Ras transformation uncouples the kinesin-coordinatedcellular nutrient responseElma Zaganjora,1, Lauren M. Weila, Joshua X. Gonzalesa, John D. Minnaa,b, and Melanie H. Cobba,2

aDepartment of Pharmacology and bHamon Cancer Center for Therapeutic Oncology Research, University of Texas Southwestern Medical Center, Dallas,TX 75390-9041

Contributed by Melanie H. Cobb, June 12, 2014 (sent for review March 13, 2014; reviewed by George S. Bloom and Anning Lin)

The kinesin family members (KIFs) KIF2A and KIF2C depolymerizemicrotubules, unlike the majority of other kinesins, which trans-port cargo along microtubules. KIF2A regulates the localization oflysosomes in the cytoplasm, which assists in activation of the mech-anistic target of rapamycin complex 1 (mTORC1) on the lysosomalsurface. We find that the closely related kinesin KIF2C also influen-ces lysosomal organization in immortalized human bronchial epi-thelial cells (HBECs). Expression of KIF2C and, to a lesser extent,KIF2A in untransformed and mutant K-Ras–transformed cells isregulated by ERK1/2. Prolonged inhibition of ERK1/2 activationwith PD0325901 mimics nutrient deprivation by disrupting lyso-some organization and decreasing mTORC1 activity in HBEC, sug-gesting a long-term mechanism for optimization of mTORC1activity by ERK1/2. We tested the hypothesis that up-regulationof KIF2C and KIF2A by ERK1/2 caused aberrant lysosomal position-ing and mTORC1 activity in a mutant K-Ras–dependent cancer andcancer model. In Ras-transformed cells, however, mTORC1 activityand lysosome organization appear independent of ERK1/2 andthese kinesins although ERK1/2 activity and the kinesins are re-quired for Ras-dependent proliferation and migration. We concludethat mutant K-Ras repurposes these signaling and regulatory pro-teins to support the transformed phenotype.

Complex signaling pathways are engaged to coordinate cellu-lar growth and proliferation as dictated by stimulatory signals,

such as abundant nutrients and growth factors, or inhibitory sig-nals, such as limiting nutrients or other cell stresses. Disturbancesin coordination of signaling events are often observed in pathol-ogies such as cancer (1). Some cancers use oncogenic mutations ofthe small GTPase K-Ras to promote their survival and pro-liferation (2–6). Mutant K-Ras activates central signaling cascades,among them the phosphatidylinositol 3-kinase (Ras-PI3K) and theextracellular signal-regulated kinase ERK1/2 (Ras-ERK) path-ways, to facilitate continued proliferation under suboptimal con-ditions and bypass inhibitory signals from energy stress pathways.Mechanistic target of rapamycin (mTOR), a serine/threonine

protein kinase in a PI3K-related family, is a central regulator ofevents leading to cell growth and proliferation (7–12). It carries outthis role primarily as part of a complex [mTOR complex 1(mTORC1)], and, in this capacity, it is exquisitely responsive tonutrient availability. In certain cancers, positive regulatory signals tomTORC1 can originate from mutant K-Ras, even in nutrient-poorsettings (13, 14). Because cancers often use mTORC1 to optimizecellular functions to promote growth and survival, we have com-pared how the Ras-ERK pathway influences nutrient sensing andmTORC1 in untransformed immortalized cells and in cancer cells.Kinesins are motor proteins that transport cargo along micro-

tubules. Currently, more than 40 kinesins have been found inmammals, most of which have primary functions as cargo carriers.Kinesin family member (KIF) 2A is one of a handful of kinesinsthat apparently do not transport cargo, but instead depolymerizemicrotubules. In HeLa cells, KIF2A was shown to regulate thelocation of lysosomes in the cytoplasm, which is thought to be es-sential for mTORC1 activation (15–18). We find that KIF2A andthe closely related kinesin KIF2C (also known as MCAK) alsocontrol lysosomal organization in human bronchial epithelial cells.

We recently determined that expression of KIF2C and KIF2Ais up-regulated by the Ras-ERK pathway (19). Therefore, weasked whether ERK1/2 and these kinesins are required to main-tain lysosomal organization in immortalized human bronchialepithelial cells (HBECs) and also whether the elevated expres-sion of KIF2C and KIF2A accounts for the nutrient insensitivityof lysosomal positioning and mTORC1 activity in a mutantK-Ras–dependent cancer and cancer model.

ResultsKIF2A and KIF2C Are Required for the Amino Acid-DependentLocalization of mTORC1 to Lysosomes in Immortalized HBEC. To ex-plore the relationships among KIF2A and KIF2C expression,ERK1/2, and lysosomal organization, we used a model system toassess signaling differences between immortalized cells beforeand after transformation under relatively well-defined conditions.We examined the characteristics of a nontumor human bronchialepithelial cell line (HBEC) that was immortalized by expressionof cyclin-dependent kinase 4 (CDK4) and telomerase (hTERT),yielding HBEC3KT (number distinguishes different cell donors)as described (20). These cells maintain epithelial morphology andexpress E-cadherin but not N-cadherin (19). We compared thebehavior of this immortalized untransformed cell line to an iso-genic non-small cell lung cancer (NSCLC) model that was gen-erated by introducing mutant K-Ras and a stable knockdown ofp53 to yield HBEC3KTRL53 (21, 22). We also examined thepatient-derived colorectal cancer cell HCT116, which has anoncogenic K-Ras mutation.To examine the effects of nutrient restriction, cells were placed

in Earl’s balanced salt solution (EBSS), which contains glucosebut lacks serum and amino acids. Because mTORC1 is, at least in

Significance

The kinesin family member 2C (KIF2C) that regulates microtu-bule depolymerization, is up-regulated in a number of humancancer cells in a Ras- and ERK1/2-dependent manner. In thisstudy, we find that KIF2C expression positively regulates sig-naling pathways downstream of Ras, ERK1/2, and mechanistictarget of rapamycin complex 1 in nontransformed cells to co-ordinate a cellular nutrient response. Transformation by on-cogenic Ras overrides the need for this level of organizationand repurposes ERK1/2 and KIF2C signaling events. Therefore,this study provides a link for how cytoskeletal transformationmay interconnect with cancer metabolism.

Author contributions: E.Z. and M.H.C. designed research; E.Z., L.M.W., and J.X.G. per-formed research; J.D.M. contributed new reagents/analytic tools; E.Z., L.M.W., J.D.M.,and M.H.C. analyzed data; and E.Z. and M.H.C. wrote the paper.

Reviewers: G.S.B., University of Virginia; and A.L., University of Chicago and ShanghaiInstitute of Biochemistry and Cell Biology.

The authors declare no conflict of interest.1Present address: Department of Cell Biology, Harvard Medical School, Boston, MA 02115.2To whom correspondence should be addressed. Email: melanie.cobb@utsouthwestern.edu.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1411016111/-/DCSupplemental.

10568–10573 | PNAS | July 22, 2014 | vol. 111 | no. 29 www.pnas.org/cgi/doi/10.1073/pnas.1411016111

Dow

nloa

ded

by g

uest

on

Dec

embe

r 13

, 202

0

part, activated through interaction with Rag small GTPases andthe Ragulator complex on lysosomes (16), we examined how re-moval of amino acids and serum affected lysosomes and mTORlocalization. After 60–90 min of starvation in EBSS, lysosomes inHBEC3KT lost the typical perinuclear concentrated pattern andinstead gained a more diffuse and symmetric distribution, similarto what has been observed in HEK293 and HeLa cells followingremoval of amino acids (15, 16, 23) (Fig. 1 A and B). mTOR wasalso dispersed and less concentrated on lysosomes in HBEC3KTcells, and mTORC1 activity was also reduced (Fig. 1C), as mea-sured by a decrease in phosphorylation of ribosomal protein S6.ERK1/2 were also less active after an hour in EBSS. Readdition ofamino acids or complete medium with amino acids and serumreactivated both ERK1/2 and mTORC1 (Fig. 1C).Knockdown studies revealed that the microtubule-depolymerizing

kinesin KIF2A is required for lysosomal organization and mTORC1positioning on lysosomes (15). We recently found that KIF2Aand the closely related kinesin KIF2C are both up-regulated inmany cancers and that their persistent expression is supported byoncogenic K-Ras and ERK1/2 (19). As previously found in HeLaand 293 cells, depletion of KIF2A from HBEC3KT preventedtypical lysosomal organization, and mTOR also remained diffuse

and less well-localized with lysosomes (Fig. S1 A, B, and F). BecauseKIF2C has many functions that overlap with KIF2A (19), we testedits effects as well and found that depletion of KIF2C also preventedrestoration of lysosomal organization or mTOR activity by aminoacids (Fig. 1 A–C and Fig. S1 D and G), suggesting that more thanone kinesin participates in lysosomal positioning. Although KIF2AsiRNA and KIF2C siRNA caused a reduction in mTORC1 activityin amino acid-replete cells, knockdown of either or both had littleor no effect on mTORC1 activity in cells in serum-containingmedium (Fig. 1C and Fig. S1 F and G). ERK1/2 activation inHBEC3KT was also reduced following KIF2A or KIF2C siRNA.

Inhibition of the ERK1/2 Pathway Alters Lysosome Positioning andReduces mTORC1 Activity in Immortalized Cells. mRNAs encodingKIF2C and, to a lesser extent, KIF2A are decreased in bothHBEC3KT and HBEC3KTRL53 by extended exposure toPD0325901, which inhibits the enzymes MEK1 and MEK2 thatactivate ERK1/2 (19). These and other studies indicated thatERK1/2 stimulate KIF2A/C expression. Therefore, we exploredthe possibility that ERK1/2 may also influence organelle orga-nization in immortalized cells through their ability to control theexpression of proteins that alter microtubule dynamics includingthese kinesins. We inhibited ERK1/2 activation by exposure ofcells to the MEK inhibitor PD0325901 briefly or for up to 2 d.We showed previously that serum can stimulate ERK1/2 im-mediately after washing out PD0325901 even if cells have beentreated with the inhibitor for 2 d (19). In HBEC3KT, prolongedblockade of the ERK pathway with the MEK inhibitor had sig-nificant effects on mTOR and LAMP2 localizations, causingdispersal of both (Fig. 2 A and B). mTORC1 activity was notaffected by inhibition of ERK1/2 activation for 30 min but wasalmost completely inhibited in cells treated with the MEK in-hibitor for 2 d (Fig. 2C), suggesting that inhibition of mTORC1required changes in expression of protein. Similar observationswere made in HBEC30KT, a lung cell line from a different do-nor (Fig. S2A) and in H358, another cancer cell line (Fig. S2B).To confirm that these effects on mTORC1 and lysosomes

were comparable with those triggered by pathways that respondto limiting nutrients, we activated the AMP-activated proteinkinase AMPK with the AMP-like molecule AICAR (5-amino-imidazole-4-carboxamide-1-β-D-ribofuranoside). The activities ofAMPK and mTORC1 are intertwined on multiple levels (11, 24–29). AMPK can inhibit mTORC1 and can induce autophagy toaid in replenishing nutrients. Inhibition of mTORC1 as a conse-quence of AMPK activation occurred in HBEC3KT as early as6 h after treatment with AICAR (Fig. S3 A and B). Metformin,which activates AMPK by inhibition of complex I of the re-spiratory chain and perhaps other mechanisms (30), also inhibitedmTORC1 in HBEC3KT, but to a lesser extent in HBEC3KTRL53(Fig. S3 A–C). Exposure to metformin also appeared to dispersemTOR and lysosomes in HBEC3KT, but not in HBEC3KTRL53(Fig. S3C). These data are consistent with the idea that oncogene-induced transformation prevents a comprehensive shutdown ofsignaling pathways coordinated in normal cells in response tonutrient limitation.

Cancer-Cell Transformation Unlinks the Nutrient Sensitivity of mTORLocalization with Lysosomes. Fast growth and limited vasculariza-tion are just some of the environmental hurdles that provide achallenge to cancer-cell survival. Signaling in cancer cells must bealtered compared with normal cells to adjust to the high demandfor nutrients and other stresses. In HCT116 or HBEC3KTRL53in EBSS, mTOR remained concentrated in the proximity of thelysosomal marker LAMP2 (Fig. 3A). In nutrient-rich medium,these cells exhibited a more dispersed lysosomal phenotype thanuntransformed cells. Lysosomes in both cell types showed a range ofchanges in organization but did not acquire the morphology typicalof starved HBEC3KT. ERK1/2 were inactivated as quickly as 15

A

C

mTOR LAMP2

LAMP2

MERGE

MERGE siC

ontro

l

HBEC3KT

mTOR

siK

IF2C

EB

SS

AA

mTOR LAMP2

LAMP2

MERGE

MERGEmTOR

EB

SS

AA

mTOR LAMP2 MERGEB

ERK1/2

pERK1/2

pS6 (S235/236)

S6

siControl siKIF2A siKIF2C siKIF2A/C

EBSS AA M EBSS AA M EBSS AA M EBSS AA MHBEC3KT

KIF2A

1 9 32 4 2 49 1 3 35 1 4 37 pERK/ERK

1 3 11 1 1 8 1 2 11 1 2 9 pS6/S6

Fig. 1. KIF2C influences lysosomal organization and mTORC1 activity inHBEC3KT cells. (A) HBEC3KT cells were treated with control or KIF2C siRNAtwice 48 h apart for a total of 96 h. Cells were then starved for 90 min andrestimulated with amino acids (combination of each amino acid contained inL-glutamine-free DMEM plus 0.5 mM L-glutamine) for 30 min. Cells werecoimmunostained for endogenous mTOR (red) and endogenous LAMP2(green). (B) The eight-bit grayscale images from Fig. 1A were analyzed afterthresholding to measure the area of mTOR localization after restimulationwith amino acids by obtaining the ratio of mTOR staining per whole-cellarea. (C) Cells were starved treated as above with control, KIF2C, or KIF2AsiRNA, starved, and then stimulated with amino acids (AA) or fresh mediumwith serum (M). Lysate proteins were resolved on gels, transferred tomembranes, and immunoblotted with antibodies to the indicated proteins.

Zaganjor et al. PNAS | July 22, 2014 | vol. 111 | no. 29 | 10569

CELL

BIOLO

GY

Dow

nloa

ded

by g

uest

on

Dec

embe

r 13

, 202

0

min following transfer of HBEC3KT to EBSS (Fig. 3B); mTORC1was inactivated within 45 min in EBSS. On the other hand, inHBEC3KTRL53 or HCT116 cells, ERK1/2 phosphorylation wasnot reduced after an hour in EBSS, and relatively small decreasesin mTORC1 were observed (Fig. 3 B and C).The mTOR localization pattern in HBEC3KTRL53 was also

not substantially altered by prolonged inhibition with the MEKinhibitor PD0325901, nor was mTORC1 activity greatly reducedby the drug, suggesting that oncogenic transformation overridesthe changes in organelle compartments required for signalingin normal cells after growth pathway inhibition (Fig. 2 A–C andFig. S1 C and E). Also noted, ERK2 expression relative to ERK1is decreased in these cells, most clearly observed using an ERK2-specific antibody. These findings are consistent with the idea thatassociation of mTOR with lysosomes permits mTOR to maintain

an activated state in cancer cells even in nutrient-poor and growthpathway-inhibited conditions (Fig. S2C). Inhibition of MEK/ERKalso had no effect on mTORC1 in H358 lung-cancer cells thathave oncogenic K-Ras mutation as well as loss of p53 (Fig. S3B).In HBEC3KT, mTORC1 could be inhibited by various

means, including starvation, inhibition of growth pathways, en-ergy stress, and loss of the kinesins KIF2A or KIF2C. BecauseHBECKTRL53 and HCT116 cells were resistant to mTORC1inhibition by any of these interventions, we hypothesized thatoverriding mechanisms mediated by oncogenic transformationcontributed to the capacity of these cells to withstand thesechallenges. Activation of growth-factor pathways may providemore options to overcome amino acid depletion to maintainmTORC1 activity. To test this idea, we asked whether normalcells would be resistant to mTORC1 down-regulation due to lossof these kinesins in the presence of abundant nutrients and growthfactors. Following KIF2A or KIF2C siRNA in HBEC3KT, starvedcells were restimulated with medium containing amino acids andserum. In contrast to the results observed with amino acids alone,restimulation with amino acids plus serum activated mTORC1even in cells with reduced KIF2A and KIF2C expression (M lanes

B

HBEC3KT HBEC3KTRL53

ERK2pS6 (S240/244)

pERK1/230 min 2 d 30 min 2 d

PD PD

AmTOR mTOR

LAMP2 LAMP2

MERGE MERGE

Control PD

HBEC3KT

mTOR

LAMP2

MERGE

mTOR

LAMP2

MERGE

Control PD

HBEC3KTRL53

mTO

RLA

MP

2M

ER

GE

E

C

1 1 0.1 0.4 0.2 0.1 1 - 0.3 0.2 0.03 0.03 pERK/ERK

1 1 1.7 1.3 0.1 0.1 1.1 - 1.1 1.1 0.7 0.5 pS6/ERK

Fig. 2. Inhibition of ERK1/2 inhibits mTORC1 in immortalized bronchial epi-thelial cells but not in Raf-transformed cells. (A) HBEC3KT and HBEC3KTRL53cells were treated for 2 d with 100 nM PD0325901 and immunostained as inFig. 1A. (B) Images were analyzed as in Fig. 1 to measure the area of lysosomaldistribution (LAMP2 staining) by obtaining the ratio of the organelle area perwhole-cell area. (C) HBEC3KT and HBEC3KTRL53 cells were treated for 30 minor 2 d with 100 nM PD0325901. Lysate proteins were immunoblotted with theindicated antibodies.

A

B

C

HBEC3KTRL53

mTOR mTOR

Control 30 min EBSS

LAMP2

MERGE

LAMP2

MERGE

HCT116

mTOR

LAMP2

MERGE

mTOR

LAMP2

MERGE

Control 120 min EBSS

0 0 15 30 45 60 0 0 15 30 45 60 pS6 (S240/244)

pERK1/2

KIF2A

HBEC3KT HCT116

EBSS(min)

mTO

RLA

MP

2M

ER

GE

ERK1/2

pS6 (S240/244)

pERK1/2

S6

0 30 60 0 30 60

HBEC3KT HBEC3KTRL53

EBSS(min)

1 0 0.01 1 0.8 0.8 pERK/ERK

1 0.8 0.2 1 0.7 0.5 pS6/S6

1 1 1.5 1.1 0.4 0.1 1.1 0.9 1.1 0.9 0.7 0.5 pS6/KIF2A

0.9 1.1 0.01 0.04 0.02 0.12 0.9 1.1 1.7 0.7 1.1 1 pERK/KIF2A

Fig. 3. Cancer cells with K-Ras mutations are less sensitive to starvation thannormal cells. (A) HBEC3KTRL53 and HCT116 cells were starved in EBSS for30 min or 120 min, as indicated, and coimmunostained as in Fig. 1A. (B)HBEC3KT and HBEC3KTRL53 cells were starved in EBSS for 0 min, 30 min, or60 min. Lysate proteins were resolved on gels followed by immunoblottingwith the indicated antibodies. (C) HBEC3KT and HCT116 cells were starved inEBSS for the specified times and immunoblotted as above.

10570 | www.pnas.org/cgi/doi/10.1073/pnas.1411016111 Zaganjor et al.

Dow

nloa

ded

by g

uest

on

Dec

embe

r 13

, 202

0

in Fig. 1C). Consistent with these results, mTOR was more lo-calized to lysosomes in the presence of serum plus amino acidseven following loss of KIF2A or KIF2C in both HBEC3KT andHBEC3KTRL53 (Fig. S1H). Loss of KIF2A or KIF2C was dis-pensable for restimulation of ERK1/2 and mTORC1 in HBEC3KTor HBEC3KTRL53 by EGF, suggesting that KIF2A/C may beimportant for signaling through amino acid-regulated pathwaysbut not in the presence of high concentrations of growth factors(Fig. S4 A–C).

Elevated Expression of KIF2A and KIF2C Increased mTORC1 and ERK1/2Activity in HBEC and HeLa Cells. In HBEC3KTRL53, loss of KF2Aor KIF2C had little effect on the localization of lysosomes ormTOR whether or not amino acids were present. We examinedheterologous expression of KIF2C in HBEC3KT. BecauseKIF2A expresses poorly in HBEC cells, we also examined theeffects of forced KIF2A/C expression in HeLa cells. Expressionof KIF2A and KIF2C increased mTORC1 and ERK1/2 activitiesin cells in medium or EBSS (Fig. 4 A–C). These results supportthe idea that changes in KIF2A/C expression affect the lysosomalcompartment and mTOR activity.

Effects of Depletion of K-Ras and KIF2C on Autophagy. ReducedmTORC1 activity often results in autophagy to replenish nutrientsneeded for survival. mTORC1 usually inhibits autophagy underconditions in which nutrients such as glucose and amino acids orgrowth factors are abundant by phosphorylating the kinase ULK1,which is involved in autophagosome membrane formation (31).One indicator of accelerated autophagy is the conversion of theubiquitin-like protein LC3-I to its lipidated form, LC3-II, whichaccumulates on autophagosome membranes (32, 33). LC3-II wasincreased in HBEC3KTRL53 (Fig. 5A). Because a number of Rasfunctions are executed by PI3K, we tested the extent to which thispathway contributes to sustained autophagy. Accumulation ofLC3-II was decreased by LY294002, which inhibits the PI3Kpathway, and more strongly by the MEK inhibitor PD0325901,suggesting that oncogenic Ras can promote autophagy throughboth of these effectors (Fig. S5A). We hypothesized that de-pleting mutant K-Ras would reverse the autophagy observed inK-Ras–transformed cells. Instead, autophagy was further en-hanced by K-Ras knockdown, as assessed by loss of p62,a marker of autophagic flux, which is sequestered by autopha-gosomes and degraded in lysosomes when autophagy is induced(Fig. 5B and Fig. S5B) (34, 35). Because loss of KIF2A wassuggested to induce autophagy, as a consequence of decreasedmTORC1 activity (15), we tested whether loss of KIF2C alsopromotes autophagy. Even in HBEC3KTRL53, where loss ofKIF2C had little effect on mTORC1 activity, we observed anenhancement of autophagy, as assessed by a decrease in p62(Fig. 5C, lanes 1 and 2; compare lanes 7 and 8).

DiscussionWe demonstrate that prolonged inhibition of the ERK1/2 pathwayalters lysosomal distribution and suppresses mTORC1 in a man-ner that mimics nutrient deprivation. ERK1/2 are well known tohave regulatory inputs via TSC1/2 and Raptor to growth factor-stimulated mTOR activities both directly and through the down-stream effector Rsk (7–12, 36). ERKs have also been shown to betargeted to endosomes through a MEK partner 1 (MP1, geneLAMTOR3) complex (37), a component of the Ragulator com-plex, which is required for mTORC1 activity at the lysosome (16).Here, we define a long-term requirement for ERK1/2-dependentgene regulation in control of lysosomal organization in normalcells that is necessary for optimum mTORC1 performance.Among ERK1/2-regulated genes in both untransformed and

mutant K-Ras–transformed cells are the kinesins KIF2A andKIF2C (19). These KIFs are overrepresented in lung tumors, areinduced by K-Ras in an ERK1/2-dependent manner, and enhance

migration of Ras-transformed cancer cells. Because KIF2A waspreviously shown to be important for mTORC1 accumulation atlysosomes, we hypothesized that the insensitivity of mTORC1 andlysosome organization to growth factor and amino acid depriva-tion in Ras-transformed cells resulted from high expression ofKIF2A/C. In agreement with published work (15), we establishedthat KIF2A is important for lysosomal positioning in the im-mortalized lung-cell model and further showed that the relatedkinesin KIF2C also performs this function. Knocking down eitherof these kinesins inhibited mTORC1 stimulation by amino acidsand altered lysosomal organization in HBEC.In contrast, neither ERK1/2 activation nor these kinesins

were required for lysosomal organization and mTORC1 activityin Ras-transformed cells. The kinesin requirement could alsobe bypassed in untransformed cells under nutrient-repleteconditions in the presence of high concentrations of growthfactors found in serum. Despite the fact that these proteinshave the capacity to induce a functional organization of lyso-somes in untransformed cells, in Ras-transformed cells, thesekinesins are disconnected from this function. Wild-type andmutant K-Ras appear to induce morphological changes andperpetuate mTORC1 activity by distinct mechanisms in un-transformed and cancer cells. In addition, a signaling hierarchyis apparent in untransformed cells that leads to mTORC1 ac-tivation with a requirement for KIF2A and KIF2C that isdetected only in growth factor-poor conditions. Because two

ERK1/2pS6 (S240/244)

FLAG

pS6 (S235/236)

pERK1/2

EBSS EBSSControl FLAG-KIF2A

HeLa

HBEC3KTKIF2C

pS6 (235/236)ERK1/2

pERK1/2

PD PDControl FLAG-KIF2C

HeLa

FLAG

pS6 (S235/236)

pERK1/2

PD PDControl FLAG-KIF2A

ERK1/21 1.6 0.04 - 4.8 6.8 - 0.4 pERK/ERK

0.4 1.6 0.2 0.2 3.7 3.6 0.3 0.3 pERK/ERK

1.1 0.9 0.9 0.9 0.8 0.8 0.6 0.4 pS6/ERK

0.9 1.1 0.9 1 2 1.2 1.6 1 pS6/ERK

1 1 1.5 1.8 3.2 2.8 4.8 4.1 1.9 1.9 2.4 2.4 pERK/ERK

1.2 0.8 0.05 0.08 2 2 0.6 0.6 2.3 2.3 0.2 0.08 pS6/S6

EBSSFLAG-KIF2C

A

B

C

Fig. 4. KIF2A and KIF2C overexpression in HeLa and HBEC3KT activatespERK and pS6. (A) FLAG-KIF2A were expressed in HeLa cells that werestarved in EBSS or for 90 min. (B) Cells expressing FLAG-KIF2A were treatedwith PD0325901 for 90 min. (C) FLAG-KIF2C was expressed in HBEC3KT cellsthat were treated with PD0325901 for 90 min. Lysate proteins were resolvedon gels, followed by immunoblotting with the indicated antibodies.

Zaganjor et al. PNAS | July 22, 2014 | vol. 111 | no. 29 | 10571

CELL

BIOLO

GY

Dow

nloa

ded

by g

uest

on

Dec

embe

r 13

, 202

0

kinesins share this function, we considered the possibility thata third kinesin, closely related to KIF2A and -C, may alsocollaborate to position organelles. Expression of this third mi-crotubule depolymerizing family member, KIF2B, is low toundetectable in these cells, and its expression is increased, notdecreased, by the MEK inhibitor (19), suggesting that it is nota candidate for such a function.Microtubule instability, as is observed in K-Ras–transformed

cells, can be regulated by many proteins in addition to depoly-merizing kinesins. Other factors proposed to account for K-Ras–induced microtubule instability include stathmin, a destabilizingfactor, and reduced expression of the stabilizing factors discslarge 1 (Dlg1), RASSF1A, and adenomatous polyposis coli (APC)(38–41). ERK1/2 MAPKs themselves were linked to changes inmicrotubule dynamics many years ago (42–44). Lysosomes arelocalized at the microtubule organizing center (MTOC), requiremicrotubule dynamics to maintain proper organization (45), andwould be expected to have different distributions in cells dependingon microtubule dynamics. Ras-dependent changes in microtubuledynamics affected by these other molecules may bypass KIF2A/Cand account for distinct lysosomal distributions in HBEC3KTand HBEC3KTRL53.We also noted that manipulating KIF2A/C expression affected

ERK1/2 activity. Activity differences may also be a consequence

of the effects of these kinesins on microtubule dynamics. ERK ac-tivity was originally identified using a microtubule-associated pro-tein as substrate (46), and, in early studies, ERK activity was shownto be increased by microtubule depolymerization (47). From 30% to50% of the cytosolic ERK1/2 was estimated to be microtubulebound (48). Decreases in amounts of microtubule-depolymerizingkinesins should increase the average extent of microtubule poly-merization and thereby increase the number of microtubule bindingsites present at any given time to scaffold or sequester the kinases.We have found that KIF2A knockdown impaired ERK1/2 activa-tion. Likewise, overexpression of KIF2A/C in HeLa cells decreasedmicrotubule polymerization and increased ERK1/2 activity. Forthese reasons, we suggest that Ras-ERK induction of KIF2A/C mayform a positive loop to perpetuate some level of ERK1/2 activity byincreasing the unbound fraction of the proteins.The best-studied functions of KIF2A and KIF2C are in reg-

ulation of the mitotic spindle during mitosis (49–51). Despite thelarge number of kinesins in cells, KIF2A/C also have multiplefunctions in interphase cells, including effects on migration andautophagy (19, 52). As noted for loss of KIF2A (15), loss ofKIF2C also induces autophagy even in HBEC3KTRL53 in whichmTORC1 remains activated. This finding was unexpected be-cause of the minimal effects of KIF2A/C on lysosomal organi-zation in Ras-transformed cells and also because drugs thatinhibit microtubule dynamics, such as nocodazole and taxol,prevent starvation-induced autophagy (53). It will be importantto assess further whether or not the induction of autophagy bydepletion of KIF2A/C is related to their ability to increase mi-crotubule dynamics, their impact on ATP concentration, or otheras yet unidentified actions.

MethodsAntibodies. The following antibodies were used: KIF2A (cat. no. ab37005,Abcam; cat. no. NB500, Novus Biologicals), a-tubulin (cat. no. T6199; Sigma),Actin (cat. no. MAB1501MI; Millipore), FLAG (cat. no. F3165; Sigma), pS6 (cat.no. 5364; Cell Signaling Technology), pERK (cat. no. M8159, Sigma; cat. no.4377; Cell Signaling), LC3 (cat. no. PM036; MBL), p62 (cat. no. SC28359; SantaCruz Biotechnology), LAMP2 (cat. no. ab25631; Abcam), pAKT (cat. no. 4060;Cell Signaling), KIF2C (cat. no. A300-807A; Bethyl Labs), mTOR (cat. no. 2983;Cell Signaling), Ras (sc-166691; Santa Cruz), and ERK1/2 (Y691) and ERK2(C357) as previously described) (54).

Cell Culture. Immortalized HBEC3KT, HBEC30KT, and HBEC3KT53 cells werecultured in keratinocyte serum-free medium (KSFM) (Invitrogen) supple-mented with 5 ng/mL epidermal growth factor and 50 μg/mL bovine pituitaryextract according to the manufacturer’s recommendations. HBEC3KTRL53,HCT116, and H358 were cultured in RPMI-1640 medium supplemented with5% (vol/vol) heat-inactivated FBS and 2 mM L-glutamine. Cells were grownat 37 °C in a humidified atmosphere of 5% CO2.

Immunofluorescence. Cells were fixed with 4% paraformaldehyde (vol/vol) inTris-buffered saline TBS for 10 min and permeabilized with 0.1% Triton X-100for 5 min. After blocking with 10% normal goat serum (vol/vol) at roomtemperature for 1 h, cells were incubated with the indicated antibodies at4 °C overnight. Cells were incubated with Alexa Fluor-conjugated secondaryantibody at room temperature for 1 h, mounted, and imaged. FluorescentZ-stacks (0.2 mm) were acquired and deconvolved using the Deltavision RTdeconvolution microscope. Images were processed using ImageJ software.The eight-bit grayscale images were analyzed after thresholding to measurethe area of lysosomal distribution (LAMP2 staining, Fig. 2A) by obtaining theratio of the organelle area per whole-cell area (Fig. 2B), or mTOR localizationafter restimulation with amino acids (Fig. 1A) by obtaining the ratio of mTORstaining per whole-cell area (Fig. 1B).

siRNA. Cells were transfected for from 48 h to 96 h as indicated withdsRNA oligonucleotides using Lipofectamine RNAiMax according to themanufacturer’s protocol (Invitrogen). The following target sequencesfor KIF2A were used: GAAAACGACCACUCAAUAA (Thermo Scientific)and GACCCTCCTTCAAGAGATA (Thermo Scientific). For KIF2C, the fol-lowing were used: GCAAGCAACAGGUGCAAGU (Thermo Scientific) and

p62ERK1/2

EBSS BafA EBSS BafA

siControl siKIF2CHBEC3KTRL53

ALC3 II

HBEC3KT HBEC3KTRL53ERK1/2

B

C

ERK1/2pS6 (240/244)

p62

pERK1/2

PD

siControl siK-Ras

LAMP2

PDHBEC3KTRL53

0.8 1.2 0.2 0.2 0.8 0.7 0.3 0.3 pERK/ERK

0.9 1.1 3.2 2.7 0.3 0.3 0.2 0.2 p62/ERK

0.9 1.1 1.2 0.9 0.7 1 1 1 pS6/ERK

0.9 1.1 2 2 1.3 2.1 2 2.3 LAMP2/ERK

Fig. 5. Loss of KIF2C induces autophagy. (A) LC3-II was immunoblottedin equal amounts of lysates from HBEC3KT and HBEC3KTRL53. (B) Ras wasdepleted from HBEC3KTRL53 with siRNA. Cells were then treated for 3 d with100 nM PD0325901, followed by immunoblotting of lysates as indicated. (C)KIF2C was depleted with siRNA from HBEC3KTRL53. Cells were starved for 12h in EBSS or treated with 10 nM Baflomycin A and immunoblotted withantibodies to p62. ERK1/2 was used as the loading control.

10572 | www.pnas.org/cgi/doi/10.1073/pnas.1411016111 Zaganjor et al.

Dow

nloa

ded

by g

uest

on

Dec

embe

r 13

, 202

0

GGCAUAAGCUCCUGUGAAU (Thermo Scientific). For Ras, the followingwas used: GGAGGGCUUUCUUUGUGUA (Thermo Scientific).

Cell Harvest. Cells were lysed on ice in 50mMHepes (pH 7.5), 150mMNaCl, 1.5mM MgCl2, 1 mM EGTA, 0.2 mM Na3VO4, 100 mM NaF, 50 mM β-glycer-ophosphate, 10% glycerol, 0.1% Triton X-100, 1.6 μg/mL aprotinin, 0.1 mMphenylmethylsulfonyl fluoride, and 10 μg/mL each of Nα-p-tosyl-L-lysinechloromethyl ketone, Nα-p-tosyl-L-arginine methyl ester, pepstatin A, andleupeptin. Lysates were frozen in N2 (liquid) and thawed on ice, followed bycentrifugation for 15 min at 16,000 × g in a microcentrifuge at 4 °C. Lysateswere subsequently boiled in Laemmli sample buffer (2% SDS, 10% glycerol,5% β-mercaptoethanol, 0.01% bromophenol blue, 50 mM Tris·HCl) andsubjected to polyacrylamide gel electrophoresis in SDS.

Immunoblotting and Li-Cor Imaging. Gels were transferred to nitrocelluloseand incubated with the indicated antibodies. Primary antibodies were

detected by fluorescently labeled secondary antibodies (Fluor 800-labeledIgG or Fluor 680-labeled IgG secondary antibody), using the LI-COR Odysseydual-color system. To quantify the relative intensity of bands, we subtractedthe background and measured the ratio of band of interest to total loadingcontrol. Ratios were set equal to 1. If duplicates are shown, their averagewasset equal to 1.

ACKNOWLEDGMENTS. We thank Stina Singel (Department of Cell Biology,University of Texas Southwestern Medical Center) and members of theM.H.C. laboratory for helpful discussions and Dionne Ware for administra-tive assistance. This work was supported by National Institutes of HealthGrants R37DK34128 (to M.H.C.) and P50CA70907 (to J.D.M.), a grant fromthe Cancer Prevention and Research Institute of Texas (to J.D.M.), WelchFoundation Grant 1243 (to M.H.C.), and a grant from the LongenbaughFoundation (to J.D.M.). For E.Z. and L.M.W., this article represents partialfulfillment for the PhD degree.

1. Vander Heiden MG, Cantley LC, Thompson CB (2009) Understanding the Warburgeffect: The metabolic requirements of cell proliferation. Science 324(5930):1029–1033.

2. Karnoub AE, Weinberg RA (2008) Ras oncogenes: Split personalities. Nat Rev Mol CellBiol 9(7):517–531.

3. White MA, et al. (1995) Multiple Ras functions can contribute to mammalian celltransformation. Cell 80(4):533–541.

4. Peifer M, et al. (2012) Integrative genome analyses identify key somatic driver mu-tations of small-cell lung cancer. Nat Genet 44(10):1104–1110.

5. Sekido Y, Fong KM, Minna JD (2003) Molecular genetics of lung cancer. Annu RevMed 54:73–87.

6. Pollock CB, Shirasawa S, Sasazuki T, Kolch W, Dhillon AS (2005) Oncogenic K-RAS isrequired to maintain changes in cytoskeletal organization, adhesion, and motility incolon cancer cells. Cancer Res 65(4):1244–1250.

7. Sengupta S, Peterson TR, Sabatini DM (2010) Regulation of the mTOR complex 1pathway by nutrients, growth factors, and stress. Mol Cell 40(2):310–322.

8. Yecies JL, Manning BD (2011) mTOR links oncogenic signaling to tumor cell metab-olism. J Mol Med (Berl) 89(3):221–228.

9. Avruch J, et al. (2009) Amino acid regulation of TOR complex 1. Am J Physiol Endo-crinol Metab 296(4):E592–E602.

10. Dazert E, Hall MN (2011) mTOR signaling in disease. Curr Opin Cell Biol 23(6):744–755.11. Inoki K, Kim J, Guan KL (2012) AMPK and mTOR in cellular energy homeostasis and

drug targets. Annu Rev Pharmacol Toxicol 52:381–400.12. Ma XM, Blenis J (2009) Molecular mechanisms of mTOR-mediated translational con-

trol. Nat Rev Mol Cell Biol 10(5):307–318.13. Guo JY, et al. (2011) Activated Ras requires autophagy to maintain oxidative me-

tabolism and tumorigenesis. Genes Dev 25(5):460–470.14. Guo JY, White E (2013) Autophagy is required for mitochondrial function, lipid metab-

olism, growth, and fate of KRAS(G12D)-driven lung tumors.Autophagy 9(10):1636–1638.15. Korolchuk VI, et al. (2011) Lysosomal positioning coordinates cellular nutrient re-

sponses. Nat Cell Biol 13(4):453–460.16. Sancak Y, et al. (2010) Ragulator-Rag complex targets mTORC1 to the lysosomal

surface and is necessary for its activation by amino acids. Cell 141(2):290–303.17. Flinn RJ, Yan Y, Goswami S, Parker PJ, Backer JM (2010) The late endosome is essential

for mTORC1 signaling. Mol Biol Cell 21(5):833–841.18. Narita M, et al. (2011) Spatial coupling of mTOR and autophagy augments secretory

phenotypes. Science 332(6032):966–970.19. Zaganjor E, et al. (2013) Ras regulates kinesin 13 family members to control cell mi-

gration pathways in transformed human bronchial epithelial cells. Oncogene,10.1038/onc.2013.486.

20. Ramirez RD, et al. (2004) Immortalization of human bronchial epithelial cells in theabsence of viral oncoproteins. Cancer Res 64(24):9027–9034.

21. Sato M, et al. (2006) Multiple oncogenic changes (K-RAS(V12), p53 knockdown, mu-tant EGFRs, p16 bypass, telomerase) are not sufficient to confer a full malignantphenotype on human bronchial epithelial cells. Cancer Res 66(4):2116–2128.

22. Sato M, et al. (2013) Human lung epithelial cells progressed to malignancy throughspecific oncogenic manipulations. Mol Cancer Res 11(6):638–650.

23. Wauson EM, et al. (2012) The G protein-coupled taste receptor T1R1/T1R3 regulatesmTORC1 and autophagy. Mol Cell 47(6):851–862.

24. Corradetti MN, Inoki K, Bardeesy N, DePinho RA, Guan KL (2004) Regulation of theTSC pathway by LKB1: Evidence of a molecular link between tuberous sclerosiscomplex and Peutz-Jeghers syndrome. Genes Dev 18(13):1533–1538.

25. Hahn-Windgassen A, et al. (2005) Akt activates the mammalian target of rapamycinby regulating cellular ATP level and AMPK activity. J Biol Chem 280(37):32081–32089.

26. Gwinn DM, et al. (2008) AMPK phosphorylation of raptor mediates a metaboliccheckpoint. Mol Cell 30(2):214–226.

27. Schneider A, Younis RH, Gutkind JS (2008) Hypoxia-induced energy stress inhibits themTOR pathway by activating an AMPK/REDD1 signaling axis in head and necksquamous cell carcinoma. Neoplasia 10(11):1295–1302.

28. Egan D, Kim J, Shaw RJ, Guan KL (2011) The autophagy initiating kinase ULK1 is reg-ulated via opposing phosphorylation by AMPK and mTOR. Autophagy 7(6):643–644.

29. Faubert B, et al. (2013) AMPK is a negative regulator of the Warburg effect andsuppresses tumor growth in vivo. Cell Metab 17(1):113–124.

30. Hardie DG, Ross FA, Hawley SA (2012) AMP-activated protein kinase: A target fordrugs both ancient and modern. Chem Biol 19(10):1222–1236.

31. Kim J, Kundu M, Viollet B, Guan KL (2011) AMPK and mTOR regulate autophagythrough direct phosphorylation of Ulk1. Nat Cell Biol 13(2):132–141.

32. Rubinsztein DC, Codogno P, Levine B (2012) Autophagy modulation as a potentialtherapeutic target for diverse diseases. Nat Rev Drug Discov 11(9):709–730.

33. Lamb CA, Dooley HC, Tooze SA (2013) Endocytosis and autophagy: Shared machineryfor degradation. BioEssays 35(1):34–45.

34. Wooten MW, et al. (2008) Essential role of sequestosome 1/p62 in regulating accu-mulation of Lys63-ubiquitinated proteins. J Biol Chem 283(11):6783–6789.

35. Ichimura Y, Kominami E, Tanaka K, Komatsu M (2008) Selective turnover of p62/A170/SQSTM1 by autophagy. Autophagy 4(8):1063–1066.

36. Roux PP, Ballif BA, Anjum R, Gygi SP, Blenis J (2004) Tumor-promoting phorbol estersand activated Ras inactivate the tuberous sclerosis tumor suppressor complex via p90ribosomal S6 kinase. Proc Natl Acad Sci USA 101(37):13489–13494.

37. Nada S, et al. (2009) The novel lipid raft adaptor p18 controls endosome dynamics byanchoring the MEK-ERK pathway to late endosomes. EMBO J 28(5):477–489.

38. Humbert PO, et al. (2008) Control of tumourigenesis by the Scribble/Dlg/Lgl polaritymodule. Oncogene 27(55):6888–6907.

39. van Es JH, Giles RH, Clevers HC (2001) The many faces of the tumor suppressor geneAPC. Exp Cell Res 264(1):126–134.

40. Donninger H, Vos MD, Clark GJ (2007) The RASSF1A tumor suppressor. J Cell Sci 120(Pt18):3163–3172.

41. Fotiadou PP, Takahashi C, Rajabi HN, Ewen ME (2007) Wild-type NRas and KRasperform distinct functions during transformation. Mol Cell Biol 27(19):6742–6755.

42. Gotoh Y, et al. (1991) In vitro effects on microtubule dynamics of purified Xenopus Mphase-activated MAP kinase. Nature 349(6306):251–254.

43. Drechsel DN, Hyman AA, Cobb MH, Kirschner MW (1992) Modulation of the dynamicinstability of tubulin assembly by the microtubule-associated protein tau. Mol BiolCell 3(10):1141–1154.

44. Harrison RE, Turley EA (2001) Active erk regulates microtubule stability in H-ras-transformed cells. Neoplasia 3(5):385–394.

45. Matteoni R, Kreis TE (1987) Translocation and clustering of endosomes and lysosomesdepends on microtubules. J Cell Biol 105(3):1253–1265.

46. Ray LB, Sturgill TW (1987) Rapid stimulation by insulin of a serine/threonine kinase in3T3-L1 adipocytes that phosphorylates microtubule-associated protein 2 in vitro. ProcNatl Acad Sci USA 84(6):1502–1506.

47. Shinohara-Gotoh Y, Nishida E, Hoshi M, Sakai H (1991) Activation of microtubule-associated protein kinase by microtubule disruption in quiescent rat 3Y1 cells. Exp CellRes 193(1):161–166.

48. Reszka AA, Seger R, Diltz CD, Krebs EG, Fischer EH (1995) Association of mitogen-activated protein kinase with the microtubule cytoskeleton. Proc Natl Acad Sci USA92(19):8881–8885.

49. Ganem NJ, Compton DA (2004) The KinI kinesin Kif2a is required for bipolar spindleassembly through a functional relationship with MCAK. J Cell Biol 166(4):473–478.

50. Manning AL, et al. (2007) The kinesin-13 proteins Kif2a, Kif2b, and Kif2c/MCAK havedistinct roles during mitosis in human cells. Mol Biol Cell 18(8):2970–2979.

51. Zhu C, et al. (2005) Functional analysis of human microtubule-based motor proteins,the kinesins and dyneins, in mitosis/cytokinesis using RNA interference. Mol Biol Cell16(7):3187–3199.

52. Sanhaji M, Friel CT, Wordeman L, Louwen F, Yuan J (2011) Mitotic centromere-associated kinesin (MCAK): A potential cancer drug target. Oncotarget 2(12):935–947.

53. Aplin A, Jasionowski T, Tuttle DL, Lenk SE, Dunn WA, Jr (1992) Cytoskeletal elementsare required for the formation and maturation of autophagic vacuoles. J Cell Physiol152(3):458–466.

54. Boulton TG, Cobb MH (1991) Identification of multiple extracellular signal-regulatedkinases (ERKs) with antipeptide antibodies. Cell Regul 2(5):357–371.

Zaganjor et al. PNAS | July 22, 2014 | vol. 111 | no. 29 | 10573

CELL

BIOLO

GY

Dow

nloa

ded

by g

uest

on

Dec

embe

r 13

, 202

0