specific vesicles while tuberin is localized to the Golgi,the interaction of humurtin and tuberin may facilitate vesicular docking. In the setting of functional inactivation of either hamartin ortuberin. this docking would not occur. Consistent with this model, anaccumulation of microvesicles is seen in certain tuberous sclerosistumor cells, including renal angiomyolipomas and pulmonary lym-

phangiomyomatosis (21).In conclusion, our data suggest that hamartin is associated with

cytoplasmic vesicles and point to a role for tuberin and hamartin invesicular trafficking. If aberrant vesicular trafficking leads to tumor-

igenesis in TSC, it would be a novel tumor suppressor gene mechanism.

Acknowledgments

We are grateful to Jon Boyd for help with the confocal microscopy:Alphonso Bellacosa. Andy Godwin, and Warren Kruger for reviewing themanuscript: and Alfred Knudson for helpful discussions and advice.

References

1. Gomez. M. Phenotypes of Ihe tuberous sclerosis complex with a revision of diagnostic criteria. Ann. N. Y. Acad. Sci.. 6/5; 1-7. 1991.

2. Roach. E.. Smith. M.. Hultenlocher. P.. Bhal. M.. Alcorn. D.. and Hawley. L.Diagnostic criteria: tuberous sclerosis complex. J. Child Neurol., 7: 221-224, 1992.

3. van Slegtenhorsl. M., de Hoogt, R.. Hermans. C.. Nellist. M.. Janssen. B.. Verhoef.S.. Lindhout. D., van den Ouweland. A., Halley. D.. Young. J.. Burley. M.. Jeremiah.S.. Woodward. K.. Nahmias. J.. Fox. M.. Ekong. R.. Osborne. J.. Wolfe. J.. Povey. S.,Sncl!. R.. Cheadle. J.. Jones. A.. Tachalaki. M.. Ravine. D.. Sampson. J.. Reeve. M..Richardson, P., Wilmer. R.. Munro. C.. Hawkins. T.. Sepp. T.. Ali. J.. Ward, S.,Green. A.. Yates, J.. Kwialkowska. J.. Henskc. E.. Short. M.. Haines. J.. Jozwiak. S..and Kwialkowski. D. Identification of the tuberous sclerosis gene 7~5C/ on chromo

some 9q34. Science (Washington DC). 277.- 805-808, 1997.

4. European Chromosome 16 Tuberous Sclerosis Consortium. Identification and characterization of Ihe tuberous sclerosis gene on chromosome 16. Cell, 75: 1305-1315.1993.

5. Carbonara. C.. Longa, L., Grosso. E.. Borrone. C.. Garre. M.. Brisigotti. M., andBigone. N. 9q34 loss of hetcrozygosity in a tuberous sclerosis astrocyloma suggestsa growth suppressor-like activity also for the TSCI gene. Hum. Mol. Genet., 3:1829-1832. 1994.

6. Green. A.. Smith, M.. and Yates. J. Loss of heterozygosity on chromosome 16pl3.3in hamartomas from tuberous sclerosis patients. Nat. Genet.. 6: 193-196, 1994.

7. Green, A., Johnson, P., and Yates. J. The tuberous sclerosis gene on chromosome9q34 acts as a growth suppressor. Hum. Mol. Genet., J: 1833-1834, 1994.

8. Henske. E.. Scheithauer. B.. Short. M.. Wollmann. R., Nahmias, J.. Homigold, N..van Sleglenhorst. M.. Welsh. C., and Kwiatkowski. D. Allelic loss is frequent intuberous sclerosis kidney lesions hut rare in brain lesions. Am. J. Hum. Genet.. 59:400-406. 1996.

9. Geist. R., and Gutmann. D. The tuberous sclerosis 2 gene is expressed at high levelsin Ihe cerebellum and developing spinal cord. Cell Growth Differ.. 6: 1477-1483,

1995.10. Wienecke. R.. Maize. J.. Reed. J.. deGunzburg. J.. Yeung, R., and DeClue. J.

Expression of the TSC2 product tuberin and its target Rapi in normal human tissues.Am. J. Pathol., 150: 43-50, 1997.

11. Wienecke. R., Konig. A., and DeClue. J. Identification of tuberin, Ihe tuberoussclerosis-2 product. Tuberin possesses specific Rapi GAP activity. J. Biol. Chem.,270: 16409^16414. 1995.

12. Xiao. G.. Shoarinejad. F.. Jin. F.. Golemis, E., and Yeung, R. The tuberous sclerosis-2

gene product, tuberin. functions as a RabSGAP in modulating endocytosis. J. Biol.Chem.. 272: 6097-6100. 1997.

13. Wienecke, R.. Maize. J.. Shoarinejad, F.. Vass. W., Reed, J.. Bonifacino, J.,Resau. J.. DeGunzburg. J.. Yeung. R.. and DeClue. J. Co-localizalion of the TSC2

product tuberin wilh its target Rapi in the Golgi apparatus. Oncogene, 13:913-923, 1996.

14. van Slegtenhorst. M.. Nellist. M.. Nagelkerken, B.. Cheadle. J.. Snell. R.. van denOuweland. A.. Reuser. A.. Sampson. J.. Halley. D.. and van der Sluijs. P. Interactionbetween hamartin and tuberin. the TSCI and TSC2 gene products. Hum. Mol. Genet..7: 1053-1057, 1998.

15. Stenmark. G., Vitale, G., Ullrich, O., and Zerial. M. Rabaptin-5 is a direct effectorof the small GTPase Rab5 in endocytic membrane fusion. Cell, 83: 423-432,1995.

16. Soucek. T., Pusch. O.. Wienecke, R., DeClue. J.. and Hengstschlager. M. Role of thetuberous sclerosis gene-2 product in cell cycle control. J. Biol. Chem.. 272: 29301-

29308. 1997.17. Henry. K.. Yuan. X.. Koszewski. N.. Ondo. H.. Kwiatkowski. D.. and Noonan. D.

Tuberous sclerosis gene 2 product modulates transcription mediated by steroidhormone receptor family members. J. Biol. Chem.. 27.?: 20535-20539. 1998.

18. Kobayashi. R.. Hirayama. Y.. Kobayashi, E.. Kubo. Y.. and Hiño.O. A germlineinsertion in Ihe tuberous sclerosis (Tsc2) gene gives rise to the Eker rat model ofdominantly inherited cancer. Nat. Genet.. 9: 70-74. 1995.

19. Yeung. R.. Xiao. G-H.. Jin. F.. Lee. W-C. Testa. J.. and Knudson. A. Predispositionto renal carcinoma in the Eker ral is determined by germ-line mutation of the tuberoussclerosis (TSC2) gene. Proc. Nati. Acad. Sci. USA. 91: 11413-11416. 1994.

20. Xiao, G., Jin, F.. and Yeung. R. Germ-line TSC2 mutation in a dominantly inheritedcancer model defines a novel family of rat intracisternal-A particle elements. Oncogene, //: 81-87. 1995.

21. Bacchi. C.. Bonetti. F.. Pea. M.. Martignoni. G., and Gown. A. HMB-45: a review.Appi. Immunohistochem.. 4: 73-85. 1996.

4770

Research. on February 27, 2021. © 1998 American Association for Cancercancerres.aacrjournals.org Downloaded from

1998;58:4766-4770. Cancer Res   Tracey L. Plank, Raymond S. Yeung and Elizabeth Petri Henske  Cytoplasmic VesiclesInteracts with Tuberin and Appears to Be Localized to

) Gene,TSC1Hamartin, the Product of the Tuberous Sclerosis 1 (

  Updated version

  http://cancerres.aacrjournals.org/content/58/21/4766

Access the most recent version of this article at:

   

   

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/58/21/4766To request permission to re-use all or part of this article, use this link

Research. on February 27, 2021. © 1998 American Association for Cancercancerres.aacrjournals.org Downloaded from