Pharmacological Inhibition of the Ubiquitin Ligase RNF5 Rescues F508del-CFTR in Cystic Fibrosis Airway Epithelia

Article

Pharmacological Inhibition of the Ubiquitin Ligase

RNF5 Rescues F508del-CFTR in Cystic Fibrosis

Airway Epithelia

Graphical Abstract

Authors

Elvira Sondo, Federico Falchi,

Emanuela Caci, ..., Roberto Ravazzolo,

Andrea Cavalli, Nicoletta Pedemonte

Correspondence

andrea.cavalli@unibo.it (A.C.),

nicoletta.pedemonte@unige.it (N.P.)

In Brief

Sondo et al. used a computational

approach to identify an inhibitor, named

inh-02, for RNF5 ubiquitin ligase. RNF5

detects the misfolding of a mutant CFTR

in cystic ﬁbrosis. Inh-2 decreases

ubiquitylation and rescues F508del-CFTR

on human primary bronchial epithelia.

This work validates RNF5 as a drug target

for cystic ﬁbrosis.

Highlights

d

Homology modeling/virtual ligand screening identiﬁed RNF5

inhibitor inh-2

Inh-2 rescues F508del-CFTR activity on human primary

bronchial epithelia

Inh-2 decreases ubiquitylation and increases half-life of

F508del-CFTR

Inh-2 modulates RNF5 downstream pathways

Sondo et al., 2018, Cell Chemical Biology 25, 1–15

July 19, 2018 ª 2018 Elsevier Ltd.

https://doi.org/10.1016/j.chembiol.2018.04.010

Please cite this article in press as: Sondo et al., Pharmacological Inhibition of the Ubiquitin Ligase RNF5 Rescues F508del-CFTR in Cystic Fibrosis

Airway Epithelia, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.04.010

Cell Chemical Biology

Article

Pharmacological Inhibition of the Ubiquitin

Ligase RNF5 Rescues F508del-CFTR

in Cystic Fibrosis Airway Epithelia

Elvira Sondo,¹^,⁶Federico Falchi,²^,³^,⁶Emanuela Caci, Loretta Ferrera, Elisa Giacomini, Emanuela Pesce,

1

3

1

Valeria Tomati, Sine Mandrup Bertozzi, Luca Goldoni, Andrea Armirotti, Roberto Ravazzolo, Andrea Cavalli,²^,³^,

1

4

1,5

*

1

,7,

and Nicoletta Pedemonte

*

1

U.O.C. Genetica Medica, Istituto Giannina Gaslini, Via Gerolamo Gaslini 5, Genova 16147, Italy

²Dipartimento di Farmacia e Biotecnologie, University of Bologna Alma Mater Studiorum, Bologna 40126, Italy

³D3 Compunet, Fondazione Istituto Italiano di Tecnologia, Genova 16163, Italy

⁴Analytical Chemistry Facility, Fondazione Istituto Italiano di Tecnologia, Genova 16163, Italy

⁵DINOGMI Department, University of Genova, Genova 16132, Italy

⁶These authors contributed equally

⁷Lead Contact

*Correspondence: andrea.cavalli@unibo.it (A.C.), nicoletta.pedemonte@unige.it (N.P.)

https://doi.org/10.1016/j.chembiol.2018.04.010

SUMMARY

of phenylalanine 508 (F508del) is the most common CF mutation

ꢀ90% of all CF alleles). F508del-CFTR shows a misfolding

(

In cystic ﬁbrosis (CF), deletion of phenylalanine 508 defect, which leads to reduced stability and/or defective interac-

(

F508del) in the CFTR channel is associated with mis- tions between separate domains of the CFTR protein. The

mutant channel is retained in the ER and prematurely degraded

folding and premature degradation of the mutant

protein. Among the known proteins associated with

F508del-CFTR processing, the ubiquitin ligase

RNF5/RMA1 is particularly interesting. We previously

demonstrated that genetic suppression of RNF5

in vivo leads to an attenuation of intestinal patholog-

ical phenotypes in CF mice, validating the relevance

of RNF5 as a drug target for CF. Here, we used a

by the ubiquitin/proteasome system. Plasma membrane expres-

sion is thus reduced (trafﬁcking defect) (Du et al., 2005). In addi-

tion, activity of the mutant channel is severely decreased due to a

reduced open-channel probability (gating defect) (Dalemans

et al., 1991).

Druggability of F508del-CFTR has been demonstrated by

several studies. At present, two such molecules are commer-

cially available for CF patients. Ivacaftor, previously known as

computational approach, based on ligand docking VX-770, is a potentiator, i.e., a molecule that increases the activ-

and virtual screening, to discover inh-02, a drug- ity of CFTR mutants with gating defects (Yu et al., 2012). A com-

like small molecule that inhibits RNF5. In in vitro ex- bination of ivacaftor plus lumacaftor, also known as VX-809, is

the second drug that has been approved to treat CF patients ho-

periments, treatment with inh-02 modulated ATG4B

and paxillin, both known RNF5 targets. In immortal-

ized and primary bronchial epithelial cells derived

from CF patients homozygous for the F508del muta-

tion, long-term incubation with inh-02 caused signif-

icant F508del-CFTR rescue. This work validates

RNF5 as a drug target for CF, providing evidence to

mozygous for the F508del mutation. Lumacaftor is a corrector,

i.e., a molecule that improves the maturation and trafﬁcking of

F508del-CFTR to the membrane (Wainwright et al., 2015).

Different mechanisms of action of correctors have been pro-

posed. Correctors could act as pharmacological chaperones

or as proteostasis regulators. Pharmacological chaperones act

on mutant CFTR by stabilizing speciﬁc CFTR domains and/or

support its druggability.

by improving interactions of separate CFTR domains (Okiyoneda

et al., 2013). Proteostasis regulators modulate the proteostasis

environment, leading to beneﬁcial effects on CFTR processing

(Balch et al., 2008).

INTRODUCTION

A single corrector is generally not considered sufﬁcient to pro-

Cystic ﬁbrosis (CF) is the most frequent genetic disease in mote a therapeutically relevant F508del-CFTR rescue. Rather,

Caucasian populations (1 affected in every ꢀ3,000 births). It is CF drug therapy is thought to require more than one pharmaco-

caused by loss-of-function mutations in the gene that encodes logical chaperone or a combination with a proteostasis regulator.

the CF transmembrane conductance regulator (CFTR). CFTR is Novel drug targets and their modulators must therefore be iden-

a cAMP-regulated chloride channel expressed at the apical tiﬁed in order to design new CF therapies. One of the most prom-

membrane of many types of epithelial cells. CF is a multi-organ ising targets is the ubiquitin ligase RNF5/RMA1 (Tomati et al.,

disease, characterized by chronic lung infection and deteriora- 2015; Sondo et al., 2017). RNF5/RMA1 is an E3 ubiquitin ligase

tion of lung function, pancreatic insufﬁciency, male infertility, acting at early stages of CFTR biosynthesis, promoting its degra-

and meconium ileus (Pilewski and Frizzell, 1999). The deletion dation (Younger et al., 2006). Recently, our group demonstrated

Cell Chemical Biology 25, 1–15, July 19, 2018 ª 2018 Elsevier Ltd.

1

Please cite this article in press as: Sondo et al., Pharmacological Inhibition of the Ubiquitin Ligase RNF5 Rescues F508del-CFTR in Cystic Fibrosis

Airway Epithelia, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.04.010

Figure 1. Generation of a Homology Model of RNF5 RING Domain for Ligand Docking Based Screening and Biological Evaluation of Putative

RNF5 Inhibitors as Mutant CFTR Correctors

(

A) Example of a single RNF5 conformation.

B) Superimposition of the 10 RNF5 protein conformations selected for the virtual screening. The protein is represented as cartoon and line, the zinc ions as

spheres. See also Figure S1 and Table S1.

-

(C) Representative traces of the YFP functional assay on CFBE41o cells expressing F508del-CFTR.

D and E) Results of the compound library screening performed on F508del-CFTR-expressing CFBE41o cells, using the YFP functional assay. Each dot rep-

-

(

resents CFTR activity following treatment with a single compound at 10 mM (D) or at 2 mM (E).

(

legend continued on next page)

2

Cell Chemical Biology 25, 1–15, July 19, 2018

Please cite this article in press as: Sondo et al., Pharmacological Inhibition of the Ubiquitin Ligase RNF5 Rescues F508del-CFTR in Cystic Fibrosis

Airway Epithelia, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.04.010

that genetically suppressing the ubiquitin ligase RNF5/RMA1 cell line has been extensively used by our group, in combination

in vivo attenuates the pathological phenotypes due to intestinal with the microﬂuorimetric assay based on the HS-YFP, to identify

malabsorption in F508del/F508del mice, and concomitantly in- and characterize various CFTR rescue maneuvers, including pro-

creases CFTR activity in intestinal epithelial cells (Tomati et al., teostasis regulators (Pedemonte et al., 2011; Sondo et al., 2011;

2015). This previous work validates RNF5 as a drug target for Pesce et al., 2015, 2016; Tomati et al., 2015, 2018). Twenty-four

CF. However, to date, no RNF5 inhibitors have been reported. hours after plating, cells were treated with test compounds (2-

Here, we used a computational approach, based on ligand and 10-mM ﬁnal nominal concentration). The vehicle alone

docking and virtual screening, to discover inh-02, a drug-like (DMSO) and corrector VX-809 (1 mM) were used as negative

small molecule that inhibits RNF5. Long-term incubation with and positive controls, respectively. In parallel, we also transfected

inh-02 caused signiﬁcant F508del-CFTR rescue in immortalized cells with small interfering RNA (siRNA) molecules targeting

and primary bronchial epithelial cells from CF patients homozy- RNF5. After an additional 24 hr, we assessed F508del-CFTR ac-

gous for the F508del mutation. We characterized inh-2’s mech- tivity in the plasma membrane by measuring the rate of HS-YFP

anism of action, conﬁrming that it decreases ubiquitylation of quenching caused by iodide inﬂux into cells (Figure 1C). For

mutant CFTR, causing stabilization of the mature form of each compound, we evaluated the resulting CFTR activity,

CFTR. In addition, inh-2 acts on ATG4B and paxillin, which are normalized to the activity measured in negative controls. Results

known targets downstream of RNF5 (Didier et al., 2003; Kuang of the screenings are displayed in Figures 1D and 1E. CFTR activ-

et al., 2012).

ity scores (from the screening performed at 10 mM) were then put

into an ordered distribution (Figure 1F). Given the high reproduc-

ibility and robustness of the assay, we set an activity threshold of

RESULTS

1.4. Thus, compounds were considered active when CFTR activ-

Homology Modeling and Virtual Ligand Screening to

Identify RNF5 Inhibitors

ity was 40% higher than the control. The primary screening high-

lighted 43 putative hits, i.e., compounds able to elicit signiﬁcant

From a computational standpoint, human RNF5 is a very chal- F508del-CFTR rescue and increased anion transport based on

lenging target. To date, there are no structures available in the assessment of HS-YFP quenching, similar to what was observed

PDB, and no known inhibitors. Moreover, there is very low identity following silencing of RNF5 with siRNA molecules (90%–95% in-

with similar proteins in the PDB for homology modeling en- crease in F508del-CFTR function over negative control cells). We

-

deavors. Therefore, we used two complementary approaches used the YFP functional assay on F508del-CFTR CFBE41o cells

to identify potential RNF5 inhibitors. First, we used homology to test the compounds at different concentrations. Surprisingly,

modeling to build the 3D structure of RNF5. We started from we observed an extremely low hit-conﬁrmation rate, with only

the RNF5 sequence downloaded from the UniProt, and from a two compounds having a clear dose-dependent effect. We

list of potential templates (Table S1) with identity between 28% named these compounds inh-2 and inh-5 (Figure 1G, left panel).

and 42%. After 250 ns of molecular dynamics (MD) simulations, The 50% effective concentration (EC50) values for the two com-

1

0 protein conformations were clustered and eventually retained pounds were equal to 2.6 mM for inh-2 and 3.4 mM for inh-5.

Figure 1). Then, the ten conformations were submitted to drugg- Very narrow dose-response curves were obtained for the other

(

ability tests to identify suitable pockets for virtual ligand hits. These other compounds were active at only one or two con-

screening. All the ligands from the Life Chemicals database (about centrations, and showed marked cytotoxic effects at higher con-

5

00K) were docked into these pockets with GLIDE (SP scoring centrations. To evaluate the possible effects of cell background

function). The 1,000 top-scoring compounds for each protein on the activity of the hits, we tested the compounds at increasing

conformation were stored. The selected compounds (10,000) concentrations on another cell model, Fischer rat thyroid (FRT)

were visually inspected and clusterized. A total of 1,623 suitable cells stably expressing F508del-CFTR and the HS-YFP

compounds were identiﬁed. To obtain more chances of success (Figure 1G, right panel). The results showed that only inh-2 had

of the virtual screening, a second and complementary approach a clear dose-dependent effect, with an EC50 equal to 2.2 mM,

-

was used. We built a diversity set of 1,000 compounds to obtain comparable to that obtained on CFBE41o cells. According to

the most diverse compounds that span the entire descriptor previous results from our group (Tomati et al., 2015) RNF5 sup-

space of the LifeChemicals database, using the Canvans soft- pression displayed an additive effect with VX-809. Thus, we eval-

-

ware on the basis of the MOLPRINT 2D ﬁngerprints. Finally, uated mutant CFTR rescue in CFBE41o cells treated with inh-2

2

,623 molecules were purchased and biologically tested.

and VX-809 (alone and combined together) and we found that

the two rescue maneuvers were additive (Figure 1H).

Biological Evaluation of Test Compounds as F508del-

CFTR Correctors

Compounds were bought in larger quantities (LifeChemical)

and analyzed by NMR to conﬁrm their chemical structures

As a primary screen, we tested the compounds for their ability to (Figure 2A). In parallel, inh-2 was also re-synthesized in house.

-

rescue the F508del-CFTR trafﬁcking defect in CFBE41o bron- During the synthesis, we observed the formation of a small

chial epithelial cells, stably co-expressing F508del-CFTR and amount of an undesired byproduct (roughly 10% compared

the halide-sensitive yellow ﬂuorescent protein (HS-YFP). This with inh-2). This compound was a structural analog of inh-2

(F) Ordered distribution of the CFTR activity scores displayed in (C).

-

(G) Dose-response relationships using the YFP assay on CFBE41o and FRT cells.

-

(H) Additivity of inh-2 and VX-809 in F508del-CFTR CFBE41o cells (YFP assay).

Data are expressed as means ± SEM, n = 4–6. Symbols indicate statistical signiﬁcance versus control: **p < 0.01.

Cell Chemical Biology 25, 1–15, July 19, 2018

3

Please cite this article in press as: Sondo et al., Pharmacological Inhibition of the Ubiquitin Ligase RNF5 Rescues F508del-CFTR in Cystic Fibrosis

Airway Epithelia, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.04.010

Figure 2. Validation of the F508del-CFTR Corrector Activity of the Putative RNF5 Inhibitors Derived from Primary Screening

(

A) Chemical structures of selected hits and analog-1.

-

B) F508del-CFTR activity measured on CFBE41o cells, using the YFP assay after 24 hr treatment with inh-2 and analog-1 alone and in combination (combo).

Activity was normalized for the activity in DMSO-treated cells.

C) Transepithelial electrical resistance measurements performed on F508del-CFTR-expressing FRT cells treated with indicated compounds. The bar graph

shows the delta between the values of electrical resistance measured before and after CFTR inhibition, converted into its reciprocal conductance.

(

-

ꢁ

(

D) Time course of CFTR rescue by inh-2. CFBE41o cells were incubated for different times with inh-2 (5 mM) at 37 C. F508del-CFTR activity was measured using

the YFP assay.

-

E) Bar graph showing the effect of test compounds (inh-2, analog-1, and inh-5: 5 mM; VX-809: 1 mM) on proliferation of CFBE41o cells.

(

-

F and G) Bar graphs showing the number of living (F) and apoptotic (G) CFBE41o cells following 24 hr treatment with indicated compounds.

-

H) CFBE41o cell viability as determined by the MTT cytotoxicity assay. Data are expressed as means ± SEM, n = 4–6. Symbols indicate statistical signiﬁcance

versus control: **p < 0.01, *p < 0.05.

(named analog-1). This compound was separated, puriﬁed, and with the activity of original hit inh-2. Analog-1 had no F508del-

collected for further evaluation (Figure 2A). CFTR corrector activity. In addition, by combining analog-1

F508del-CFTR corrector activity of analog-1 was analyzed in and inh-2, we observed that inh-2’s ability to rescue mutant

-

CFBE41o cells (by measuring YFP quenching) and compared CFTR was severely decreased in the presence of its analog

4

Cell Chemical Biology 25, 1–15, July 19, 2018

Please cite this article in press as: Sondo et al., Pharmacological Inhibition of the Ubiquitin Ligase RNF5 Rescues F508del-CFTR in Cystic Fibrosis

Airway Epithelia, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.04.010

(

Figure 2B). To further evaluate the ability of these compounds to cells from four F508del homozygous subjects, using two

rescue F508del-CFTR, we performed transepithelial electrical different electrophysiological techniques: the transepithelial

resistance measurements, using FRT cells stably expressing electrical resistance and potential difference measurements,

F508del-CFTR. FRT cells were grown on permeable supports and short-circuit current recordings in Ussing chamber (Figure 3).

until they developed into tight epithelia. FRT epithelia were Bronchial epithelia were then treated for 24 hr with DMSO, or

treated for 24 hr with test compounds or vehicle alone (DMSO) with inh-2, analog-1, or inh-5 (each compound at 5 mM) or with

and then assayed. Transepithelial electrical resistance was the combination of inh-2 and analog-1. Epithelia were then as-

measured for each epithelium before and after acute stimulation sayed. Transepithelial electrical resistance and potential differ-

with forskolin (20 mM) plus genistein (50 mM) to fully activate ence were measured for each epithelium before and after addi-

CFTR. It was then measured after addition of the selective tion of amiloride (10 mM) to block ENaC-mediated sodium

CFTR inhibitor PPQ-102 (30 mM; see Tradtrantip et al., 2009) to absorption. They were then measured after full CFTR activation

completely block CFTR activity. The delta between the values with forskolin (20 mM) plus genistein (50 mM). Finally, they were

of electrical resistance before and after CFTR inhibition was measured after CFTR inhibition due to addition of PPQ-102

then converted into its reciprocal conductance and graphed in (30 mM). From the values of transepithelial resistance and poten-

Figure 2C. In agreement with the results obtained using the tial difference measured before and after CFTR inhibition, we

-

YFP assay on FRT and CFBE41o cells, F508del-CFTR rescue calculated the CFTR-dependent transepithelial conductance

by inh-2 signiﬁcantly increased the transepithelial conductance. and equivalent short-circuit current for each condition (Fig-

In contrast, epithelia treated with inh-5, analog-1, or the combi- ure 3A). Long-term treatment of CF primary bronchial epithelia

nation of analog-1 plus inh-2 displayed a conductance similar to with inh-2 signiﬁcantly increased both CFTR-mediated conduc-

that measured in negative controls (Figure 2C).

The time course of F508del-CFTR correction by inh-2 was additivity was observed upon co-treatment with inh-2 and VX-

tance and equivalent short-circuit current. Disappointingly, no

-

analyzed in CFBE41o cells (by measuring YFP quenching) at 809. Corrector activity of inh-2 was completely stopped by co-

different time points. F508del-CFTR correction was seen as incubation with analog-1. Epithelia treated with inh-5 displayed

early as 1 hr after compound addition, with maximal effect after CFTR-dependent activity non-signiﬁcantly different from that of

4

–6 hr (Figure 2D).

DMSO-treated epithelia (Figure 3A). Considering that F508del-

CFTR rescue by inh-2 reached a maximal effect after 4–6 hr in

-

CFBE41o cells (Figure 2D), we tested whether inh-2 and inh-5

Effect of Putative RNF5 Inhibitors on Cell Proliferation,

Viability, and Apoptosis

We then evaluated ﬁrst the effect of chronic treatment with test ment. However, no signiﬁcant rescue was detected under this

could be effective in human bronchial epithelia upon 6 hr treat-

compounds (at the concentration tested for corrector activity) condition (Figure 3A).

-

on the proliferation of CFBE41o cells. Therefore we plated

In parallel, F508del-CFTR homozygous bronchial epithelia

-

YFP-expressing CFBE41o cells and evaluated cell proliferation were incubated for 24 hr with inh-2 or inh-5 (5 mM) and with

for 48 hr following treatment with test compounds or vehicle vehicle alone (DMSO). Epithelia were mounted in Ussing cham-

alone (DMSO). Treatment with inh-2 caused a modest but statis- bers to measure chloride secretion by short-circuit current anal-

+

tically signiﬁcant increase in cell proliferation at 4 hr and 20 hr af- ysis (Figure 3B). After blocking Na current with amiloride, cells

ter plating, without preventing cells from reaching conﬂuency treated with DMSO showed little response to the membrane-

(

Figure 2E). Conversely, cells treated with analog-1 showed a permeable cAMP analog CPT-cAMP, VX-770, or to selective

decreased proliferation at 48 hr after plating (Figure 2E). We CFTRinh-172 (Ma et al., 2002). Incubation with VX-809 for

then evaluated the effect of chronic treatment with high concen- 24 hr resulted in the appearance of signiﬁcant chloride current,

trations of test compounds on the proliferation and apoptosis of as seen by the increased current after CPT-cAMP and VX-770,

-

CFBE41o cells. Cell nuclei were counterstained with Hoechst and inhibition by CFTRinh-172. Incubation with inh-2 for 24 hr

3

3342 and propidium iodide to visualized total and apoptotic increased chloride current, although to a lesser extent than VX-

cell count, respectively (Figures 2F and 2G). Long-term treat- 809 (Figure 3B). Epithelia co-treated with inh-2 and VX-809 dis-

ment with analog-1 (20 and 100 mM) caused a signiﬁcant played CFTR-dependent activity non-signiﬁcantly different from

decrease in cell proliferation, and an increase in the number of that of VX-809 alone (Figure 3B). Epithelia treated with inh-5 dis-

À

apoptotic cells, as compared with control condition (Figures 2F played CFTR-mediated Cl secretion, which was not signiﬁ-

and 2G). Similar effect was observed upon treatment with VX- cantly increased over DMSO-treated epithelia (not shown).

8

09 (100 mM). Finally, we performed the cell viability assay based

on the metabolic conversion of the 3-(4,5-dimethylthiazol-2-yl)- Putative RNF5 Inhibitor inh-2 Decreases Ubiquitylation

,5-diphenyltetrazolium bromide (MTT) into a purple-colored for- of F508del-CFTR thus Increasing its Half-Life in

2

mazan product (Figure 2H). The MTT assay highlighted a cyto- Immortalized Bronchial Epithelial Cells

toxic effect of incubation for 24 hr with analog-1 (20 and It has been demonstrated that RNF5 acts at early stages of CFTR

1

00 mM) as well as of VX-809 (100 mM) (Figure 2H).

biosynthesis, promoting its ubiquitylation and therefore its

degradation through the proteasome (Younger et al., 2006).

-

Putative RNF5 Inhibitor inh-2 Rescues F508del-CFTR

Activity on Human Primary Bronchial Epithelia Derived

from CF Patients

Thus, we evaluated ubiquitylation of mutant CFTR in CFBE41o

cells following 24 hr treatment with DMSO (vehicle), inh-2, inh-

5, and VX-809 (Figure 4A). Subsequently, cells were treated for

The ability of inh-2 to rescue F508del-CFTR was assayed on 4 hr with DMSO alone or with MG-132 (10 mM; to block protea-

well-differentiated primary cultures of human bronchial epithelial somal degradation) and then lysed. Cell lysates were

Cell Chemical Biology 25, 1–15, July 19, 2018

5

Please cite this article in press as: Sondo et al., Pharmacological Inhibition of the Ubiquitin Ligase RNF5 Rescues F508del-CFTR in Cystic Fibrosis

Airway Epithelia, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.04.010

Figure 3. Correction of F508del-CFTR Activity in Human Primary Bronchial Epithelia by inh-2

(A) Transepithelial electrical resistance measurements performed on human primary bronchial epithelia derived from four different CF patients homozygous for

the F508del mutation. Epithelia were treated for 24 hr (or 6 hr) with test compounds: inh-2 (5 mM), inh-5 (5 mM), analog-1 (an-1; 5 mM), VX-809 (1 mM), and VX-809

plus inh-2. The bar graphs show the delta between the values of electrical resistance measured before and after CFTR inhibition, converted into its reciprocal

conductance (upper panels), and the equivalent short-circuit current (lower panels).

(B) Representative traces and bar graphs summarizing data (means ± SEM) from the Ussing chamber recordings of human primary bronchial epithelia from a

homozygous F508del patient (BE91), treated with inh-2 (5 mM), VX-809 (1 mM), and VX-809 plus inh-2 (VX + inh-2). Symbols indicate statistical signiﬁcance versus

control: **p < 0.01, *p < 0.05.

immunoprecipitated using an anti-CFTR antibody and then sub- SDS-PAGE followed by western blotting to evaluate CFTR

jected to SDS-PAGE followed by western blotting to evaluate expression (Figure 4E). As shown in Figures 4E and 4F, the

CFTR expression and ubiquitylation status (Figures 4A–4D). expression of mutant CFTR (both band B and C) decreases

The three test compounds rescued mutant CFTR, as shown by over time. However, when cells were treated with VX-809, the

the appearance of the mature form of CFTR (band C; Figure 4A, half-life of the mature form of mutant CFTR was signiﬁcantly pro-

CFTR blot), and evidenced also by the analysis of intensity pro- longed by 3-fold (from 1.25 to 4 hr; Figure 4F). Treatment with

ﬁles (Figure 4B). Treatment with MG-132 caused the appearance inh-2 alone was less effective than VX-809, causing a modest

of CFTR forms at high molecular weight (at 250–300 kDa) that 50% increase in mature CFTR half-life (from 1.25 to 1.8 hr; Fig-

increased in the presence of test compounds, as evidenced by ure 4F). Interestingly, when cells were co-treated with VX-809

the analysis of intensity proﬁles (Figure 4C). Similarly, MG-132 and inh-2, the half-life was further extended up to 6 hr (Figure 4F).

caused the accumulation of ubiquitylated CFTR, particularly

abundant in the absence of rescue maneuvers (Figure 4A, Ub Evaluation of the Effect of RNF5 Inhibitors on RNF5

blot), and decreased upon treatment with RNF5 inhibitors or Downstream Pathways

VX-809 (Figures 4A and 4D), demonstrating that both the RNF5 Our data were obtained in a bronchial epithelial cell line and vali-

inhibitors and VX-809 decreases ubiquitylation of mutant CFTR dated in human primary bronchial epithelia. They indicated that

putative RNF5 inhibitors may indeed be an important therapeutic

-

in CFBE41o cells.

We then evaluated the degradation rate of mutant CFTR in the tool for rescuing F508del-CFTR. We thus sought to conﬁrm the

absence or in presence of inh-2 and VX-809, alone or combined putative mechanism of action by investigating the effects of

-

together. To this aim, we treated F508del-CFTR CFBE41o cells long-term treatment with inh-2 and analog-1 on RNF5-associ-

with test compounds and, after 24 hr, we blocked protein syn- ated cellular readouts. ATG4B is one of the known targets of

thesis by adding cycloheximide to the medium. We then lysed RNF5. Its ubiquitylation is mediated (at least partially) by RNF5

cells at different time points, and cell lysates were subjected to ligase activity (Kuang et al., 2012). We therefore performed in vivo

6

Cell Chemical Biology 25, 1–15, July 19, 2018

Please cite this article in press as: Sondo et al., Pharmacological Inhibition of the Ubiquitin Ligase RNF5 Rescues F508del-CFTR in Cystic Fibrosis

Airway Epithelia, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.04.010

Figure 4. Reduced Ubiquitylation and Increased Half-Life of Mutant CFTR by RNF5 Inhibitors

-

A) Biochemical analysis of CFTR ubiquitylation and expression pattern in CFTR immunoprecipitates from CFBE41o cells after 24 hr treatment with inh-2 (5 mM),

(

inh-5 (5 mM), and VX-809 (1 mM) in the absence or in the presence of MG-132 (10 mM; last 4 hr) to block proteasomal degradation.

B–D) Analysis of intensity proﬁles of CFTR and ubiquitin in the absence (B) or in the presence of MG-132 (C and D).

(

legend continued on next page)

Cell Chemical Biology 25, 1–15, July 19, 2018

7

Please cite this article in press as: Sondo et al., Pharmacological Inhibition of the Ubiquitin Ligase RNF5 Rescues F508del-CFTR in Cystic Fibrosis

Airway Epithelia, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.04.010

ubiquitylation experiments to evaluate levels of ATG4B ubiquity- fusion of autophagosomes with lysosomes, MDC ﬂuorescence

-

lation in F508del-CFTR-expressing CFBE41o cells following increases due to the acidic environment. Therefore, F508del-

-

4 hr treatment with inh-2, inh-5 (both at 5 mM), or, as negative CFTR-expressing CFBE41o cells were transfected with a scram-

2

control, vehicle alone (DMSO). In parallel, we also silenced cells bled siRNA or an siRNA molecule targeting RNF5, and then

with the anti-RNF5 siRNA molecule. After treatment, cells were treated for 24 hr with torin-1 (20 nM), SAR-405 (2 mM), test com-

lysed, and lysates were immunoprecipitated using an anti- pounds (5 mM), VX-809 (1 mM), or DMSO alone (vehicle). The cells

ATG4B antibody. The immunoprecipitants were then subjected were then loaded with MDC and evaluated by high-content

to SDS-PAGE followed by western blotting to evaluate the pres- confocal imaging to detect signal spots (corresponding to auto-

ence of ubiquitylated ATG4B (Figure 5A). The blotting detected phagic vacuoles) in each cell (Figure 6C). Determination of signal

two ATG4B bands, corresponding to differently ubiquitylated spots clearly demonstrated that the number of autophagic vacu-

forms, at 52 and 57 KDa (presumably corresponding to mono- oles was signiﬁcantly increased following RNF5 silencing, or after

and biubiquitylated ATG4B), as already described by Kuang treatment with torin-1 or inh-2. Combination of RNF5 silencing

et al. (2012). As expected, ATG4B ubiquitylation is markedly plus inh-2 or VX-809 plus inh-2 resulted in an increased number

reduced by RNF5 silencing compared with the control condition of spots, similar to the one observed upon incubation with inh-2

(

Figure 5A). Similarly, incubation with inh-2 decreased the alone. In contrast, incubation with SAR-405 signiﬁcantly

ATG4B ubiquitylation level, while inh-5 was not effective decreased the number of autophagic vacuoles (Figure 6C).

Figure 5A). To further demonstrate the ability of the test com- Restoring the basal autophagy has been proposed as a rescue

(

pounds to inhibit ATG4B ubiquitylation, we performed in vivo strategy for the F508del-CFTR mutant (Tosco et al., 2016). We

ATG4B ubiquitylation experiments in F508del-CFTR-expressing therefore sought to understand whether the F508del-CFTR

-

CFBE41o cells following transient transfection of a plasmid cod- rescue was caused by the increase in basal autophagy, which

ing for a ubiquitin with a triple-hemagglutinin tag (Ub-3xHA). The we observed following genetic suppression or pharmacological

day after transfection, cells were treated with inh-2 and, in paral- inhibition of RNF5. We reasoned that an autophagy inhibitor

lel, with analog-1 (Figure 5B). Ubiquitin overexpression did not would block the rescue of mutant CFTR by RNF5 suppression.

-

alter ATG4B expression level. ATG4B was detected as a marked Thus, F508del-CFTR-expressing CFBE41o cells were trans-

band at 52 KDa (corresponding to mono-ubiquitylated ATG4B) fected with an NT or RNF5-targeting siRNAs. The cells were

and a faint band at 46 KDa (corresponding to the non-ubiquity- then treated for 24 hr with DMSO alone or with SAR-405 (2 mM)

lated protein). However, ubiquitin overexpression signiﬁcantly to inhibit autophagy. In a parallel experiment, F508del-CFTR-ex-

-

increased the polyubiquitylated forms, with bands at 76 and 85 pressing CFBE41o cells were treated for 24 hr with inh-2 (5 mM) or

KDa, as already reported by Kuang et al. (2012).

VX-809 (1 mM) or DMSO alone (vehicle), in the absence or pres-

ATG4B is reported to regulate basal autophagy, while RNF5 ence of SAR-405 (2 mM). We then evaluated mutant CFTR activity

knockdown increases basal autophagy by decreasing ATG4B with the YFP-based assay (Figure 6E). Interestingly, co-treatment

ubiquitylation and thus its degradation (Kuang et al., 2012). We with SAR-405 did not antagonize F508del-CFTR rescue following

exploited known readouts of basal autophagy in F508del- RNF5 silencing or inhibition. This demonstrates that the modula-

-

CFTR-expressing CFBE41o cells transfected with NT- or tion of the autophagic pathway is a secondary effect of RNF5 sup-

RNF5-targeting siRNA (30 nM) or following 24 hr treatment pression, but not the mechanism by which CFTR rescue is

with test compounds (5 mM). Western blot tests were performed achieved (Figure 6E).

on whole cell lysates to quantify expression levels of LC3IIB and

We then focused on paxillin, another known RNF5 target (Did-

p62, two well-known autophagy markers (Klionsky et al., 2016). ier et al., 2003). Paxillin is a multidomain protein that localizes pri-

Treatment with torin-1 (20 nM), a known autophagy inducer, marily to sites of cell adhesions, known as focal adhesions,

markedly increased the LC3IIB band’s intensity (Figure 6A). Incu- which link the extracellular matrix to the actin cytoskeleton.

bation with inh-2 alone, or downregulation of RNF5 in the RNF5 regulates paxillin ubiquitylation and localization, and

absence or in the presence of inh-2, signiﬁcantly increased to therefore affects the actin cytoskeleton (Didier et al., 2003).

the same extent the intensity of the LC3IIB band, while analog- Therefore, we performed a texture analysis of the actin cytoskel-

1

-

did not change band intensity (Figure 6A). In agreement with eton following treatment of CFBE41o F508del-CFTR cells upon

these results, SAR-405 (2 mM), a potent inhibitor of the autopha- different experimental conditions. After 24 hr incubation with test

gic pathway, signiﬁcantly increased expression of p62 (which compounds (5 mM) or vehicle (DMSO), cells were ﬁxed, and the

accumulates when the autophagic process is inhibited), while actin cytoskeleton was labeled using phalloidin conjugated to

torin-1 had no effect (Figure 6B). Incubation with analog-1 signif- Alexa Fluor 647 ﬂuorophore (Figure 7A). As calculated by Hara-

icantly increased the p62 band’s intensity, while treatment with lick’s formula, actin signal contrast (calculated on 4 adjacent

inh-2 or RNF5 downregulation were not effective (Figure 6B).

px) was signiﬁcantly altered by treatment with compounds. In

We then decided to evaluate the effect of inh-2 and its analog particular, inh-2 decreased the signal contrast (4 px), while

on the formation of autophagic vacuoles. Monodansylcadaverine analog-1 increased it (Figure 7B).

(

MDC) is a speciﬁc marker for autolysosomes (Biederbick et al., RNF5 is also reported to affect cell motility by regulating

995). Indeed, MDC accumulates inside autophagosomes. After paxillin. In particular, RNF5 knockdown increases cell motility

1

-

(

E) Immunoblot detection of mutant CFTR in whole lysates derived from CFBE41o cells treated with inh-2 (5 mM), VX-809 (1 mM), or inh-2 plus VX-809, at different

time points following CHX-induced block of protein synthesis.

F) Quantiﬁcation of mutant CFTR (band B and band C) half-life in experiments detailed in (E).

Data are expressed as means ± SEM, n = 3.

(

8

Cell Chemical Biology 25, 1–15, July 19, 2018

Please cite this article in press as: Sondo et al., Pharmacological Inhibition of the Ubiquitin Ligase RNF5 Rescues F508del-CFTR in Cystic Fibrosis

Airway Epithelia, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.04.010

Figure 5. Evaluation of the Activity of Putative Hits as RNF5 Inhibitors: Effect on Downstream Target ATG4B

-

A) Biochemical analysis of the ubiquitylation status of ATG4B in CFBE41o cells treated with the indicated compounds (5 mM), or following RNF5 silencing.

-

B) Biochemical analysis of the ubiquitylation status of ATG4B, following overexpression of ubiquitin (Ub-3xHA) in F508del-expressing CFBE41o cells were

(

transiently transfected with a tagged ubiquitin (Ub-3xHA) and treated with the indicated compounds (5 mM).

Data are expressed as means ± SEM, n = 3. Symbols indicate statistical signiﬁcance versus control: *p < 0.05, **p < 0.01.

Cell Chemical Biology 25, 1–15, July 19, 2018

9

Please cite this article in press as: Sondo et al., Pharmacological Inhibition of the Ubiquitin Ligase RNF5 Rescues F508del-CFTR in Cystic Fibrosis

Airway Epithelia, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.04.010

Figure 6. Evaluation of the Activity of Putative Hits as RNF5 Inhibitors: Effect on ATG4B-Mediated Basal Autophagy

-

A) Biochemical analysis of LC3IIB and p62 in CFBE41o cells reversed transfected with NT- or RNF5-targeting siRNA or treated for 24 hr with DMSO, inh-2 (5 mM;

(

e), analog-1 (5 mM; g), VX-809 (1 mM; h), VX-809 plus inh-2 (i), and inh-5 (5 mM), or treated for 3 hr with torin-1 (20 nM), or SAR-405 (2 mM).

-

B and C) Quantiﬁcation of LC3IIB and p62 expression in F508del-expressing CFBE41o cells treated as indicated in (A) Expression was normalized for the control

(

condition (NT-siRNA or DMSO).

-

D) Autophagic vacuoles in F508del-expressing CFBE41o cells transfected with NT-siRNA (or RNF5 siRNA) and treated with DMSO (a), torin-1 (20 nM; b), SAR-

(

405 (2 mM; c), DMSO + anti-RNF5 siRNA (d), inh-2 (5 mM; e), anti-RNF5 siRNA plus inh-2 (f), analog-1 (5 mM; g), VX-809 (1mM; h), VX-809 plus inh-2 (i), and inh-5

(5 mM; j), visualized after loading with MDC (50 mM). Scale bar, 100 mm.

(

legend continued on next page)

10 Cell Chemical Biology 25, 1–15, July 19, 2018

Please cite this article in press as: Sondo et al., Pharmacological Inhibition of the Ubiquitin Ligase RNF5 Rescues F508del-CFTR in Cystic Fibrosis

Airway Epithelia, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.04.010

Figure 7. Evaluation of the Activity of Puta-

tive Hits as RNF5 Inhibitors: Effect on Down-

stream Target Paxillin

(

A and B) Representative images and bar graph

showing results of texture analysis of actin cyto-

skeleton labeled with phalloidin conjugated to

Alexa Fluor 647, in F508del-CFTR-expressing

-

CFBE41o cells treated with the indicated com-

pounds (5 mM). Scale bar, 50 mm.

(

C and D) Representative images and bar graphs

showing results of scratch wound healing assay

performed on CFPAC-1 cells following 24 hr

treatment with the indicated compounds (5 mM), in

the absence (upper panel) or presence (lower

panel) of mitomycin C (30 mM). Scale bar, 250 mm.

Data are expressed as means ± SEM. Symbols

indicate statistical signiﬁcance versus respective

control condition: **p < 0.01.

to allow or block, respectively, cell pro-

liferation. In both conditions, gap ﬁlling

was signiﬁcantly improved by treatment

with inh-2 (Figures 7C and 7D).

DISCUSSION

A critical goal in CF research is the phar-

macological rescue of processing and

trafﬁcking defects caused by the CFTR

F508del mutation. This outcome may be

achieved using pharmacological chaper-

ones that directly interact with mutant

CFTR to improve its folding and enhance

its stability, or by blocking the activity of

proteins that perturb CFTR trafﬁcking

and promote its premature degradation.

Indeed, several proteins have been identi-

ﬁed as potentially useful drug targets for a

CF therapy based on proteostasis modu-

lation (Balch et al., 2008; Amaral and

Farinha, 2013; Sondo et al., 2017). Previ-

ously, our group compared the therapeu-

by decreasing paxillin ubiquitylation and thus its degradation tic relevance of modulating these proteins using gene-silencing

(

Didier et al., 2003). Therefore, we performed a scratch wound approaches, highlighting RNF5 as an interesting putative drug

healing assay to evaluate whether incubation with inh-2 and its target (Tomati et al., 2015). RNF5 is an E3 ubiquitin ligase that

analog altered cell migration/proliferation (Figures 7C and 7D). acts at early stages of F508del-CFTR biosynthesis (Grove

We used CFPAC-1 cells because they are more ﬁrmly et al., 2009). RNF5 knockdown is particularly effective in

-

attached to the substrate compared with CFBE41o cells. rescuing mutant CFTR at the functional level in immortalized

Moreover, they do not detach as a single cell sheet when bronchial epithelial cells (Tomati et al., 2015). Our observations

injured. CFPAC-1 cells were plated on 96-well plates. After were validated using genetic animal models of RNF5KO (Delau-

2

4 hr, cell layers were scratched, and cells were treated with nay et al., 2008) and CFTRDF508 (van Doorninck et al., 1995; Bij-

test compounds (5 mM) or DMSO (vehicle). Experiments velds et al., 2005). Homozygous F508del-CFTR (CFTRDF508)

were performed in the absence or presence of mitomycin C mice show consistent intestinal malabsorption (Bijvelds et al.,

(

E) Quantiﬁcation of the number of autophagic vesicles (signal spots) in cells treated as indicated in (D), normalized for the control condition (cells transfected with

the NT-siRNA and treated with DMSO).

F) F508del-CFTR activity measured on CFBE41o cells, using the YFP assay upon transfection with NT-siRNA (or RNF5 siRNA) and 24 hr treatment with indicated

-

(

compounds (2 mM SAR-405; 5 mM inh-2; 1 mM VX-809). Activity was normalized for the activity detected under control condition (cells transfected with the NT-

siRNA and treated with DMSO). Data are expressed as means ± SEM. Symbols indicate statistical signiﬁcance versus control condition: *p < 0.05, **p < 0.01.

Cell Chemical Biology 25, 1–15, July 19, 2018 11

Please cite this article in press as: Sondo et al., Pharmacological Inhibition of the Ubiquitin Ligase RNF5 Rescues F508del-CFTR in Cystic Fibrosis

Airway Epithelia, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.04.010

2

005; Freudenberg et al., 2008; Wilke et al., 2011), reduced in-

Evaluation of the mechanism of action of inh-2, leading to

testinal cAMP-induced chloride and bicarbonate transport CFTR rescue, conﬁrmed that this compound decreases ubiqui-

response (Wilke et al., 2011; Xiao et al., 2012), and a higher tylation of mutant CFTR, thus stabilizing the mature form of

secretion of fecal biliary acids (Bijvelds et al., 2005; Freudenberg CFTR (but not the immature one, band B). This is in contrast

et al., 2008; Wilke et al., 2011). In vivo RNF5 suppression amelio- with what has been described, for example, for FAU suppression

rated these defects in CFTRDF508 mice (Tomati et al., 2015), (another possible target involved in F508del-CFTR maturation)

providing a strong basis for developing molecules to inhibit that extends the half-life of both mature and immature CFTR

RNF5 activity.

Here, we used computational methods to generate a ho-

(Tomati et al., 2018).

To further investigate the mechanism of action of inh-2 (and

mology model of RNF5 RING domain (starting from the its analog), we exploited known cellular targets of RNF5, other

structures of the RING domains of other proteins) to perform than CFTR. First, we examined the ubiquitylation level of

virtual screening based on ligand docking. In parallel, we ATG4B, which is mediated by RNF5 (Kuang et al., 2012).

used molecular ﬁngerprinting to select a diversity set of In vivo ubiquitylation experiments conﬁrmed that inh-2 inhibits

compounds.

RNF5 ligase activity, decreasing the ubiquitylated forms of

To increase the chances of successfully identifying RNF5 in- ATG4B. Interestingly, treatment with analog-1 increased the

hibitors, we adopted two complementary computational ap- ubiquitylation status of ATG4B. Treatment with inh-5 did not

proaches. Careful computational modeling was required for affect ATG4B ubiquitylation. We can speculate that inh-5 in-

the very low sequence identity with the templates available for hibits a ubiquitin ligase involved in CFTR ubiquitylation (in

the homology modeling, and for the very high ﬂexibility of protein. immortalized but not in primary bronchial cells), different

In particular, we used ten different protein conformations in the from RNF5.

virtual screening protocol to deal with the high protein ﬂexibility.

ATG4B is reported to be a regulator of basal autophagy

Using these conformations, a ﬁrst set of 1,623 ligands was (Kuang et al., 2012). Treatment of cells with inh-2 increased

selected by high-throughput docking. A second diversity set of the basal level of autophagy, as demonstrated by functional

1,000 ligands based on molecular ﬁngerprint chemical diversity (increased MDC signal spots) and biochemical evidence

was extracted. (increased intensity of LC3IIB band and decreased intensity of

We used the functional assay based on HS-YPF to evaluate p62 band). These results are in agreement with the ﬁnding

the biological activity of selected compounds as F508del- that cysteine protease ATG4B is responsible for regulating

CFTR correctors. Our group has used this assay extensively LC3 processing (Kuang et al., 2012). Treatment with analog-1

to identify and characterize CFTR correctors (Pedemonte caused the opposite effects (compared to inh-2). Restoring

et al., 2005, 2010, 2011). In total, 43 hits were identiﬁed in the the basal autophagy has been proposed as a rescue strategy

-

primary screening on CFBE41o cells. The hit rate of 1.5% for the F508del-CFTR mutant (Tosco et al., 2016). Here, how-

was very high compared with that obtained in previous screen- ever, the increase in basal autophagy appeared to be a

ings (Pedemonte et al., 2005, 2010), suggesting that the secondary effect of RNF5 inhibition, rather than the cause of

compound set was enriched by molecules acting as F508del- mutant CFTR rescue.

correctors. Surprisingly, we observed an extremely low hit-

We also evaluated the effect of inh-2 and analog-1 on another

conﬁrmation rate, with only two compounds having a clear RNF5 target, paxillin (Didier et al., 2003). Paxillin localizes in focal

dose-dependent effect. The remaining 41 compounds, how- adhesions, which link extracellular matrix to the actin cytoskel-

ever, should not be considered false positives. Their corrector eton (Turner, 2000). By increasing paxillin ubiquitylation, RNF5

activity was consistently detected but only at one or two of overexpression alters the actin cytoskeleton and decreases

the concentrations tested, with toxic effects at higher concen- cell motility (Didier et al., 2003). Thus, we analyzed the texture

trations. These disappointing results could be explained by pattern of the actin cytoskeleton, which was consistently altered

their putative mechanism of action, involving inhibition of ubiq- by treatment with both inh-2 and analog-1, although with oppo-

uitin ligases. Indeed, while RNF5 suppression has no apparent site effects. Scratch/wound healing assays conﬁrmed that inh-2

negative effect in vitro or in vivo, as shown in this and in previous increases cell motility, improving gap ﬁlling in the cell layer,

reports (Delaunay et al., 2008; Tomati et al., 2015), primary hits which is consistent with inhibition of RNF5 ubiquitin ligase activ-

could also inhibit activity of ligases other than RNF5, leading to ity. In contrast, treatment with analog-1 signiﬁcantly delayed

toxic side effects.

We used two different electrophysiological techniques to

gap ﬁlling.

In summary, our ﬁndings demonstrate that inh-2 acts as an

evaluate the best hits as F508del-CFTR correctors on human RNF5 inhibitor, modulating RNF5 downstream targets. We

primary bronchial epithelia. The compound named inh-2 cannot exclude that inh-2 could also affect other targets; how-

rescued the F508del-CFTR processing defect, increasing ever, the biological effects that we observed after treatment

CFTR-mediated current. Surprisingly, a close analog of inh-2 with inh-2 are consistent with what has been described as

(

analog-1), as well as inh-5, had no activity as correctors. Co- RNF5-dependent. Thus, inh-2 could inhibit for example other

treatment with inh-2 and analog-1 resulted in a complete loss ubiquitin ligases involved in the same pathways of RNF5. We

of F508del-CFTR rescue by inh-2. In the case of inh-5, we hypothesize that inh-2 directly binds to RNF5 RING domain

can speculate that lack of consistency between results ob- thus inhibiting RNF5 ligase activity. The proposed binding

tained in immortalized and primary bronchial cells could be mode (Figure S1) based on docking simulations of inh-2 into

due to possible effects on other targets that, in primary cells, the homology model of RNF5 shows an H-bond between our

hinder mutant CFTR rescue.

compound and Arg73, two p-p interactions with Trp48 and

12

Cell Chemical Biology 25, 1–15, July 19, 2018

Please cite this article in press as: Sondo et al., Pharmacological Inhibition of the Ubiquitin Ligase RNF5 Rescues F508del-CFTR in Cystic Fibrosis

Airway Epithelia, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.04.010

His52, and some hydrophobic interactions with Leu55, Val38, STAR+METHODS

and Val76. Despite this binding mode should be considered

rather preliminary, it can offer the possibility to rationally modify Detailed methods are provided in the online version of this paper

and optimize inh-2 through a medicinal chemistry campaign, and include the following:

which can in turn help in better understanding the binding

d KEY RESOURCES TABLE

mode and optimizing the homology model.

Experimental evidence suggests also that analog-1 could be

d CONTACT FOR REAGENTS AND RESOURCE SHARING

d EXPERIMENTAL MODEL AND SUBJECT DETAILS

an RNF5 activator, eliciting the opposite effects on RNF5 tar-

gets downstream as compared with inh-2. Indeed, the ability

d METHOD DETAILS

B Homology Modeling

of inh-2 to rescue the F508del-CFTR trafﬁcking defect is

completely reversed by co-treatment with analog-1. Notably,

B Molecular Dynamics

B Life Chemicals Database Preparation

although pharmacological modulation of RNF5 has a signiﬁcant

impact on cell motility, basal autophagy, and possibly other cell

B Virtual Screening by High-Throughput Docking

B Diversity Set Selection

processes, mice lacking RNF5 expression display no altered

phenotypes (Delaunay et al., 2008). This suggests that in vivo

B Compound Library Screening

B Synthesis of Compounds

RNF5 inhibition could be exploited. In addition, RNF5 knockout

mice have been shown to be resistant to group A streptococcus

B Synthetic Procedures to Obtain Compounds inh-2,

analog-1, and Intermediates 1-2

infection, in part due to enhanced autophagy-mediated clear-

ance of invading bacteria by macrophages (Kuang et al.,

B Puriﬁcation of inh-2 and analog-1

B NMR Purity Assessment and Structural Elucidation of

2

012). Indeed, in vivo RNF5 regulation of autophagy is involved

inh-2 and analog-1

B Final Purity and Chemical Structures Data for inh-2 and

analog-1

in the innate host defense against bacterial infection. This could

be considered a positive secondary effect of RNF5 inhibition in

CF patients having chronic lung infection. However, it has also

been demonstrated that RNF5 negatively regulates virus-trig-

gered induction of type I IFNs and cellular antiviral response

by targeting both VISA and MITA (Zhong et al., 2010). Thus,

in vivo RNF5 inhibition could lead to deregulated cellular anti-

viral response.

B Gene Silencing by siRNA Transfection

B Fluorescence Assay for CFTR

B Transepithelial Electrical Resistance and Potential Dif-

ference Measurements (TEER/PD)

B Short-Circuit Current Recordings

B CFTR Immunoprecipitation (IP) Assay

B Western Blot

Taken together, our ﬁndings demonstrate that RNF5 inhibi-

tion can rescue the F508del-CFTR trafﬁcking defect, not

only in heterologous cell lines or in a murine CF model, but

also in primary culture of human bronchial epithelia, the

main target tissue for CF treatment. These ﬁndings validate

RNF5 as a drug target for CF, and provide evidence to support

its druggability.

B Proliferation Study

B MTT Tetrazolium Cell Viability Assay

B In Vivo Ubiquitylation Assay

B Labeling of Autophagic Vacuoles with Monodansylca-

daverine (MDC)

B Analysis of Actin Cytoskeleton

B Scratch Wound Healing Assay

d QUANTIFICATION AND STATISTICAL ANALYSIS

SIGNIFICANCE

In cystic ﬁbrosis (CF), deletion of phenylalanine 508

(F508del) in the CFTR channel is associated with misfolding

SUPPLEMENTAL INFORMATION

and premature degradation of the mutant protein. Correc-

tors are molecules that improve the maturation and traf-

ﬁcking of F508del-CFTR to the membrane. Correctors could

act by improving CFTR folding and processing or as proteo-

stasis regulators, i.e., modulating the proteostasis environ-

ment thus leading to beneﬁcial effects on CFTR processing.

Among the known proteins associated with F508del-CFTR

processing, the ubiquitin ligase RNF5/RMA1 is particularly

interesting. Genetic suppression of RNF5 in vivo leads to

an attenuation of intestinal pathological phenotypes in CF

mice, validating the therapeutic relevance of this target for

CF pathology. Here, we used a computational approach to

discover inh-02, a drug-like small molecule that inhibits

RNF5. Inh-2 rescues F508del-CFTR processing defect in pri-

mary bronchial epithelia from CF patients. Analysis of inh-2

mechanism of action conﬁrmed that it decreases CFTR

ubiquitylation, thus increasing its half-life. This work vali-

dates RNF5 as a drug target for CF, providing evidence to

support its druggability.

Supplemental Information includes one ﬁgure and one table and can be

found with this article online at https://doi.org/10.1016/j.chembiol.2018.

0

4.010.

ACKNOWLEDGMENTS

The authors thank Dr. Aldamaria Puliti, Dr. Renata Bocciardi, and Dr. Luis J. V.

Galietta for critical readings of the manuscript and helpful suggestions. The au-

thors thank Mr. Giuseppe Nicora and Mr. Carlo Pedemonte for their valuable

assistance with the Ussing chamber equipment. This work was supported

by Ricerca Corrente from the Italian Ministry of Health (Linea1), by ERA-Net

for Research Programmes on Rare Diseases (E-RARE 3; ‘‘INSTINCT’’), by

grants from the Fondazione per la Ricerca sulla Fibrosi Cistica to A.C. (FFC

no. 2/2015 with the contribution of ‘‘Delegazioni FFC di Verona e di Tradate

Gallarate,’’ ‘‘La Notte dei Sapori’’; and ‘‘Audemars Piguet Italia’’) and to N.P.

(

FFC no. 5/2012 with the contribution of ‘‘Danone S.p.A.’’; FFC no. 9/2017

with the contribution of ‘‘Delegazione FFC di Genova e Gruppo di Sostegno

FFC di Savona Spotorno,’’ "Un ﬁore per Valeria [Assemini – Cagliari],’’ ‘‘Gruppo

di Sostegno FFC di Vigevano,’’ ‘‘Delegazione FFC della Valdadige,’’ and ‘‘Del-

egazione FFC di Lodi’’).

Cell Chemical Biology 25, 1–15, July 19, 2018 13

Please cite this article in press as: Sondo et al., Pharmacological Inhibition of the Ubiquitin Ligase RNF5 Rescues F508del-CFTR in Cystic Fibrosis

Airway Epithelia, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.04.010

AUTHOR CONTRIBUTIONS

Kuang, E., Okumura, C.Y., Sheffy-Levin, S., Varsano, T., Shu, V.C., Qi, J.,

Niesman, I.R., Yang, H.J., L o´ pez-Ot ı´ n, C., Yang, W.Y., et al. (2012).

Regulation of ATG4B stability by RNF5 limits basal levels of autophagy and in-

ﬂuences susceptibility to bacterial infection. PLoS Genet. 8 , e1003007.

E.S. performed the biological evaluation of selected compounds, with the help

of V.T. F.F. generated the homology model of RNF5 RING domain and per-

formed the virtual screening. E.G. synthesized the test compounds, which

were puriﬁed and quality-checked by S.M.B., L.G., and A.A. E.P. performed

the cadaverine assay. Primary bronchial epithelial cells were cultured by

E.C., who also performed the short-circuit current measurements. L.F. per-

formed transepithelial electrical resistance measurements on cell lines and pri-

mary bronchial epithelia. A.C. supervised the computational studies. N.P.

analyzed biological data. R.R., N.P., and A.C. planned the study. N.P. wrote

the manuscript, aided by F.F., E.S., R.R., and A.C.

Ma, T., Thiagarajah, J.R., Yang, H., Sonawane, N.D., Folli, C., Galietta, L.J.,

and Verkman, A.S. (2002). Thiazolidinone CFTR inhibitor identiﬁed by high-

throughput screening blocks cholera toxin-induced intestinal ﬂuid secretion.

J. Clin. Invest. 110 , 1651–1658.

Okiyoneda, T., Veit, G., Dekkers, J.F., Bagdany, M., Soya, N., Xu, H., Roldan,

A., Verkman, A.S., Kurth, M., Simon, A., et al. (2013). Mechanism-based

corrector combination restores D F508-CFTR folding and function. Nat.

Chem. Biol. 9 , 444–454.

Pedemonte, N., Tomati, V., Sondo, E., Caci, E., Millo, E., Armirotti, A.,

Damonte, G., Zegarra-Moran, O., and Galietta, L.J. (2011). Dual activity of ami-

noarylthiazoles on the trafﬁcking and gating defects of the cystic ﬁbrosis trans-

membrane conductance regulator chloride channel caused by cystic ﬁbrosis

mutations. J. Biol. Chem. 286 , 15215–15226.

DECLARATION OF INTERESTS

The authors have declared that no conﬂict of interest exists.

Received: November 14, 2017

Revised: February 22, 2018

Accepted: April 9, 2018

Pedemonte, N., Lukacs, G.L., Du, K., Caci, E., Zegarra-Moran, O., Galietta,

L.J., and Verkman, A.S. (2005). Small-molecule correctors of defective

DeltaF508-CFTR cellular processing identiﬁed by high-throughput screening.

J. Clin. Invest. 115 , 2564–2571.

Published: May 10, 2018

REFERENCES

Pedemonte, N., Tomati, V., Sondo, E., and Galietta, L.J.V. (2010). Inﬂuence of

cell background on pharmacological rescue of mutant CFTR. Am. J. Physiol.

Cell Physiol. 298 , C866–C874.

Amaral, M.D., and Farinha, C.M. (2013). Rescuing mutant CFTR: a multi-task

approach to a better outcome in treating cystic ﬁbrosis. Curr. Pharm. Des.

Pesce, E., Bellotti, M., Liessi, N., Guariento, S., Damonte, G., Cichero, E.,

Galatini, A., Salis, A., Gianotti, A., Pedemonte, N., et al. (2015). Synthesis

and structure-activity relationship of aminoarylthiazole derivatives as correc-

tors of the chloride transport defect in cystic ﬁbrosis. Eur. J. Med. Chem.

99 , 14–35 .

19 , 3497–3508.

Balch, W.E., Morimoto, R.I., Dillin, A., and Kelly, J.W. (2008). Adapting proteo-

stasis for disease intervention. Science 319 , 916–919.

Biederbick, A., Kern, H.F., and Els a € sser, H.P. (1995). Monodansylcadaverine

(

MDC) is a speciﬁc in vivo marker for autophagic vacuoles. Eur. J. Cell Biol.

Pesce, E., Gorrieri, G., Sirci, F., Napolitano, F., Carrella, D., Caci, E., Tomati, V.,

Zegarra-Moran, O., di Bernardo, D., and Galietta, L.J. (2016). Evaluation of a

systems biology approach to identify pharmacological correctors of the

mutant CFTR chloride channel. J. Cyst. Fibros. 15 , 425–435 .

6

6 , 3–14.

Bijvelds, M.J.C., Bronsveld, I., Havinga, R., Sinaasappel, M., de Jonge, H.R.,

and Verkade, H.J. (2005). Fat absorption in cystic ﬁbrosis mice is impeded

by defective lipolysis and post-lipolytic events. Am. J. Physiol. Gastrointest.

Liver Physiol. 288 , 646–653.

Pilewski, J.M., and Frizzell, R.A. (1999). Role of CFTR in airway disease.

Physiol. Rev. 79 , S215–S255 .

Dalemans, W., Barbry, P., Champigny, G., Jallat, S., Dott, K., Dreyer, D.,

Crystal, R.G., Pavirani, A., Lecocq, J.P., and Lazdunski, M. (1991). Altered

chloride ion channel kinetics associated with the delta F508 cystic ﬁbrosis

mutation. Nature 354 , 526–528.

Reyn e` s, C., Host, H., Camproux, A.C., Laconde, G., Leroux, F., Mazars, A.,

Deprez, B., Fahraeus, R., Villoutreix, B.O., and Sperandio, O. (2010).

Designing focused chemical libraries enriched in protein-protein interaction in-

hibitors using machine-learning methods. PLoS Comput. Biol. 6 , e1000695.

Delaunay, A., Bromberg, K.D., Hayashi, Y., Mirabella, M., Burch, D., Kirkwood,

B., Serra, C., Malicdan, M.C., Mizisin, A.P., Morosetti, R., et al. (2008). The ER-

bound RING ﬁnger protein 5 (RNF5/RMA1) causes degenerative myopathy in

transgenic mice and is deregulated in inclusion body myositis. PLoS One

Scudieri, P., Caci, E., Bruno, S., Ferrera, L., Schiavon, M., Sondo, E., Tomati,

V., Gianotti, A., Zegarra-Moran, O., Pedemonte, N., et al. (2012). Association of

TMEM16A chloride channel overexpression with airway goblet cell meta-

plasia. J. Physiol. 590 , 6141–6155 .

3

, e1609.

Sondo, E., Tomati, V., Caci, E., Esposito, A.I., Pfeffer, U., Pedemonte, N., and

Galietta, L.J. (2011). Rescue of the mutant CFTR chloride channel by pharma-

cological correctors and low temperature analyzed by gene expression

proﬁling. Am. J. Physiol. Cell Physiol. 301 , C872–C885.

Didier, C., Broday, L., Bhoumik, A., Israeli, S., Takahashi, S., Nakayama, K.,

Thomas, S.M., Turner, C.E., Henderson, S., Sabe, H., and Ronai, Z. (2003).

RNF5, a RING ﬁnger protein that regulates cell motility by targeting paxillin

ubiquitination and altered localization. Mol. Cell Biol. 23 , 5331–5345.

Sondo, E., Pesce, E., Tomati, V., Marini, M., and Pedemonte, N. (2017). RNF5,

DAB2 and friends: novel drug targets for cystic ﬁbrosis. Curr. Pharm. Des. 23 ,

176–186 .

Du, K., Sharma, M., and Lukacs, G.L. (2005). The DeltaF508 cystic ﬁbrosis mu-

tation impairs domain-domain interactions and arrests post-translational

folding of CFTR. Nat. Struct. Mol. Biol. 12 , 17–25.

The UniProt Consortium. (2017). UniProt: the universal protein knowledge-

Freudenberg, F., Broderick, A.L., Yu, B.B., Leonard, M.R., Glickman, J.N., and

Carey, M.C. (2008). Pathophysiological basis of liver disease in cystic ﬁbrosis

employing a DeltaF508 mouse model. Am. J. Physiol. Gastrointest. Liver

Physiol. 294 , 1411–1420.

base. Nucleic Acids Res. 46 , 2699 .

Tomati, V., Sondo, E., Armirotti, A., Caci, E., Pesce, E., Marini, M., Gianotti, A.,

Jeon, Y.J., Cilli, M., Pistorio, A., et al. (2015). Genetic inhibition of the ubiquitin

ligase RNF5 attenuates phenotypes associated to F508del cystic ﬁbrosis mu-

tation. Sci. Rep. 5 , 12138 .

Grove, D.E., Rosser, M.F.N., Ren, H.Y., Naren, A.P., and Cyr, D.M. (2009).

Mechanisms for rescue of correctable folding defects in CFTRDelta F508.

Mol. Biol. Cell 20 , 4059–4069.

Tomati, V., Pesce, E., Caci, E., Sondo, E., Scudieri, P., Marini, M., Amato, F.,

Castaldo, G., Ravazzolo, R., Galietta, L.J.V., and Pedemonte, N. (2018).

High-throughput screening identiﬁes FAU protein as a regulator of mutant

cystic ﬁbrosis transmembrane conductance regulator channel. J. Biol.

Chem. 293 , 1203–1217.

Haralick, R.M., Shanmugam, K., and Dinstein, I. (1973). Textural features for

image classiﬁcation. IEEE Trans. Syst. Man Cybern. 3 , 610–621.

Klionsky, D.J., Abdelmohsen, K., Abe, A., Abedin, M.J., Abeliovich, H.,

Acevedo Arozena, A., Adachi, H., Adams, C.M., Adams, P.D., Adeli, K., et al.

Tosco, A., De Gregorio, F., Esposito, S., De Stefano, D., Sana, I., Ferrari, E.,

Sepe, A., Salvadori, L., Buonpensiero, P., Di Pasqua, A., et al. (2016). A novel

treatment of cystic ﬁbrosis acting on-target: cysteamine plus epigallocatechin

(2016). Guidelines for the use and interpretation of assays for monitoring auto-

phagy (3rd edition). Autophagy 12 , 1–222.

14 Cell Chemical Biology 25, 1–15, July 19, 2018

Please cite this article in press as: Sondo et al., Pharmacological Inhibition of the Ubiquitin Ligase RNF5 Rescues F508del-CFTR in Cystic Fibrosis

Airway Epithelia, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.04.010

gallate for the autophagy-dependent rescue of class II-mutated CFTR. Cell

Wilke, M., Buijs-Offerman, R.M., Aarbiou, J., Colledge, W.H., Sheppard, D.N.,

Touqui, L., Bot, A., Jorna, H., de Jonge, H.R., and Scholte, B.J. (2011). Mouse

models of cystic ﬁbrosis: phenotypic analysis and research applications.

J. Cyst. Fibros. 10 , S152–S171.

Death Differ. 8 , 1380–1393.

Tradtrantip, L., Sonawane, N.D., Namkung, W., and Verkman, A.S. (2009).

Nanomolar potency pyrimido-pyrrolo-quinoxalinedione CFTR inhibitor re-

duces cyst size in a polycystic kidney disease model. J. Med. Chem. 52 ,

Wu, P.S., and Otting, G. (2005). Rapid pulse length determination in high-res-

6

447–6455.

Turner, M.R. (1986). Texture discrimination by Gabor functions. Biol. Cybern.

5 , 71–82 .

olution NMR. J. Magn. Reson. 176 , 115–119.

Xiao, F., Li, J., Singh, A.K., Riederer, B., Wang, J., Sultan, A., Park, H., Lee,

M.G., Lamprecht, G., Scholte, B.J., et al. (2012). Rescue of epithelial HCO3-

secretion in murine intestine by apical membrane expression of the cystic

ﬁbrosis transmembrane conductance regulator mutant F508del. J. Physiol.

5

Turner, C.E. (2000). Paxillin and focal adhesion signalling. Nat. Cell Biol. 12 ,

E231–E236 .

5

90 , 5317–5334.

van Doorninck, J.H., French, P.J., Verbeek, E., Peters, R.H., Morreau, H.,

Bijman, J., and Scholte, B.J. (1995). A mouse model for the cystic ﬁbrosis delta

F508 mutation. EMBO J. 14 , 4403–4411.

Younger, J.M., Chen, L., Ren, H.Y., Rosser, M.F., Turnbull, E.L., Fan, C.Y.,

Patterson, C., and Cyr, D.M. (2006). Sequential quality-control checkpoints

triage misfolded cystic ﬁbrosis transmembrane conductance regulator. Cell.

126 , 571–582.

Wainwright, C.E., Elborn, J.S., Ramsey, B.W., Marigowda, G., Huang, X.,

Cipolli, M., Colombo, C., Davies, J.C., De Boeck, K., Flume, P.A., et al.;

TRAFFIC and TRANSPORT Study Groups (2015). Lumacaftor-ivacaftor in pa-

tients with cystic ﬁbrosis homozygous for phe508del CFTR. N. Engl. J. Med.

Yu, H., Burton, B., Huang, C.J., Worley, J., Cao, D., Johnson, J.P., Jr., Urrutia,

A., Joubran, J., Seepersaud, S., Sussky, K., et al. (2012). Ivacaftor potentiation

of multiple CFTR channels with gating mutations. J. Cyst. Fibros. 11 , 237–245.

3

73 , 220–231.

Zhong, B., Zhang, Y., Tan, B., Liu, T.T., Wang, Y.Y., and Shu, H.B. (2010). The

E3 ubiquitin ligase RNF5 targets virus-induced signaling adaptor for ubiquiti-

nation and degradation. J. Immunol. 184 , 6249–6255.

Wider, G., and Dreier, L. (2006). Measuring protein concentrations by NMR

spectroscopy. J. Am. Chem. Soc. 128 , 2571–2576 .

Cell Chemical Biology 25, 1–15, July 19, 2018 15

Please cite this article in press as: Sondo et al., Pharmacological Inhibition of the Ubiquitin Ligase RNF5 Rescues F508del-CFTR in Cystic Fibrosis

Airway Epithelia, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.04.010

STAR+METHODS

KEY RESOURCES TABLE

REAGENT or RESOURCE

SOURCE

IDENTIFIER

Antibodies

Rabbit monoclonal anti-ATG4B (D1G2R)

Rabbit polyclonal anti-LC3B

Cell Signaling Technology

Sigma

Cat#13507

Cat#L7543; RRID: AB_796155

Cat#5114

Rabbit polyclonal anti-SQTM1/p62

Mouse monoclonal anti-RNF5 antibody (22B3)

Mouse monoclonal anti-GAPDH (6C5)

Mouse monoclonal anti-Ub (P4D1)

Cell Signaling Technology

Santa Cruz Biotechnology

Cat#sc-81716; RRID: AB_2238618

Cat#sc-32233; RRID: AB_627679

Cat#sc-8017; RRID: AB_628423

Cat#sc-8017-HRP

Horseradish peroxidase (HRP)-conjugated mouse

monoclonal anti-Ub (P4D1-HRP)

Mouse monoclonal anti-CFTR (596)

Cystic Fibrosis Foundation

Therapeutics

ID: 596

Mouse monoclonal anti-CFTR (24-1)

R&D Systems

Millipore

Cat#MAB25031; RRID: AB_2260673

Cat#05-369; RRID: AB_309699

Mouse monoclonal anti Na+/K+ ATPase

a1 (C464.6)

Monoclonal antibody anti- HA.11 (16B12)

HRP-conjugated anti-mouse IgG

HRP-conjugated anti-rabbit IgG

HRP-conjugated anti-HA (16B12)

Chemicals, Peptides, and Recombinant Proteins

VX-809

Biolegend

Abcam

Cat#901514; RRID: AB_2565336

Cat#ab6789; RRID: AB_955439

Cat#P0448; RRID: AB_2617138

Cat#901519; RRID: AB_2686981

DAKO

Biolegend

Selleck Chemicals

Calbiochem

Cat#S1565

Cat#S1144

Cat#S7139

Cat#219677

Cat#F6886

Cat#G6649

VX-770

CFTR inhibitor 172

PPQ102

Forkolin

Sigma

Genistein

Sigma

Experimental Models: Cell Lines

Rat: Fischer rat thyroid (FRT) cells

-

Human: CFBE41o (unknown age and gender)

Gaslini cell bank

N/A

Laboratory of John P Clancy

-

Human: CFBE41o wt-CFTR (unknown age and

gender)

-

Human: CFBE41o F508del-CFTR (unknown age

Laboratory of John P Clancy

Gaslini cell bank

ATCC

N/A

and gender)

Human: Primary CF-HBE F508del/F508del

Donor ID: BE86

Donor ID: BE91

Donor ID: BE93

Donor ID: BE111

Cat#CRL-1918

(

unknown age and gender)

Human: Primary CF-HBE F508del/F508del

unknown age and gender)

Human: Primary CF-HBE F508del/F508del

unknown age and gender)

Human: Primary CF-HBE F508del/F508del

unknown age and gender)

(

Human: CFPAC-1 (26 yrs; male)

Oligonucleotides

Anti-RNF5 siRNA

Life Technologies

Cat#HSS184327

ID: MED GC

siRNA negative control

Recombinant DNA

Plasmid: Ub-3xHA

VectorBuilder

custom

(Continued on next page)

e1 Cell Chemical Biology 25, 1–15.e1–e8, July 19, 2018

Please cite this article in press as: Sondo et al., Pharmacological Inhibition of the Ubiquitin Ligase RNF5 Rescues F508del-CFTR in Cystic Fibrosis

Airway Epithelia, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.04.010

Continued

REAGENT or RESOURCE

Software and Algorithms

Protein Preparation Wizard

SOURCE

IDENTIFIER

Schr o¨ dinger

https://www.schrodinger.com/protein-

preparation-wizard

Desmond

Schr o¨ dinger

https://www.schrodinger.com/desmond

https://www.schrodinger.com/LigPrep

LigPrep

Schr o¨ dinger

PPI-HitProﬁler tool

Reyn e` s et al., 2010

http://www.cdithem.fr/online.php?

lg=en&service=ppiHitProﬁler

SiteMap

Glide

Schr o¨ dinger

https://www.schrodinger.com/sitemap

https://www.schrodinger.com/glide

https://www.schrodinger.com/canvas

Canvas

CONTACT FOR REAGENTS AND RESOURCE SHARING

Further information and requests for reagents may be directed to, and will be fulﬁlled by the lead contact and corresponding author

for biological studies Nicoletta Pedemonte (nicoletta.pedemonte@unige.it ) or by the corresponding author for computational studies

Andrea Cavalli (andrea.cavalli@unibo.it ).

EXPERIMENTAL MODEL AND SUBJECT DETAILS

-

We generated CFBE41o and FRT cells stably expressing F508del-CFTR and the halide-sensitive yellow ﬂuorescent protein

-

HS-YFP) YFP-H148Q/I152L, as previously described (Sondo et al., 2011). As culture media, we used MEM for CFBE41o cells

(

and F-12 Coon’s modiﬁcation for FRT cells, both supplemented with 10% fetal calf serum, 2 mM L-glutamine, 100 U/ml penicillin,

-

and 100 mg/ml streptomycin. For YFP-based assays of CFTR activity, CFBE41o or FRT cells were plated (50,000 cells/well) on

clear-bottom 96-well black microplates (Corning Life Sciences, Acton, MA).

The isolation, culture, and differentiation methods of primary bronchial epithelial cells have previously been described in detail

(

No information about gender and age of the transplanted patients was provided to cell ﬁnal users. For this study, cells were obtained

Scudieri et al., 2012). In brief, epithelial cells were obtained from mainstem bronchi of CF individuals undergoing lung transplant.

ꢁ

from four CF patients (homozygous for F508del mutation). Bronchi were set overnight at 4 C in a solution containing protease XIV to

detach cells. Epithelial cells were then cultured in a serum-free medium (LHC9 mixed with RPMI 1640, 1:1) supplemented with hor-

mones and supplements that favor cell-number ampliﬁcation. For cells deriving from CF patients, the culture medium contained a

complex mixture of antibiotics (usually colistin, piperacillin, and tazobactam) to eradicate bacteria in the ﬁrst days. The Ethics

Committee of the Istituto Giannina Gaslini speciﬁcally approved the collection of bronchial epithelial cells and their study to investi-

gate the mechanisms of transepithelial ion transport, following the guidelines of the Italian Ministry of Health. Each patient provided

informed consent to the study using a form that was also approved by the Ethics Committee.

2

To obtain differentiated epithelia, cells were seeded at high density on porous membranes (500.000 cells for 1 cm Snapwell in-

2

serts, Corning, code 3801, for Ussing chamber studies, or 200.000 cells for 0.33 cm mini-transwell inserts, Corning, code 3379,

for TEER/PD measurements). After 24 hr, the serum-free medium was replaced with DMEM/Ham’s F12 containing 2% fetal bovine

serum plus hormones and supplements. Differentiation of cells into a tight epithelium was checked by measuring transepithelial elec-

trical resistance and potential difference with an epithelial voltohmmeter (EVOM1, World Precision Instruments). The medium was

replaced daily on both sides of permeable supports up to 8–10 days (liquid-liquid culture, LLC). Subsequently the apical medium

was totally removed, and the cells received nutrients only from the basolateral side (air-liquid culture, ALC). This condition favored

a further differentiation of the epithelium. Cells were maintained under ALC for 2-3 weeks before experiments.

To test putative RNF5 inhibitors, compounds were added to the basolateral medium 24 hr prior to the experiments to achieve the

required concentration. Control epithelia were treated with vehicle alone (DMSO).

METHOD DETAILS

Homology Modeling

Starting from the RNF5 sequence (Q99942) downloaded from the UniProt (The UniProt Consortium, 2017) , the structure was built

with the Prime software (Schr o¨ dinger Release 2016-3: Prime, Schr o¨ dinger, LLC, New York, NY, 2016) on the basis of the structures

(

PDB: 1JM7_A, 4A4B_A, 2Y1M_A, 2Y1N_A,1FBV_A, 4A49_A, 3VGO_A, 2LDR_A, 2JMD_A, 4AYC_A, 2ECJ_A, 2ECI_A, 2XEU_A,

2

YSL_A, 3NG2_A, 3HCT_A, 3HCU_A, 4AP4_A, 3HCS_A, 3LRQ_A, 2YSJ_A, 2ECN_A, 4AUQ_B, 3FL2_A, 2ECW_A, 4FWE_A,

D8T_A, 2EA6_A, 2H0D_A, 2CKL_A, 2ECT_A, 3RPG_B, 2CT2_A, 2ECV_A, 3L11_A, 2KIZ_A), using ClustalW for sequence alignment

and composite/chimera-based as the knowledge-based building method.

Cell Chemical Biology 25, 1–15.e1–e8, July 19, 2018 e2

Please cite this article in press as: Sondo et al., Pharmacological Inhibition of the Ubiquitin Ligase RNF5 Rescues F508del-CFTR in Cystic Fibrosis

Airway Epithelia, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.04.010

Molecular Dynamics

The model of RNF5 was optimized with the Protein Preparation Wizard (Schr o¨ dinger Release 2016-3: Protein Preparation Wizard,

Schr o¨ dinger, LLC, New York, NY, 2016), and then subjected to molecular dynamics (MD) simulations.

We performed an exhaustive sampling of the orientations of groups where the H-bond network needs to be optimized. Finally, the

protein structure was reﬁned to relieve steric clashes with a restrained minimization with the OPLS2005 force ﬁeld until a ﬁnal RMSD

˚

of 0.30 A with respect to the input protein coordinates.

An MD simulation of the resulting protein structure was performed using Desmond (Schr o¨ dinger Release 2016-3: Desmond

Molecular Dynamics System, D. E. Shaw Research, New York, NY, 2016). The protein was neutralized using sodium counterions.

˚

The protein and the counterions were immersed in an orthorhombic periodic TIP4P water bath that extended about 10 A in each

direction. After an initial default relaxation protocol, an MD production run was performed for 250 ns with a time step of 2 fs.

Life Chemicals Database Preparation

The Life Chemicals (http://www.lifechemicals.com ) database collected directly from the vendor was prepared with LigPrep

(

Schr o¨ dinger Release 2016-3: LigPrep, Schr o¨ dinger, LLC, New York, NY, 2016).

The 2D structures were converted into 3D structures and, for each entry, all stereoisomers were generated. The resulting mole-

cules were submitted to Epik, and all the tautomers and ionization states at pH 7.0 ± 2.0 were calculated. Finally, we deleted dupli-

cates, compounds with more than 2 chiral centers, Pan-Assay Interference Compounds (PAINS), compounds with Michael acceptor

groups, and frequent hitters.

To enrich the database with potential Protein-Protein Interaction Inhibitors, it was ﬁltered with the PPI-HitProﬁler tool (Reyn e` s et al.,

010) using the ‘‘soft’’ methods.

2

Virtual Screening by High-Throughput Docking

From the last 100 ns of MD simulations, 10 protein conformations were collected by clustering on the basis of the RMSD. For all the

collected structures, the most druggable pocket was searched with SiteMap (Schr o¨ dinger Release 2016-3: SiteMap, Schr o¨ dinger,

LLC, New York, NY, 2016). In this pocket, all the ligands from the Life Chemicals database (about 500K) were docked with the soft-

ware Glide (Schr o¨ dinger Release 2016-3: Glide, Schr o¨ dinger, LLC, New York, NY, 2016) (SP) and the best 1000 top-scoring com-

pounds for each selected protein conformation were stored. All the selected compounds (10,000) were further ﬁltered. First, the bind-

ing modes were visually inspected and the compounds were discarded based on a visual inspection analysis. In particular, molecules

that did not show any clear chemical interactions with the target were considered false positive. This step, which depends on the

chemical intuition and experience of the user, can help addressing the well-known weakness of docking scoring functions. Then,

duplicate compounds were eliminated and the remaining compounds were grouped into 1,000 chemical clusters on the basis of

the MOLPRINT 2D ﬁngerprints using Tanimoto distance. Finally, we selected one or two representative of each cluster (depending

on the cluster size) ending up to a total of 1,623 compounds, which were eventually purchased.

Diversity Set Selection

As a ﬁrst step, we removed from the Life Chemicals database the 1,623 compounds selected by the previous virtual screening

campaign. Then, to increase chemical diversity of the ﬁnal set of compounds to submit to biological testing, we identiﬁed further

1

,000 molecules that spanned the entire descriptor space of the Life Chemical database. This step was carried out by means of

the Canvans software (Schr o¨ dinger Release 2016-3: Canvas, Schr o¨ dinger, LLC, New York, NY, 2016) on the basis of the MOLPRINT

D ﬁngerprints, using ‘‘sphere’’ as the diversity selection method.

2

Compound Library Screening

-

Twenty-four hours after seeding on 96-well plates, CFBE41o or FRT cells stably expressing F508del-CFTR and the HS-YFP were

treated with test compounds at the desired concentration, one compound per well. Cells treated with vehicle alone (DMSO) and

-

with corrector VX-809 (1 mM), respectively, served as negative and positive controls. In parallel, as an additional control, CFBE41o

cells were also reverse-transfected with siRNA targeting RNF5 (30 nM ﬁnal concentration).

Synthesis of Compounds

Scheme 1 illustrates the synthetic route to inh-2. Reaction of N-phenylbenzamidine with benzyl isothiocianate (both commercially

available) afforded thiourea 1, which underwent bromine oxidation yielding thiadiazolium salt 2. The latter was treated with acetoni-

trile (ACN) in the presence of triethylamine (TEA), yielding 1,2,4-thiadiazolylidene inh-2 as the main product, and analog-1 as by-

product.

General experimental details. Reaction progress was monitored by TLC on pre-coated silica gel plates (Kieselgel 60 F254, Merck)

and visualized by UV254 light. Flash column chromatography was performed on silica gel (particle size 40-63 mM, Merck). If required,

solvents were distilled prior to use. All reagents were obtained from the commercial sources and used without further puriﬁcation.

When stated, reactions were carried out in an inert atmosphere. For simplicity, solvents and reagents are indicated as follows: aceto-

2

nitrile (ACN), dichloromethane (DCM), diethyl ether (Et O), dimethylsulfoxide (DMSO), ethanol (EtOH), ethyl acetate (EtOAc), meth-

anol (MeOH), triethylamine (TEA). Compounds were named using the naming algorithm developed by CambridgeSoft Corporation

and used in Chem-BioDraw Ultra 15.0.

e3 Cell Chemical Biology 25, 1–15.e1–e8, July 19, 2018

Please cite this article in press as: Sondo et al., Pharmacological Inhibition of the Ubiquitin Ligase RNF5 Rescues F508del-CFTR in Cystic Fibrosis

Airway Epithelia, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.04.010

Scheme 1. Synthetic Route to inh-2

QC analysis of both compounds was performed by UPLC/MS on a Waters ACQUITY UPLC/MS system comprising a Single

Quadropole Detector (SQD) Mass Spectrometer equipped with an Electrospray Ionization interface and a Photodiode Array Detector.

Electrospray ionization in positive and negative mode was applied in the mass scan range 100-500 Da. The PDA range was 210-

4

00 nm. The separation was carried out on an ACQUITY UPLC BEH C18 column (100x2.1 mm ID, particle size 1.7 mm) with a

VanGuard BEH C18 pre-column (5x2.1 mm ID, particle size 1.7 mm).

4 2

The mobile phase was 10 mM NH OAc in H O at pH 5 adjusted with AcOH (A) and 10 mM NH OAc in MeCN-H O (95:5) at pH 5 (B).

4

2

A linear gradient was applied starting at 50% B (initial hold for 0.2 min) to 100% B in 6 min with a ﬂow rate = 0.5 ml/min.

Accurate mass data of both compounds were acquired on a Waters Synapt G2 mass spectrometer, coupled to the same chro-

matographic separation described above. High-resolution mass spectra were acquired in positive ion mode and recalibrated in

real time using leucine encephalin (2 ng/ml in water/ACN) as reference compound. NMR spectra of both compounds were acquired

on either a Bruker Avance III 400 MHz spectrometer equipped with a Broad Band Inverse probe (BBI) at 300.0 K, using 150 mL of

10 mM DMSO-6 solutions, into 3 mm disposable tubes, or using a Varian Gemini at 400 MHz and 100 MHz, respectively. Chemical

1

shifts (dH) were reported relative to TMS as the internal standard. The impurity identity was conﬁrmed by superimposing the H NMR

spectrum with that of synthesized compounds, whose structure had previously been assigned in 2D experiments, acquired as fol-

13

1

lows: H- H COSY (COrrelation SpectroscopY): 4 transients, 2048 data points, 128 increments, H- C HSQC (multiplicity edited

Heteronuclear Single Quantum Coherence): 8 transients, 1024 data points, 256 increments.

Synthetic Procedures to Obtain Compounds inh-2, analog-1, and Intermediates 1-2

(

1E,NZ)-N-(4-benzyl-3-methyl-1,2,4-thiadiazol-5(4H)-ylidene)-N’-phenylbenzimidamide inh-2 and (1E,NZ)-N-(3-methyl-4-phenyl-

1

,2,4-thiadiazol-5(4H)-ylidene)-N’-phenylbenzimidamide analog-1.

To a suspension of thiadiazolium bromide salt 2 (100 mg, 0.23 mmol, 1 equiv) in ACN (2 mL), TEA (99 mL, 0.71 mmol, 3 equiv) was

added. The mixture became a clear solution, which was reﬂuxed for 30 min, then quenched with ice. The mixture was extracted with

DCM, washed with water, dried over Na SO , and the solvent evaporated in vacuum. The crude product was puriﬁed via ﬂash silica

gel column chromatography (petroleum ether/EtOAc in 95/5 ratio), then washed with pentane and ﬁltered, to give a mixture (45 mg,

1%) of inh-2 and analog-1 as pale-yellow powder, which was further puriﬁed by preparative HPLC (see below).

2

4

5

(

isothiocyanate (795 mL, 6 mmol, 1 equiv) in 1,2-dichloroethane dry (10 mL) was heated at 55 C for 16 h, then cooled to room tem-

E)-N-(benzylcarbamothioyl)-N’-phenylbenzimidamide 1. A mixture of N-phenylbenzamidine (1.18 g, 6 mmol, 1 equiv) and benzyl

ꢁ

perature, and the solvent was evaporated. The resulting residue was puriﬁed via ﬂash silica gel column chromatography (petroleum

1

ether/EtOAc in 80/20 ratio) to give 1 as an off-white solid (830 mg, 40% yield). H NMR (400 MHz, Chloroform-d) d 12.37 (s, 1H), 8.05

(

s, 1H), 7.44 – 7.26 (m, 7H), 7.25 – 7.18 (m, 3H), 7.09 (t, J = 7.9 Hz, 2H), 6.98 – 6.88 (m, 1H), 6.67 – 6.53 (m, 2H), 4.96 (d, J = 5.3 Hz, 2H)

+

-

+

ppm. MS (ES ) 368 (345 + Na ) (ES ) 344 (345 – H )

Z)-N-(2,3-diphenyl-1,2,4-thiadiazol-5(2H)-ylidene)-1-phenylmethanamine hydrobromide 2. To a solution of thiourea 1 (460 mg,

.30 mmol, 1 equiv) in a mixture of DCM/EtOAc (4/5.2 mL), a 0.5 M solution of bromine (133 mL, 2.60 mmol, 2 equiv) in EtOAc

(

1

was added dropwise at room temperature. The thiadiazole hydrobromide salt started to precipitate during the addition. Petroleum

ꢁ

ether was added, and the mixture was left to stand at 3 C for 12 h, then ﬁltered, washed with additional petroleum ether, and dried in

vacuum to afford 2 as a yellow solid (491 mg, 89%). 1H NMR (400 MHz, DMSO-d6) d 10.12 (d, J = 6.0 Hz, 1H), 7.72 – 7.47 (m, 8H), 7.47

+

7.34 (m, 6H), 7.34 – 7.26 (m, 1H), 4.93 (d, J = 6.0 Hz, 2H) ppm. MS (ES+): 344 (M+ H ).

–

Puriﬁcation of inh-2 and analog-1

The puriﬁcation by preparative HPLC/MS was performed on a Waters LC/MS Autopuriﬁcation system comprising a 3100 Single

Quadrupole Mass Spectrometer equipped with an Electrospray Ionization Interface and a 2998 Photodiode Array Detector. The

HPLC system included a 2747 Sample Manager, 2545 Binary Gradient Module, System Fluidic Organizer, and 515 HPLC Pump.

Electrospray ionization in positive and negative mode was used in the mass scan range 100-500 Da. The PDA range was

2

10-400 nm. The puriﬁcation was performed on an XBridgeTM Prep C18 OBD column (100x19 mm ID, particle size 5 mm) with an

XBridgeTM Prep C18 (10x19 mm ID, particle size 5 mm) Guard Cartridge. The mobile phase was H O (A) and MeCN (B). A linear

gradient was applied starting at 70% B (initial hold for 0.5 min) to 90% B in 7 min with a ﬂow rate = 20 ml/min.

2

Cell Chemical Biology 25, 1–15.e1–e8, July 19, 2018 e4

Please cite this article in press as: Sondo et al., Pharmacological Inhibition of the Ubiquitin Ligase RNF5 Rescues F508del-CFTR in Cystic Fibrosis

Airway Epithelia, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.04.010

NMR Purity Assessment and Structural Elucidation of inh-2 and analog-1

NMR purity was assessed with the PULCON (PUlse Length Based Concentration Determination) external standard procedure (Wider

and Dreier, 2006), by comparing a signal of the molecule with the signal of a certiﬁed standard (Maleic Acid, Sigma Aldrich, NMR

reference standard), acquired in an independent spectrum. Brieﬂy, 16 transients were accumulated over a spectral width of 20.55

ꢁ

ppm (with the offset positioned at 6.175 ppm), after the 90 pulse optimization (Wu and Otting, 2005), at a ﬁxed receiver gain (64),

using 30 s of interpulse delay and no steady state scans.

The impurity identity was conﬁrmed by superimposing the 1H NMR spectrum with that of synthesized compounds, whose struc-

ture had previously been assigned using 2D experiments, acquired as follows: 1H-1H COSY (COrrelation SpectroscopY): 4 tran-

sients, 2048 data points, 128 increments, 1H-13C HSQC (multiplicity edited Heteronuclear Single Quantum Coherence): 8 transients,

1

024 data points, 256 increments.

Final Purity and Chemical Structures Data for inh-2 and analog-1

1

inh-2: H NMR (400 MHz, DMSO-d6) d 7.44 – 7.34 (m, 5H), 7.34 – 7.15 (m, 7H), 7.02 (d, J = 7.4 Hz, 1H), 6.76 (dt, J = 7.7, 1.1 Hz, 2H),

1

.56 (s, 2H), 2.41 (s, 3H) ppm. C NMR (101 MHz, dmso) d 170.7, 158.5, 153.7, 146.8, 135.9, 134.7, 129.5, 129.2 (2C), 128.9 (2C),

3

5

1

3

28.9 (2C), 128.0 (2C), 127.8, 127.0 (2C), 123.0, 122.0 (2C), 48.5, 16.22 ppm. HPLC-MS (ESI, m/z) Rt = 3.29 min. Accurate mass

+

85.1500 m/z (mass accuracy 3.4 ppm for brute formula C23H20N4S detected as (M+H) ). Purity evaluation by NMR: 92.2%

1

analog-1: H NMR (400 MHz, DMSO-d

6

) d 7.69 – 7.50 (m, 5H), 7.32 – 7.15 (m, 7H), 7.01 (t, J = 7.4 Hz, 1H), 6.79 – 6.69 (m, 2H), 2.19

1

3

(

s, 3H) ppm. C NMR (101 MHz, dmso) d 182.5, 159.1, 153.1, 147.0, 136.9, 135.0, 130.1 (2C), 129.9, 129.8, 129.5 (2C), 129.4 (2C),

28.6 (2C), 128.3 (2C), 123.5, 122.4 (2C), 17.3 ppm. HPLC-MS (ESI, m/z) Rt = 3.60 min. Accurate mass 371.1324 m/z (mass accuracy

1

+

1.6 ppm for brute formula C22H18N4S detected as (M+H) ). Purity evaluation by NMR: 88.7%

-

Gene Silencing by siRNA Transfection

-

CFBE41o cells expressing F508del-CFTR and the HS-YFP were reverse-transfected with 30 nM (ﬁnal concentration) siRNAs

(

Stealth, Life Technologies) using lipofectamine 2000 as transfection agent. Twenty-four hours after transfection and plating, the me-

ꢁ

dium was changed, and the cells were incubated at 37 C for an additional 24 h, prior to processing the cells for functional or biochem-

ical assays.

Fluorescence Assay for CFTR

At the time of the assay, cells were washed with PBS containing (in mM) 137 NaCl, 2.7 KCl, 8.1 Na

2 4 2 4 2

HPO , 1.5 KH PO , 1 CaCl , and

0

.5 MgCl . Cells were then incubated for 25 min with 60 ml of PBS plus forskolin (20 mM) and VX-770 (1 mM) to maximally stimulate

2

F508del-CFTR. Cells were then transferred to a microplate reader (FluoStar Galaxy; BMG Labtech, Offenburg, Germany) for CFTR

activity determination. The plate reader was equipped with high-quality excitation (HQ500/20X: 500 ± 10 nm) and emission (HQ535/

3

and 12 s after injection of 165 ml of an iodide-containing solution (PBS with Cl replaced by I ; ﬁnal I concentration 100 mM). Data were

0M: 535 ± 15 nm) ﬁlters for YFP (Chroma Technology). Each assay comprised a continuous 14-s ﬂuorescence reading, 2 s before

-

normalized to the initial background-subtracted ﬂuorescence. To determine I inﬂux rate, the ﬁnal 11 s of the data for each well were

ﬁtted with an exponential function to extrapolate initial slope (dF/dt). Reproducibility of results was conﬁrmed by performing three

independent experiments.

Transepithelial Electrical Resistance and Potential Difference Measurements (TEER/PD)

Both primary bronchial and FRT cells were treated with compounds at the indicated concentrations for 24 hr at 37 C and 5% CO

ꢁ

2

,

before measuring the Transepithelial Electrical Resistance (TEER) and/or Potential Difference (DP). In all experiments, TEER and DP

were evaluated with an epithelial voltohmmeter (EVOM1, World Precision Instruments).

For primary bronchial cells, we performed experiments in Coon’s modiﬁed Ham’s F-12 medium where NaHCO

ꢁ

3

was replaced with

0 mM Hepes. After incubation for 1 h with medium modiﬁed on both sides of permeable supports at 37 C, we measured TEER and

2

DP in basal conditions. Then, we added 10 mM amiloride, a speciﬁc inhibitor of epithelial sodium channel (ENaC). We then added

0 mM forskolin (FSK) and 50 mM Genistein on the epithelium’s apical side and 20 mM FSK on the baso-lateral side in order to activate

CFTR channels. Finally, we added 30 mM PPQ102, a speciﬁc inhibitor of CFTR, on the epithelium’s apical side. After each treatment,

2

ꢁ

we waited 10 min at 37 C before recording measurements. Data are expressed as means ± SEM, n = 4 independent epithelia (bio-

logical replicates) per experiments. Reproducibility of results was conﬁrmed by performing two independent experiments per patient.

For FRT cells, we performed experiments in a solution contained 130 mM NaCl, 2.7 mM KCl, 1.5 mM KH

2 4 2

PO , 1mM CaCl , 0.5 mM

MgCl , 10mM Na–HEPES (pH 7.3), and 10mM glucose. We measured TEER in basal conditions with the same solution on both sides

2

of permeable supports. Then, we added 20 mM forskolin (FSK) and 50 mM Genistein on the epithelium’s apical side and 20 mM FSK on

baso-lateral side in order to activate CFTR channels. Finally, we added 30 mM PPQ102, a speciﬁc inhibitor of CFTR, on the

ꢁ

epithelium’s apical side. After each treatment, we waited 10 min at 37 C before recording measurements. Data are expressed as

means ± SEM, n = 3.

Short-Circuit Current Recordings

Snapwell inserts carrying differentiated bronchial epithelia were mounted in a vertical diffusion chamber, resembling a Ussing cham-

ber with internal ﬂuid circulation. Both apical and basolateral hemichambers were ﬁlled with 5 ml of a solution containing (in mM): 126

e5 Cell Chemical Biology 25, 1–15.e1–e8, July 19, 2018

Please cite this article in press as: Sondo et al., Pharmacological Inhibition of the Ubiquitin Ligase RNF5 Rescues F508del-CFTR in Cystic Fibrosis

Airway Epithelia, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.04.010

NaCl, 0.38 KH

mixture containing 5% CO

2

PO

4

, 2.13 K

2

HPO

4

, 1 MgSO

4

, 1 CaCl

2

, 24 NaHCO

- 95% air, and the temperature of the solution was kept at 37 C. The transepithelial voltage was short-

3

, and 10 glucose. Both sides were continuously bubbled with a gas

ꢁ

2

circuited with a voltage-clamp (DVC-1000, World Precision Instruments) connected to the apical and basolateral chambers via Ag/

AgCl electrodes and agar bridges (1 M KCl in 1% agar). The offset between voltage electrodes and the ﬂuid resistance were adjusted

to compensate the parameters before experiments. The short-circuit current was recorded with a PowerLab 4/25 (ADInstruments)

analog-to-digital converter, connected to a Macintosh computer.

Data are expressed as means ± SEM, n = 5 independent epithelia (biological replicates) per experiments. Reproducibility of results

was conﬁrmed by performing two independent experiments. Statistical signiﬁcance was tested by parametric ANOVA, followed by

the Dunnet multiple comparisons test.

CFTR Immunoprecipitation (IP) Assay

-

CFBE41o cells stably expressing F508del-CFTR were grown to conﬂuence on 60-mm diameter dishes and treated for 24 h with the

indicated compounds in the absence or in the presence of MG-132 (10 mM) in the last 4 h. Then, cells were rinsed twice with ice-cold

2

PBS without Ca /Mg and then lysed with IP Lysis Buffer (#87788 Thermo Sci.) containing EDTA-free complete protease inhibitor

+

2+

ꢁ

(

Roche), N-ethylmaleimide (5 mM) and MG-132 (20 mM). Nuclei were pelletted by centrifugation at 12,000 rpm at 4 C for 20 min. Su-

pernatant protein concentration was calculated using the BCA assay (Euroclone) following the manufacturer’s instructions. An

aliquot of supernatant corresponding to 500 mg of protein was incubated for 1 hr with 2 mg/sample of mouse monoclonal anti-

CFTR R24-1 antibody (R&D), rocking at room temperature. Antibody-antigen mixture was precipitated with 25 ml/sample of Pierce

Protein A/G Magnetic Beads (Thermo Sci.) for 1 hr rocking at RT, following supplier instructions. Immunoprecipitated proteins

were eluted from the resin under reducing conditions with 100 ml Laemli Sample Buffer 1X, at RT. Equal amount of IP products

were analyzed by Western blotting (20 ml).

Western Blot

-

CFBE41o cells were grown to conﬂuence on 60-mm diameter dishes, treated for 24 hr with the indicated compounds or silenced

with the indicated siRNAs (30 nM ﬁnal concentration) and lysed in RIPA buffer containing a complete protease inhibitor (Roche).

ꢁ

Cell lysates were subjected to centrifugation at 12000 rpm at 4 C for 10 min. Supernatant protein concentration was calculated using

the BCA assay (Euroclone) following the manufacturer’s instructions. Equal amounts of protein (30 mg to detect LC3B and p62) were

separated onto gradient (4-20% depending on target protein molecular weight) Criterion TGX Precast gels (Bio-rad laboratories Inc.),

then transferred to nitrocellulose membrane with Trans-Blot Turbo system (Bio-rad Laboratories Inc.) and analyzed by Western blot-

ting. Proteins were detected using antibodies indicated in the STAR Methods section. The proteins were then visualized by chem-

iluminescence using the SuperSignal West Femto Substrate (Thermo Scientiﬁc). Chemiluminescence was monitored using the Mo-

lecular Imager ChemiDoc XRS System. Images were analyzed with ImageJ software (National Institutes of Health). Bands were

analyzed as Region-of-Interest (ROI), normalized against the GAPDH loading control. The molecular weight of the proteins (on the

basis of the Precision Plus Protein WesternC Standards, Bio-rad Laboratories Inc.) and the lane density proﬁles of ubiquitylated

CFTR were calculated using the band analysis program of the software Quantity one 4.6 of the Molecular Imager ChemiDoc XRS

System. Data are presented as mean ± SEM of independent experiments.

-

To evaluate F508del-CFTR half-life, CFBE41o cells were plated and, after 24 hr, treated with the indicated compounds. The

following day cells were treated with cycloheximide (CHX; 150 mg/ml) (SigmaAldrich). At different time point (0 - 2 - 4 - 6 hours), cells

were then lysed in RIPA buffer 1X and subjected to SDS-PAGE as previously described. Reproducibility of results was conﬁrmed by

performing three independent experiments.

Proliferation Study

-

CFBE41o cells stably expressing F508del-CFTR and the HS-YFP were plated at low density (5,000 cell/well) on 96-well plates suit-

able for high-content imaging. After 6 hours, cells were treated with different concentrations of test compounds or vehicle alone

DMSO). Cell proliferation was monitored for 48 h using the Opera Phenix (PerkinElmer) high-content screening system. Alternatively,

(

-

to monitor cytotoxic effect of high concentrations of test compounds, CFBE41o cells were plated (10,000 cell/well) and treated (after

hours) with test compounds or vehicle alone (DMSO). After 24 hours, cells were counterstained with Hoechst 33342 and propidium

6

iodide to visualized total cells and apoptotic cells, respectively, and imaged by using the Opera Phenix (PerkinElmer) high-content

screening system. Data are expressed as means ± SEM, n = 6. Reproducibility of results was conﬁrmed by performing three inde-

pendent experiments. Statistical signiﬁcance was tested by parametric ANOVA, followed by the Dunnet multiple comparisons test.

MTT Tetrazolium Cell Viability Assay

-

CFBE41o cells were plated (10.000 cell/well) on 96-well plates. After 6 hours, cells were treated with different concentrations of test

compounds or vehicle alone (DMSO). The following day, the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT)

ꢁ

substrate was added to the medium at a ﬁnal concentration of 0.5 mg/ml, and incubated for 3 hours at 37 C. Formazan production

(proportional to the number of viable cells) was then quantiﬁed by recording changes in absorbance at 560 nm using a plate reader

equipped for absorbance measurements (Victor3, Perkin Elmer).

Cell Chemical Biology 25, 1–15.e1–e8, July 19, 2018 e6

Please cite this article in press as: Sondo et al., Pharmacological Inhibition of the Ubiquitin Ligase RNF5 Rescues F508del-CFTR in Cystic Fibrosis

Airway Epithelia, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.04.010

In Vivo Ubiquitylation Assay

In vivo ubiquitylation assays were performed as previously described (Kuang et al., 2012). In brief, CFBE41o cells stably expressing

-

F508del-CFTR were grown to conﬂuence on 60-mm diameter dishes, treated for 24 hr with the indicated compounds or silenced with

the indicated siRNAs (30 nM ﬁnal concentration) and lysed in lysis buffer (2% SDS, 150 mM NaCl, 10 mM Tris-HCl pH 8.0) supple-

mented with 2 mM sodium orthovanadate, 50 mM sodium ﬂuoride, and protease inhibitors (Complete EDTA-free, Roche). Lysates

were then diluted 1:10 with dilution buffer (10 mM Tris-HCl pH 8.0, 150 mM NaCl, 2 mM EDTA, 1% Triton X-100). Diluted cell lysates

ꢁ

were subjected to centrifugation at 14000 rpm at 4 C for 30 min. Supernatants were pre-cleared using protein A-G magnetic beads

Thermo Fisher) under rotation for 30 min and then immunoprecipitated using an anti-ATG4B antibody. Immunoprecipitants were

(

subjected to SDS-PAGE for immunoblotting analysis. Detection of ubiquitin and target proteins was performed using the antibodies

indicated in the STAR Methods section. Reproducibility of results was conﬁrmed by performing three independent experiments.

Labeling of Autophagic Vacuoles with Monodansylcadaverine (MDC)

-

CFBE41o cells stably expressing F508del-CFTR and the HS-YFP were plated (50,000 cells/well) on good-quality clear-bottom 96-

well black microplates suitable for high-content imaging. After 24 hr, cells were treated with test compounds or DMSO (as negative

ꢁ

control). After 24 hr, cells were washed were incubated with 50 mM MDC (Sigma-Aldrich) in PBS at 37 C for 10 min (Biederbick et al.,

1

995). After incubation, cells were washed three times with PBS and immediately analyzed. High-content imaging and data analysis

were performed using an Opera Phenix (PerkinElmer) high-content screening system. Wells were imaged in confocal mode, using a

0X water-immersion objective. MDC signal was laser-excited at 405 nm and the emission wavelengths were collected between 435

4

and 550 nm.

Data analysis of MDC spot number was performed using the Harmony software (ver 4.5) of the Opera Phenix high-content system.

Data are expressed as means ± SEM, n = 3 independent experiments. Statistical signiﬁcance was tested by parametric ANOVA

followed by the Dunnet multiple comparisons test (all groups against the control group).

Analysis of Actin Cytoskeleton

-

CFBE41o cells stably expressing F508del-CFTR and the HS-YFP were plated (50,000 cells/well) on good-quality clear-bottom 96-

well black microplates suitable for high-content imaging. After 24 hr, cells were treated with test compounds or DMSO (as negative

control). After additional 24 hr, cells were washed, formalin-ﬁxed, and actin was labeled using phalloidin conjugated to Alexa Fluor

6

47. Cell nuclei were counterstained with Hoechst 33342. High-content imaging was performed using an Opera Phenix (PerkinElmer)

high-content screening system. Wells were imaged in confocal mode, using a 40X water-immersion objective. Phalloidin signal was

laser-excited at 640 nm and the emission wavelengths were collected between 650 and 760 nm. Excitation and emission wave-

lengths for visualization of Hoechst 33342 signal were 405 and between 435 and 480 nm, respectively.

Analysis of the signal texture was performed using the PhenoLogic machine-learning algorithm of the Harmony software (version

.5) of the Opera Phenix high-content system. Analysis of signal texture was based on evaluation of Haralick and Gabor features, two

4

subsets of well-known parameters (Haralick et al., 1973; Turner, 1986), and evaluation of SER (Spots, Edges, Ridges) features, devel-

oped by PerkinElmer and included in the Harmony software of Opera Phenix.

Scratch Wound Healing Assay

CFPAC-1 cells stably expressing the HS-YFP were plated (50,000 cells/well) on good-quality clear-bottom 96-well black microplates

suitable for high-content imaging. After 24 hr, cells had reached 100% conﬂuency. The cells were then pre-treated for 2 hr with mito-

mycin C (30 mM; Sigma-Aldrich) or DMSO. Cell monolayers in all wells were then scratched mechanically, using pipette tips mounted

on a 12-channel pipette. After the scratch, wells were washed with PBS to remove detached cells and debris and fresh medium was

added, containing test compounds (5 mM) or vehicle alone (DMSO; as negative control) in the absence (for cells pre-treated with

DMSO) or presence of mitomycin C (30 mM; for cells pre-treated with this compound). For each condition, 32 individually scratched

monolayers were assayed, and we performed three independent experiments. High-content imaging and data analysis for each well

was performed using an Opera Phenix (PerkinElmer) high-content screening system. Wells were imaged using a 10X air objective.

YFP signal was laser-excited at 490-500 nm and the emission wavelengths were collected between 500 and 550 nm.

Data analysis of open area percentage was performed using the Harmony software (ver 4.5) of the Opera Phenix high-content

system.

Data are expressed as means ± SEM, n = 32 independently scratched monolayers per experimental condition. Reproducibility of

results was conﬁrmed by performing three independent experiments. Statistical signiﬁcance was tested by parametric ANOVA fol-

lowed by the Dunnet multiple comparisons test (all groups against the control group).

QUANTIFICATION AND STATISTICAL ANALYSIS

We used the analysis of variance (ANOVA) followed by a post hoc test to avoid ‘‘multiple comparison errors’’ when comparing more

than 2 groups. For normally distributed quantitative variables, a parametric ANOVA was performed. When the quantitative variables

were skewed, the non-parametric ANOVA (Kruskal-Wallis test) was applied. The Kolmogorov-Smirnov test was used to evaluate the

assumption of normality.

e7 Cell Chemical Biology 25, 1–15.e1–e8, July 19, 2018

Please cite this article in press as: Sondo et al., Pharmacological Inhibition of the Ubiquitin Ligase RNF5 Rescues F508del-CFTR in Cystic Fibrosis

Airway Epithelia, Cell Chemical Biology (2018), https://doi.org/10.1016/j.chembiol.2018.04.010

-

Statistical signiﬁcance of the effect of single siRNA treatments on CFTR activity or expression in CFBE41o cells was tested by

parametric one-way analysis of variance (ANOVA) followed by the Dunnet multiple comparisons test (all groups against the control

group) as a post hoc test. For combinations of siRNAs against more than one target, we veriﬁed the statistical signiﬁcance using

ANOVA followed by the Tukey test (for multiple comparisons) as a post hoc test.

Normally distributed data are expressed as mean ± SEM, while skewed distributed data are expressed as median (min-max), and

signiﬁcances are two-sided. Differences were considered statistically signiﬁcant when p < 0.05.

Cell Chemical Biology 25, 1–15.e1–e8, July 19, 2018 e8

Article Doi

Pharmacological Inhibition of the Ubiquitin Ligase RNF5 Rescues F508del-CFTR in Cystic Fibrosis Airway Epithelia

DOI: 10.1016/j.chembiol.2018.04.010

Source and publish data:

Authors:

Article abstract of DOI:10.1016/j.chembiol.2018.04.010

Full text of DOI:10.1016/j.chembiol.2018.04.010

Products guided by the article

R&D Labs maybe for 65739-30-2

Relevant to this article

Hot Product