(A) Change in the level of miR-204, -211, and -210 in 786-O and A498 cells with reconstituted wild-type VHL as compared with the level of these miRs in VHL(−) cells. Reconstitution of human VHL (HA-VHL) is shown in a western blot.
(B) Change in miR-204 in Caki-1 with VHL knockdown as compared with cells transduced with scramble control shRNA. The western blot demonstrates knockdown of the endogenous VHL.
(C) Western blot of protein extracts form 6 human kidneys with high (lanes1-6) and low (lanes 7–11) levels of endogenous VHL. The indicated amounts of protein in extracts from 786-O cells with reconstituted VHL were blotted for comparison. Values for the normalized levels of miR-204 in each kidney are shown in a graph under the blot.
(D) Regression analysis of normalized miR-204 levels as a function of normalized VHL levels in kidneys (n=99). Data are log₂-transformed. Each dot represents a sample and the solid line represents the linear regression fit, with the Pearson correlation coefficient (r) as well as the slope and intercept of the fitted line shown in the upper left corner of the box. See also Figure S2.

(A) TEM microphotographs with quantification of the number of autophagic vesicles (AV) (top) and the cytoplasmic area (bottom) in 20 cells under the indicated conditions in 786-O VHL(−) and VHL(+) cells. Controls were treated with the same empty lentiviral constructs (EV). The arrows indicate AV. Scale bars = 1μm.
(B) Representative image of immunofluorescence staining for the endogenous ATG5-ATG12 complex in the indicated conditions in VHL(−) 786-O cells (a). (b) Quantification of the percent of cells with puncta vs. total number of cells in five independent fields containing on average 50 to 150 cells per field in each condition in 786-O and A498 cells. Controls were transfected with scramble construct. Scale bar = 50 μm.
(C) Representative image of immunofluorescence staining for endogenous ATG16L puncta in 786-O VHL(−) and VHL(+) cells under the indicated conditions. Percent of cells expressing puncta under the indicated conditions was calculated as in (B). 100 cells expressing puncta were analyzed for each condition, with the exception of non-treated controls where the number of cells with puncta was small and only 30 to 50 cells were counted. Control cells were treated with scramble construct. Scale bars are as indicated in the figure. All measurements shown in this figure were performed on cells grown in 0.1% serum and 24–30 hr after treatment with miR-204. This was approximately 24 hr before cell death was observed as shown in Figure 3B. See also Figure S4.

(A) Sequence of the wild-type (WT) miR-204 site in the 3’UTR of LC3B, and mutation (mut) of the core binding site (a). (b) The 3’UTR of LC3B with the wild-type but not mutant miR-204 site confers regulation by miR-204 to the luciferase reporter construct.
(B) Effects of an anti–miR-204 lentiviral Zip or siRNAs on LC3B protein expression in 786-O VHL(+) RCC cells shown by western blot and quantification of 3 independent experiments.
(C) Regression analysis showing negative correlation between normalized miR-204 level and normalized LC3B protein level in human ccRCC tumors with the wild-type VHL, performed as described in Figure 2D.
(D) Representative western blot showing tumors with high and low levels of LC3B. Normalized values of miR-204 levels for each tumor are shown in a graph below the blot. Expression of LC3B in 786-O cells is shown as reference.
(E) Western blot showing accumulation of LC3B-II in 786-O VHL(+) cells treated with anti-miR-204 in the absence or presence of 100 μM chloroquine.
(F) 786-O VHL(−) cells transfected with RFP-LC3B or RFP were sorted, plated, and maintained for 48 hr in medium containing 0.1% serum. Cells were then treated with 25 nM short pre-miR-204 or scramble control construct and, 48 hr later, stained by DAPI. Fraction of live cells was determined among the stained and unstained cells in both conditions.
(G) Quantification of endogenous ATG16L puncta per cell in 786-O VHL(−) cells transfected with LC3B-RFP or RFP alone and treated with miR-204 as in (F). Red cells: (+) indicates cells expressing RFP fluorescence from either construct; (−) indicates cells that did not express RFP fluorescence. Numbers of cells (n) in which puncta were counted is shown below the graph.
(H) Effect of LC3B knockdown on cell death induced by miR-204. 786-O or A498 VHL(−) cells stably transfected with two different LC3B shRNAs or scramble construct were additionally transfected with miR-204 or scramble control. Two independent pools of 786-O cells expressing LC3B sh I RNA were used (Pool 1 and Pool 2). N, non-treated cells; L, cells transfected with Lipofectamine only; Scr, cells transfected with scramble pre-miR-204 control; 204, cells transfected with 50 nM of pre-miR-204. See also Figure S6.

(A) Growth of tumors formed by 786-O VHL(−) cells in subcutaneous xenografts injected with packaged lentiviruses containing two different shRNAs against LC3B, or one against ATG5, as compared with the growth of control tumors injected with scramble sequence-containing virus. The number of mice in each experiment is shown in parentheses above the graph.
(B) Growth of tumors formed by pools of 786-O VHL(−) cells stably transfected with two different shRNAs against ATG5 or one against LC3B, and injected under the kidney capsule in nude mice. Western blots show levels of protein knockdown in stable pools of cells under conditions of 10% serum starvation. Cells were plated and grown for 48 hr in medium containing 10% serum before injection.
(C) Growth of tumors formed by pools of A498 VHL(−) cells stably transfected with two different shRNAs against LC3B and one against ATG5 and injected under the kidney capsule in nude mice. Western blots show levels of protein knockdown in stable pools of cells. Cells were treated as in panel B. Photographs of representative tumors are shown with dashed lines drawn to help visualize the tumors.
(D) There is a significant decrease in the level of miR-204 (p<0.001) and a significant increase in the level of LC3B (p<0.05) correlating with tumor grade in human ccRCC tumors. Because of the small number of tumors classified as Grades 1 and 4, the statistical comparison was made between combined subpopulations of G1+G2 vs. G3+G4 tumors. The number of tumors of each grade is provided in parentheses at the bottom. Statistics were calculated as in Figure 1B. Means ± SD are shown.
(E) Formation of tumors in orthotopic xenografts by VHL(+) 786-O and A498 with stable knockdowns of LC3C using two different LC3C shRNAs in each cell line. The table lists numbers of mice analyzed and photographs of representative tumors are shown. Dashed lines help visualize tumorigenic infiltrate of A498 cells. All scale bars = 1 cm.
(F) Schematic representation of the role of VHL in the regulation of autophagy in RCC tumor growth. See also Figure S8.

(A) Localization of miR-204 at the 5′ end of intron 6 of the host gene, TRPM3.
(B) Expression of miR-204, its paralog miR-211, and miR-210 in human ccRCC as compared with normal kidney tissue. Quantification of the indicated normalized miR levels (log₂) of the total unpaired kidney (K; n=114) and tumor (T; n=128) samples are shown. The boxes represent lower and upper quartiles separated by the median (thick horizontal line) and the whiskers extend up to the minimum and maximum values, excluding points that are outside the 1.5 interquartile range from the box (marked as circles). Means ± SD of each distribution are indicated by closed dots and crosses on the whiskers, respectively. All differences were statistically significant at p<0.001.
(C) Growth of pre-formed 786-O VHL(−) tumors in subcutaneous xenografts injected every other day over a period of 9 days with lentiviral particles of wild-type pre-miR-204 (WT) or pre-miR-204 with the 3-base-pair mutation (mut). The data are cumulative over two independent series of injections. Representative samples from each group of tumors are shown below. Scale bar = 1 cm.
(D) Representative images of H&E-stained or C9 antibody-probed sections from tumors injected with wild-type (top) or mutant (bottom) pre-miR-204. A section from a non-injected tumor stained for C9 is shown as a control. Scale bars = 200 μm.
(E) Incidence and size of RCC tumors resulting from transduced cells injected under the kidney capsules of nude mice (a). Example of H&E-stained sections of tumors formed by cells transduced with wild-type pre-miR-204 (b and c) or with mutant pre-miR-204 (d). Scale bar = 200 μm. See also Figure S1.

(A) Time line of the procedures performed during the course of experiments shown in B–E. In all experiments cells were plated at a density of 3,000 per well and the number of live cells treated with the scramble controls was between 10,000 and 20,000 at the time of collection (also see Figure 3C). Treatments were repeated at least three times in triplicate.
(B) Percentages of live cells determined by trypan blue exclusion 48 hr after administration of (a) short pre-miR-204 at the indicated concentrations as compared with the same cells treated with scramble control (Scr) at the same concentrations or (b) lentiviral miR-204 particles administered to the indicated cells as compared with the same cells treated with mutated miR-204 or empty virus (EV) in the presence of 0.1% or 10% serum. In (a) all experiments were performed in the presence of 0.1% serum.
(C) Example of a TEM microphotograph of VHL(-) cells dying from miR-204 cytotoxicity (left) and of live VHL(+) cells (right) under the same conditions. Scale bars = 2μm.
(D) Numbers of live cells when fresh medium containing 10% serum was given 6 hr before administration of miR-204 (see panel A).
(E) Numbers of live cells when medium was supplemented with 10 mM glucose after treatment with miR-204 (see panel A). See also Figure S3.

(A) Western blot analysis of LC3 orthologs in extracts from 786-O and A498 cells grown in 0.1% serum conditions. The experiment follows the same treatment sequence as shown in Figure 3A, with 100 μM chloroquine applied 24 hr after treatment with short pre-miR-204 or scramble control. The cells were collected after 1.5 hr in the case of 786-O cells and after 3 hr in the case of A498 cells, when saturating effects of chloroquine were observed based on the time-course experiments. This was approximately 24 hr before cell death was observed as shown in Figure 3B. Quantification of 3 independent experiments is shown at the bottom.
(B) Percentage of live cells measured using trypan blue analysis in 786-O and A498 VHL(+) and VHL(−) cells transfected with LC3B siRNA, where 100% corresponds to live cells counted after transfection with scramble control construct. Time line as shown in Fig. S4A. See also Figure S5.

(A) Western blot showing time course of LC3C-II induction in VHL(+) and VHL(−) 786-O and A498 cells in response to serum starvation.
(B) Quantitative RT-PCR for LC3C and LC3B mRNAs (48 hr of starvation) in VHL(+) as compared with VHL(−) cells.
(C) Quantitative RT-PCR for LC3C mRNA in VHL(−) cells where HIF-2α was knocked down by using siRNA.
(D) Quantitative RT-PCR for LC3C mRNA in VHL(+) cells where HIF was stimulated with DMOG (1 mM) for 16 hr.
(E) Effect of single LC3C or LC3B knockdown or of double LC3C/LC3B knockdown on the viability of indicated cells, as compared with scramble-treated cells. The timeline followed was the same as that shown in Figure 3A. See also Figure S7.