Degradation of Lung Adenoma Susceptibility 1, a Major Candidate Mouse Lung Tumor Modifier, Is Required for Cell Cycle Progression

We have previously identified murine lung adenoma susceptibility 1 (Las1) as the pulmonary adenoma susceptibility 1 candidate gene. Las1 has two natural alleles, Las1-A/J and Las1-B6. Las1 encodes an 85-kDa protein with uncharacter-ized biological function. In the present study, we report that Las1 is an unstable protein and the rapid destruction of Las1 depends on the ubiquitin-proteasome pathway. Las1 is a new microtubule-binding protein and Las1 associated with tubulin is not ubiquitinated. We further show that Las1-A/J is a more stable protein than Las1-B6. Las1 is expressed in the G 2 phase of the cell cycle and that ubiquitin-proteasome–mediated Las1 destruction occurs in mitosis. Overexpression of Las1-A/J inhibits normal E10 cell proliferation and induces a defective cytokinesis. The differential degradation of Las1-A/J and Las-B6 has important implications for its intracellular function and may eventually explain Las1-A/J in lung tumorigenesis.


Introduction
Lung cancer is the leading cause of death in the United States and worldwide (1).Lung cancer is largely induced by smoking; however, numerous studies also suggest there is a genetic basis for lung cancer (2,3).Studies of lung tumorigenesis using inbred mouse system have revealed that the A/J mice is more susceptible than the C57BL/6J (B6) mice toward lung carcinogens, and this has permitted identification of quantitative trait loci (QTL) by linkage analysis (4,5).The mouse pulmonary adenoma susceptibility 1 (Pas1) locus is one such QTL that has recently been fine mapped using congenic mice and shown to contain six candidate genes (4,6).Systematic characterization of Pas1 candidate genes has limited the candidates to the lung adenoma susceptibility 1 (Las1) and Kras2 genes (4).Of these two, only Las1 displays an allelic variant that leads to an amino acid change at residue 60.Las1-A/J carries Asn 60 and Las1-B6 carries Ser 60 .Preliminary studies have revealed that tumor cells expressing Las1-B6 exhibit slower rates of growth than Las1-A/J when cultured on plates and in xenograft mice (4).However, another study has shown that Las1-A/J inhibits cell growth compared with Las1-B6 (7).Therefore, the role of Las1 in lung tumorigenesis seems unclear and its molecular basis in lung cancer remains unexplored.
Ubiquitin-mediated degradation is the main nonlysosomal proteolytic pathway in eukaryotic cells (8).It plays a key role in eliminating misfolded proteins and disposing many short-lived regulatory proteins responsible for cell cycle progression, DNA repair, transcriptional regulation, signal transduction, apoptosis, and protein translocation (9,10).Ubiquitin is an abundant and highly conserved 76-residue protein and is covalently attached to a target protein at lysine residues.Polyubiquitination of a protein marks it for degradation.Two discrete steps are involved in ubiquitin-mediated protein degradation: conjugation of multiple ubiquitin molecules to the target protein and degradation of the polyubiquitin-tagged substrate by the 26S proteasome (11).Conjugation of ubiquitin is carried out by a sequence of three enzymes: an E1 ubiquitin-activating enzyme, an E2 ubiquitinconjugating enzyme, and an E3 ubiquitin ligase (12,13).Ubiquitination of a specific substrate is mainly regulated through modulation of a degradation signal, such as phosphorylation of a target protein, or through control of the activity of its cognate E3, such as by association with a specific E3 activator (13).
Microtubules are important cellular cytoskeletal structure in eukaryotic cells and are involved in many cellular processes, including maintenance of cell shape, cell polarity, intracellular transport, and mitosis (14)(15)(16).Microtubules are highly dynamic structures equilibrating between a/h-tubulin dimers and a/htubulin polymers (microtubules).The polymerization dynamics is fundamentally important to the intracellular functions of the microtubule cytoskeleton.During the mammalian somatic cell cycle, microtubules undergo dramatic rearrangements from breakdown of cytoplasmic microtubules and subsequent formation of the mitotic spindle in the G 2 -M transition to dissolution of the spindle and reformation of the cytoplasmic microtubules on the return to interphase (17)(18)(19).In addition, microtubules interact with a large number of microtubule-associated proteins (MAP), which either regulate microtubule dynamics and physical properties or are motor molecules able to move in a unidirectional manner along the surface of microtubules (20).Microtubules also undergo various post-translational modifications and therefore distribute differently and associate with distinct sets of MAPs in cells (21).
Cytokinesis is the last stage of cell division for cytoplasm and cell cortex partition.The major event in cytokinesis is to build a cleavage furrow, an actin-myosin contractile ring, which constricts inwards to partition the parent cell into two daughter cells (22,23).The actin cytoskeleton is a major player in cytokinesis because the cleavage furrow is formed by actin and myosin II; however, the correct positioning and assembly of the contractile ring requires the aster microtubules and mitotic spindle (22,23).Therefore, completion of cytokinesis requires proper coordination from the microtubule cytoskeleton to the actin cytoskeleton and the cell membrane.
In this study, we describe the characterization of a novel mouse lung tumor susceptibility gene Las1.We show that Las1 is a new microtubule-binding protein and is a ubiquitinated protein but that tubulin-bound Las1 does not get ubiquitinated.Las1 undergoes cell cycle-dependent expression in the G 2 phase.Las1-A/J and Las1-B6 display differential degradation by the 26S proteasome in mitosis with Las1-A/J being slower than Las1-B6.Consequently, Las1-A/J induces a cytokinesis defect in a normal immortal lung epithelial cell line.These observations lead us to hypothesize that Las1-A/J disturbs microtubule function in the cytokinesis phase of the cell cycle.Our results provide further insight into the role of Las1 in lung tumorigenesis.

Materials and Methods
Plasmid constructs and mutagenesis.Vectors pcDNA4-Las1-A/J and pcDNA4-Las1-B6 in which Las1-A/J or Las1-B6 allele was cloned into the HindIII and EcoRI sites of the pcDNA4/TO vector (Invitrogen).Vector pcDNA3-Flag-ubiquitin encoding an NH 2 -terminal Flag-tagged human ubiquitin was provided by Dr. K-L.Guan (University of Michigan, Ann Arbor, MI).
Immunoprecipitation and immunoblotting.Cells were lysed using lysis buffer [100 mmol/L Tris-HCl (pH 7.4), 200 mmol/L NaCl, 10 mmol/L EDTA, 10% sucrose, 1.25% NP40, protease and phosphatase inhibitors].Lysates were precleared with protein A-Sepharose (Amersham) for 1 h at 4jC and incubated with indicated primary antibodies.Immunoprecipitates were washed and resolved by SDS-PAGE, electroblotted onto 0.45 Amol/L Figure 1.Interaction of Las1 and tubulin.A, Las1 coimmunoprecipitates with h-tubulin.293T cells were transiently transfected with vector, Las1-A/J, or Las1-B6.Cell extracts were incubated with anti-h-tubulin antibody, and immunopellets were resolved by SDS-PAGE and analyzed by immunoblotting with anti-Las1 and anti-h-tubulin antibodies, respectively.B, residue 60 on Las1 is not required for Las1 binding to h-tubulin.293T cells were transiently transfected with vector, Las1-A/J, Las1-B6, or pcDNA4-Las1-D11-69.Preparation of lysates, anti-h-tubulin immunoprecipitation, and anti-Las1 immunoblotting was done as above.C, Las1 cosediments with microtubules.293T cells were transiently transfected with vector, Las1-A/J, or Las1-B6.Cells were lysed with a microtubule-stabilizing buffer.After centrifugation, pellets containing microtubules were dissolved in the same volume of SDS-PAGE sample buffer and sonicated.Samples from the supernatant (S ) and pellet (P) fractions were boiled and resolved by SDS-PAGE.Las1, microtubule, and tubulin were visualized by immunoblotting analyses with anti-Las1, antiacetylated a-tubulin, and anti-h-tubulin antibodies, respectively.D, Las1 colocalizes with microtubules in interphase cells.NIH3T3 and E10-3 cells were grown on coverslips.NIH3T3 cells on the right column were treated in culture with 2 Ag/mL nocodazole (Noc) for 1 h.All cells were then fixed with 4% paraformaldehyde, permeabilized with 0.5% Triton X-100, and then stained for Las1 (red ), h-tubulin (green ), and nucleus (blue ), as indicated.Mapping the tubulin interaction domain of Las1.A, schematic representation of Las1 and Las1 truncation mutants and a summary of the interaction between Las1, Las1 truncation mutants, and h-tubulin.Shaded box, a predicted coiled-coil region; numbers on the left, residues of Las1 retained in each mutant.D, an internal deletion; +, positive interaction determined by reciprocal coimmunoprecipitation (Co-IP ); À, negative interaction determined by reciprocal coimmunoprecipitation.ND, not determined.B, expression of Las1 and Las1 mutant proteins.293T cells were transiently transfected with vector, Las1-A/J, and Las1-A/J truncation mutants listed in A. The same amounts of cell lysates from each transfection were resolved by SDS-PAGE and analyzed by immunoblotting with anti-Las1 and anti-h-tubulin (internal loading control) antibodies, respectively.C, reciprocal coimmunoprecipitations of Las1-A/J or Las1-A/J mutants and h-tubulin.The cell lysates described in B were immunoprecipitated (IP ) with anti-h-tubulin followed by anti-Las1 immunoblotting (IB; top ) or anti-Las1 antibody immunoprecipitation followed by anti-h-tubulin immunoblotting analyses (bottom ).*, nonspecific bands.polyvinylidene difluoride membrane, and probed with indicated primary followed by secondary antibodies.Final visualization of protein signals was done with the enhanced chemiluminescent kit (Pierce).
Immunofluorescence and microscopy.Exponentially growing cells were seeded on 22-mm coverslips in six-well plates at 50% confluence and cultured for 24 h.To depolymerize microtubules, nocodazole was added to the culture with a final concentration of 5 Ag/mL for 1 h.Double immunofluorescence staining was done as follows.Cells were fixed with 4% paraformaldehyde/PBS for 10 min and permeabilized with 1% Triton X-100/ PBS for 10 min.The cells were then blocked with 3% bovine serum albumin/ 0.5% Tween 20/PBS for 1 h followed with anti-Las1 and/or anti-h-tubulin incubation for 1 h and then Alexa Fluor 594-conjugated anti-rabbit and/or Alexa Fluor 488-conjugated anti-mouse secondary antibodies (Molecular Probes) for 1 h.Coverslips were stained with 4 ¶,6-diamidino-2-phenylindole (DAPI) and mounted on slides.Images were captured by Olympus DP70 digital camera and analyzed using Olympus MicroSuite FIVE imaging software.
Measurement of Las1 half-life.COS-7 cells were transiently transfected with 0.5 Ag Las1-A/J or Las1-B6.At 36 h after transfection, cells were treated with 100 Ag/mL cycloheximide (Sigma).Cells were collected at the indicated time point and lysed with the lysis buffer.Lysates (20 Ag) were resolved by SDS-PAGE and analyzed by anti-Las1 and anti-h-tubulin immunoblotting.Films were scanned and the intensities of bands were analyzed using the software ImageJ, version 1.36b.
In vivo ubiquitination assay.COS-7 cells were transiently transfected with 1 Ag pcDNA3-Flag-ubiquitin along with 1 Ag of vector, Las1-A/J, or Las1-B6, respectively.At 36 h after transfection, 20 Amol/L of MG132 (Calbiochem) or DMSO were added into the cultures for additional 6 to 12 h as indicated.Cell lysates were subjected to immunoprecipitation using anti-Las1 antibody.Ubiquitin-Las1 conjugates were detected by immunoblotting against anti-Flag antibody (Sigma).
Cell cycle synchronization.Cell cycle synchronizations were done according to the protocols of Murray's lab. 1 Briefly, for G 0 synchronization, cells were cultured to confluent and then starved in 0.1% FBS for 48 h.For S-phase synchronization, G 0 cells were seeded at 70% confluence in 150-mm plates and cultivated in regular medium in the presence of 4 Ag/mL aphidicolin for 14 h.For G 2 -phase synchronization, the cells were continually cultured in regular medium in the presence of 4 Ag/mL Hoechst 33342 for 12 h.For M-phase synchronization, the G 0 -synchronized cells were cultivated in regular medium in the presence of 0.5 Ag/mL nocodazole for 30 h and then collected by shake off lose touched mitotic cells (25). 1 Bioprotocols (http://www.bio.com/protocolstools/).Figure 3. Ubiquitin-proteasome mediates Las1 destruction.A, Las1 is ubiquitinated (Ub ) and that tubulin-associated Las1 is not ubiquitinated.293T cells were transiently transfected with Las1-A/J (left) or Las1-A/J along with pcDNA3-FLAG-ubiquitin (right ).Cell lysates were immunoprecipitated with anti-Las1 followed by immunoblotting with anti-ubiquitin (left ) or anti-FLAG antibodies (right ), respectively.B, ubiquitinated Las1 is destructed through the 26S proteasome.COS-7 cells were transiently transfected with 1 Ag of vector, Las1-A/J, or Las1-B6.At 36 h after transfection, 20 Amol/L of MG132 or DMSO were added to the culture for additional 12 h.Cell lysates were subjected to immunoprecipitation using anti-Las1 antibody.Immunoprecipitates were analyzed by immunoblotting with anti-ubiquitin antibody.C, Las1-A/J is more stable than Las1-B6.COS-7 cells were transiently transfected with 1 Ag of vector, Las1-A/J, or Las1-B6 along with or without 1 Ag pcDNA3-FLAGubiquitin.At 36 h after transfection, cells were lysed and the same amounts of cell lysates were subjected to immunoblotting for anti-Las1 and anti-h-tubulin antibodies, respectively.D, half-life of Las1-A/J and Las1-B6.COS-7 cells were transiently transfected with 0.5 Ag of Las1-A/J or Las1-B6.At 36 h after transfection, the cells were treated with 100 Ag/mL cycloheximide and harvested at the indicated time points.Cell lysates (20 Ag) from each sample were analyzed with immunoblotting using anti-Las1 and anti-h-tubulin antibodies, respectively.The intensity of each band was digitalized by ImageJ software and normalized against h-tubulin at the same time point.The graph represents the remaining percentage of Las1 at the indicated time point by setting the remaining percentage at zero time point as 100%.Results are from one representative experiment of two done.

Results
Interaction of Las1 and tubulin.We sought to explore Las1 functions by determining which proteins form complexes with Las1.We identified tubulin as a binding protein of Las1-A/J and Las1-B6 using standard pull-down assay by passing through glutathione S-transferase (GST), GST-Las1-A/J, and GST-Las1-B6 glutathione-agarose beads with A/J mice lung extracts (Supplementary Fig. S1).We chose coimmunoprecipitation assay to confirm the interaction of Las1 and tubulin.As shown in Fig. 1A, both Las1-A/J and Las1-B6 were detected in the anti-h-tubulin immunoprecipitates but not in the absence of transfected Las1, confirming that the interaction of Las1 and tubulin is specific.There was less Las1-B6 than Las1-A/J coimmunoprecipitated with h-tubulin.We did reciprocal coimmunoprecipitation assays and got similar results (Supplementary Fig. S2).
We found that the same amounts of transfected expression constructs produced larger amounts of Las1-A/J than Las1-B6 despite the fact that these constructs differ only by one residue in the Las1.To investigate if there are some differences in Las1-A/J and Las1-B6 in binding to tubulin, we generated a mutant Las1-D11-69 in which residues 11-69, covering the predicted coiled-coil region and the single nonsynonymous residue between Las1-A/J and Las1-B6, were deleted.As shown in Fig. 1B, besides Las1-A/J and Las1-B6, Las1-D11-69 also coimmunoprecipitated with htubulin, suggesting that the coiled-coil region on Las1 is not required for tubulin binding.Therefore, Las1-A/J and Las1-B6 bind to tubulin equally.
Cosedimentation of Las1 and microtubules.To further distinguish whether Las1 is associated with tubulin and/or microtubules, we determined whether Las1 can coprecipitate with microtubules.293T cells were transiently transfected with empty vector and vectors encoding Las1-A/J and Las1-B6.Cells were lysed in a low-salt lysis buffer to stabilize microtubules.In this assay, proteins associated with microtubules are captured in the pellet fraction with microtubules.As shown in Fig. 1C, Las1-A/J and Las1-B6 were in the pellet fractions with microtubules.These results show that both Las1-A/J and Las1-B6 associate with microtubules.
Coimmunolocalization of Las1 and microtubules in interphase cells.A previous study using Myc-tagged Las1 localized Las1 to the cytosol in NIH3T3 and COS-7 cells with a diffuse staining pattern (4).In the current study, we immunostained Las1 and h-tubulin in NIH3T3 and E10-3 cells, a stable Las1-B6 cell line in which Las1-B6 is overexpressed in E10 cells.As shown in Fig. 1D, anti-Las1 immunostaining revealed a distribution of Las1 that colocalized with the microtubule cytoskeleton.To analyze whether the localization of Las1 is dependent on the integrity of microtubule cytoskeleton, we treated cells with nocodazole, a microtubule-depolymerizing drug.Our results revealed that disruption of the microtubule cytoskeleton completely abolished the Las1 fiber-like staining pattern.This result indicates that intact microtubules are required for Las1 association and further confirms that Las1 is associated with the microtubule network.
Identification of an internal Las1 fragment bound to Btubulin.The Las1 cDNA encodes a 730-residue protein.There is no known functional domain besides a predicted coiled-coil motif present at the NH 2 terminus.As both a/h-tubulin [isoelectric point (pI) 4.8/5.3]and Las1 (pI 5.3) are acidic proteins, any nonspecific binding via electrostatic interactions is unlikely.We chose Las1-A/J to delineate the tubulin-binding region of Las1.We generated a series of truncation mutants of Las1-A/J (Fig. 2A), and all these mutants but Las1 (195-263) were expressed (Fig. 2B).Interactions of the expressed Las1 mutants and tubulin were determined by coimmunoprecipitation assays.As shown in Fig. 2C, among the seven Las1 constructs examined, Las1 (1-263) was the shortest NH 2 -terminal fragment tested positive in the interactions.Las1  did not interact with h-tubulin, whereas Las1 (195-730) did interact with h-tubulin.These results suggest that the interaction domain might locate within residues 195-263 of Las1.Interestingly, Las1 (195-263) has a pI 9.3 and is very likely to be a region to interact with h-tubulin (pI 5.3).Unfortunately, there was no detectable expression of Las1 (195-263).Finally, Las1 (195-381) interacted with h-tubulin (Supplementary Fig. S3), which suggests that the residues 264-381 of Las1 may not be necessary for h-tubulin binding but may influence the stability or folding of Las1, particularly in the context of the smaller Las1 (195-263) fragment.Taken together, we conclude that the Las1 (195-381) is necessary and sufficient for h-tubulin binding.
Tubulin-associated Las1 is not ubiquitinated.The higher molecular weight smears of Las1 shown in Fig. 2B suggest that Las1 might be polyubiquitinated.To confirm this, anti-Las1 immunoprecipitates (Fig. 2C, bottom) were reprobed with antiubiquitin antibody and displayed robust polyubiquitin signals (Fig. 3A, left), indicating that Las1-A/J is a ubiquitinated protein.However, there were no detectable ubiquitin-Las1 signals from the Las1 coimmunoprecipitated with h-tubulin (Fig. 2C, top).These observations seem to suggest that Las1 associated with h-tubulin is not ubiquitinated.To confirm this, 293T cells were transiently cotransfected with pcDNA3-Flag-ubiquitin along with Las1-A/J or Las1-B6.Anti-Las1 immunoprecipitation followed by anti-Flag immunoblotting revealed robust Flag-ubiquitin-Las1 signals (Fig. 3A, right), whereas anti-h-tubulin immunoprecipitation followed by anti-Flag immunoblotting did not reveal any similar high molecular weight smears even after a much longer exposure (Supplementary Fig. S4, left).Las1 coimmunoprecipitated with h-tubulin was confirmed by anti-Las1 immunoblotting (Supplementary Fig. S4, right).These results confirm that Las1 is a Ubiquitinated Las1 is a target for proteasomal destruction.As polyubiquitination generally targets proteins for proteasomemediated destruction, we then determined the effect of the proteasome inhibitor MG132 (26) on the degradation of ubiquitinated Las1.COS-7 cells were transiently transfected with vectors encoding Las1-A/J and Las1-B6 and incubated in the presence or absence of MG132.We observed that addition of MG132 to cells significantly increased recovery of ubiquitinated Las1-A/J and Las1-B6 compared with the samples from mock (DMSO)-treated cells (Fig. 3B), suggesting that the proteasomal activity is required for ubiquitinated Las1 degradation.We also did similar assays in LM2 and E10 cells and got similar results (Supplementary Fig. S5), confirming that endogenous Las1 is polyubiquitin modified.Taken together, these results show that the degradation of Las1 is mediated by the 26S proteasome.
Las1-A/J is a more stable protein.We frequently observed larger amounts of Las1-A/J than Las1-B6 by transfecting cells with the same amounts of Las1-A/J and Las1-B6, suggesting that Las1-B6 degrades faster.To confirm this, COS-7 cells were transiently transfected with Las1-A/J and Las1-B6 along with or without pcDNA3-Flag-ubiquitin.The steady-state levels of Las1-A/J, when overproducing Flag-ubiquitin, were higher than Las1-B6 (Fig. 3C), suggesting that Las1-A/J is a more stable protein than Las1-B6.There were no obvious differences in the steady-state levels between Las1-A/J and Las1-B6 in the cells without co-overexpressing ubiquitin.We believe it is because of signal saturation.To further study the differential stabilities of Las1-A/J and Las1-B6, we determined the half-life of Las1-A/J and Las1-B6.As shown in Fig. 3D, inhibition of protein biosynthesis resulted in the loss of half the amount of total Las1-B6 within 2 h, whereas Las1-A/J was more stable with an approximate half-life of 6 h.These data agree with the above observations for Las1-A/J is degraded slower than Las1-B6.
Las1 undergoes cell cycle-dependent expression.Destruction of Las1 through the ubiquitin-proteasome pathway implies a potential regulation of Las1 expression during the cell cycle.To examine it, we did cell cycle analysis using E10-3 cells.As shown in Fig. 4A, Las1 fluctuated throughout the cell cycle.Las1 was not detected in G 0 cells, started to accumulate in early S phase, reached the highest level in the G 2 phase, and dropped to very low levels at mitosis.We also did similar experiments in E10 cells to determine whether the endogenous Las1 behaved similarly and got similar results (Supplementary Fig. S6).These results show that expression of Las1 is under cell cycle control.
Anti-Las1 immunoblotting revealed two Las1-B6 bands (Fig. 4A).Along with the fact that Las1-B6 carries Ser 60 and degrades faster, this suggests that Las1 may get phosphorylated on Ser 60 and recognized by a specific E3 ligase.To confirm this, we generated Las1-Asp 60 (mimicking constitutive phosphorylation) and Las1-Ala 60 (blocking phosphorylation) mutants and determined their steady-state levels.Las1-Asp 60 displayed a lower steady-state level than Las1-B6, but Las1-Ala 60 did not show a higher steady-state level even than Las1-B6 (Supplementary Fig. S7).These results suggest that conformational changes, and not phosphorylation at Ser 60 , account for the differential stabilities between Las1-A/J and Las1-B6.
Ubiquitin-proteasome-mediated Las1 degradation occurs in mitosis.To determine whether the ubiquitin-proteasome system is responsible for the quick degradation of Las1 in mitosis, E10-3 cells in exponential growth were synchronized in G 2 and mitosis in the presence or absence of MG132.As shown in Fig. 4B, with or without MG132 treatment, there was no Las1 high molecular weight smear appearing in the cells synchronized at G 2 ,  indicating that Las1 ubiquitination does not occur in the G 2 phase.However, upon MG132 treatment, high molecular weight forms of Las1 accumulated compared with Las1 from untreated mitoticsynchronized cells, indicating that destruction of Las1 in mitosis is mediated by the ubiquitin-proteasome system.
Overproducing Las1-A/J causes defective cytokinesis in E10 cells.We also generated E10-2 cells, a stable Las1-A/J cell line in which Las1-A/J is overexpressed in E10 cells.E10 and E10-3 cells, when cultured, reached confluence at a similar pace, but it took much longer time for E10-2 cells.To follow this, we seeded 8 Â 10 4 of E10, E10-2, and E10-3 cells into 100-mm plates.At day 5, E10 and E10-3 cells reached much higher density than E10-2 cells (Supplementary Fig. S8).In addition to having less cell numbers, E10-2 cells also displayed some morphologic changes from mostly a refractive nucleus with thin and round cytoplasm to few a flat, round, and enlarged morphology (Fig. 5A).Some E10-2 cells showed two refractive nuclear areas with close contact to each other and surrounded with cytoplasm (Fig. 5A, bottom).Anti-htubulin and DAPI staining of the E0-2 cells confirmed that these cells harbored two nuclei (Fig. 5B).We therefore examined >300 cells of E10, E10-2, and E10-3 at days 2 and 5. Less than 1% of E10 and E10-3 cells were binucleated compared with 9% (day 2) and 26% (day 5) for E10-2 cells (Fig. 5C).These results suggest a cytokinesis defect in E10-2 cells introduced by overproducing Las1-A/J.We also observed some shrunken dumbbell-shaped cells that were loosely attached and floating in the medium, suggesting that the double-nucleated E10-2 cells died eventually.

Discussion
Aberrant regulation of the cell cycle is one mechanism that leads to cancer.To understand how Las1 contributes to lung cancer, we investigate whether Las1 has characters of the cell cycle regulatory proteins.Our goal was to study the cell cycle-dependent expression of Las1 in a normal cell line.E10 is an immortalized normal lung epithelial cell line from BALB/c mice and has relative complete set of cell cycle control system (27).We chose E10 as a parental cell line and generated E10-3 for Las1 cell cycle analysis.Our data, consistent between the overexpressed and endogenous levels, revealed a clear cell cycle-dependent expression pattern of Las1 peaking in G 2 and declining in mitosis.Finding that ubiquitinproteasome mediated Las1 degradation in mitosis further strengthens the notion for Las1 being a cell cycle regulator.Our data suggest that Las1-A/J is a more stable protein than Las1-B6.Slower degradation of Las1-A/J in mitosis may disturb some normal processes in/after mitosis if Las1 plays a crucial role in cell cycle regulation.For example, ubiquitin-proteasome-mediated degradation of cyclin B is essential for progression of metaphase to anaphase.Blocking the degradation of cyclin B thus arrests cell cycle in metaphase (28).As predicted, E10-2 cells grew much more slowly than E10 and E10-3 cells.Light microscopy further revealed that many E10-2 cells (10-30%) are binucleated during exponential cell growth, suggesting a cytokinesis defect.These findings further prove that Las1 is a crucial cell cycle regulator.
We identified tubulin as Las1-interacting protein and further showed that Las1 associates with microtubules.Las1-A/J and Las1-B6 differ in one residue but display differential stabilities toward ubiquitin-proteasome-mediated degradation.However, our data suggest that there are no differences between Las1-A/J and Las1-B6 in binding to tubulin.We have further mapped the tubulin-binding domain of Las1 to residues 195-381, therefore further showing that residue 60 that differs between Las1-A/J and Las1-B6 is not required for tubulin binding.We show that the portion of Las1 associated with tubulin is not ubiquitinated.This result suggests that the association with microtubule and ubiquitination/degradation of Las1 happen sequentially with dissociation of Las1 from microtubules first.Expression in G 2 and colocalization with microtubules in interphase suggests that the interaction of Las1 and microtubules likely occurs in the G 2 phase.We have tried to determine whether Las1 binding stabilizes microtubules by treating cells with various concentrations of nocodazole but did not observe anything.Currently, the functional significance of the association of Las1 with microtubules remains unclear.The interaction of Las1 and tubulin may mediate important biological events.However, it may not be the reason for the differential susceptibilities of A/J and C57BL/6J mice to lung tumorigenesis because Las1-A/J and Las1-B6 bind to tubulin with equal affinity.
Differential degradation of Las1-A/J and Las1-B6 leads to opposite consequences in normal and tumor cells.As summarized in Fig. 6A, degradation of Las1 is required for cell cycle progression in normal cells, whereas it slows down proliferation in human (our recent study; data not shown) and mouse tumor cells (4).Taken together, our results lead to a working model in which the degradation of Las1 in tumor cells leads to activation of a yet unidentified downstream signaling pathway (X) that inhibits tumor cell growth, whereas this pathway is inactive (m) in normal cells (Fig. 6B).Identification and characterization of this pathway will further decipher the role of Las1 in lung tumorigenesis.

Figure 2 .
Figure 2.Mapping the tubulin interaction domain of Las1.A, schematic representation of Las1 and Las1 truncation mutants and a summary of the interaction between Las1, Las1 truncation mutants, and h-tubulin.Shaded box, a predicted coiled-coil region; numbers on the left, residues of Las1 retained in each mutant.D, an internal deletion; +, positive interaction determined by reciprocal coimmunoprecipitation (Co-IP ); À, negative interaction determined by reciprocal coimmunoprecipitation.ND, not determined.B, expression of Las1 and Las1 mutant proteins.293T cells were transiently transfected with vector, Las1-A/J, and Las1-A/J truncation mutants listed in A. The same amounts of cell lysates from each transfection were resolved by SDS-PAGE and analyzed by immunoblotting with anti-Las1 and anti-h-tubulin (internal loading control) antibodies, respectively.C, reciprocal coimmunoprecipitations of Las1-A/J or Las1-A/J mutants and h-tubulin.The cell lysates described in B were immunoprecipitated (IP ) with anti-h-tubulin followed by anti-Las1 immunoblotting (IB; top ) or anti-Las1 antibody immunoprecipitation followed by anti-h-tubulin immunoblotting analyses (bottom ).*, nonspecific bands.

Figure 4 .
Figure 4. Cell cycle-dependent expression of Las1.A, E10-3 cells in exponential growth were synchronized at each phase of the cell cycle as described in Materials and Methods.Cell lysates were subjected to immunoprecipitation followed by immunoblotting with anti-Las1 antibody.Arrowhead, position of Las1.HC, heavy chain of IgG; LC, light chain of IgG; AS, asynchronized cells.B, degradation of Las1 occurs in mitosis.E10-3 cells were synchronized in G 2 and M phase, respectively.MG132 (20 Amol/L) or DMSO was added to the culture for additional 6 h.Cell lysates were subjected to immunoprecipitation with anti-Las1 followed by immunoblotting analyses with anti-ubiquitin and anti-Las1 antibodies, respectively.

Figure 6 .
Figure 6.Model depicting the role of Las1 degradation in regulating cell proliferation.A, summary of the observations for Las1 degradation regulating cell proliferation in normal and tumor cells.B, a proposed model for how Las1 degradation regulates cell proliferation in normal and tumor cells.See the text for details.

Figure 5 .
Figure 5. Overexpression of Las1-A/J results in the formation of binucleate cells.E10-2 cells (8 Â 10 4 ) were seeded into 100-mm plate and cultured for up to 5 d.Live cell images were taken at days 2 and 5 during culture.A, representative image of E10-2 cells at day 5. B, E10-2 cells grown on coverslips were fixed and stained with anti-h-tubulin (green ) and DAPI (blue ).A merged image is shown.C, the percentage of binucleate E10-1, E10-2, and E10-3 at days 2 and 5 is presented.