mvae038-html.html

Cytotoxic stress caused by azalamellarin D (AzaD) interferes with cellular protein translation by

targeting the nutrient-sensing kinase mTOR

Tirawit Meerod

, Rapeepat Sangsuwan

, Kanawut Klumthong

, Bunkuea Chantrathonkul

, Nadgrita

Phutubtim

, Piyarat Govitrapong

, Somsak Ruchirawat

3,4,5

, Poonsakdi Ploypradith

3,4,5

, and Pattarawut

Sopha

1,5,*

Program in Applied Biological Sciences: Environmental Health, Chulabhorn Graduate Institute, 906

Kamphaeng Phet 6 Road, Lak Si, Bangkok 10210, Thailand

Laboratory of Natural Products, Chulabhorn Research Institute, 54 Kamphaeng Phet 6 Road, Lak Si, Bangkok

10210, Thailand

Program in Chemical Sciences, Chulabhorn Graduate Institute, 906 Kamphaeng Phet 6 Road, Lak Si,

Bangkok 10210, Thailand

Laboratory of Medicinal Chemistry, Chulabhorn Research Institute, 54 Kamphaeng Phet 6 Road, Lak Si,

Bangkok 10210, Thailand

Center of Excellence on Environmental Health and Toxicology, Office of the Permanent Secretary (OPS),

Ministry of Higher Education, Science, Research and Innovation (MHESI), Bangkok 10400, Thailand

Running title:

Azalamellarin D interferes with metabolic kinases.

ORIGINAL

UNEDITED

MANUSCRIPT

Downloaded from https://academic.oup.com/jb/advance-article/doi/10.1093/jb/mvae038/7658884 by UCSD-Philosophy user on 01 May 2024

* Corresponding author:

Dr. Pattarawut Sopha

Address: Program in Applied Biological Sciences: Environmental Health, Chulabhorn Graduate Institute, 906

Kamphaeng Phet 6 Road, Lak Si, Bangkok 10210 Thailand.

Email addresses: pattarawut@cgi.ac.th

Abbreviations:

4EBP1: eukaryotic initiation factor 4E binding protein 1, ATF4: activating transcription factor 4, AzaChi:

azalamellarin



, AzaD: azalamellarin D, Bort: bortezomib, BSA: bovine serum albumin, cCasp: cleaved

caspase, CHOP: CCAAT enhancer binding protein homologous protein, CETSA: cellular thermal shift assay,

COX4: cytochrome c oxidase subunit 4, cPARP: cleaved poly (ADP) ribose polymerase 1, DMSO: dimethyl

sulfoxide, eIF2ɑ: eukaryotic initiation factor 2 alpha, ER: endoplasmic reticulum, ETC: electron transport

chain, FBS: fetal bovine serum, GCN2: general control nonderepressible 2, HRI: heme-regulated inhibitor,

HRP: horseradish peroxidase, IC

: half maximal inhibitory concentration, ISR: integrated stress response,

KO: knockout, LamChi: lamellarin



, LamD: lamellarin D, LC3: microtubule-associated protein 1 light chain

3, mTOR: mechanistic target of rapamycin, NMR: nuclear magnetic resonance, PBS: phosphate buffered

saline, PERK: PKR-like ER kinase, PKR: protein kinase R, PMB: para-methoxybenzyl, PMSF:

phenylmethanesulfonylfluoride, S6K1: ribosomal protein S6 kinase 1, SD: standard deviation, SDS: sodium

dodecyl sulfate, SDS‒PAGE: sodium dodecyl sulfate‒polyacrylamide gel electrophoresis, ULK1: unc-51 like

autophagy activating kinase 1, VDAC: voltage-dependent anion channel, WT: wildtype

ORIGINAL

UNEDITED

MANUSCRIPT

Downloaded from https://academic.oup.com/jb/advance-article/doi/10.1093/jb/mvae038/7658884 by UCSD-Philosophy user on 01 May 2024

Abstract

Analogs of pyrrole alkaloid lamellarins exhibit anticancer activity by modulating multiple cellular

events. Lethal doses of several lamellarins were found to enhance autophagy flux in HeLa cells, suggesting

that lamellarins may modulate protein homeostasis through the interference of proteins or kinases controlling

energy and nutrient metabolism. To further delineate molecular mechanisms and their targets, our results

herein show that azalamellarin D (AzaD) cytotoxicity could cause translational attenuation, as indicated by a

change in eIF2



phosphorylation. Intriguingly, acute AzaD treatment promoted the phosphorylation of GCN2,

a kinase that transduces the integrated stress response (ISR), and prolonged exposure to AzaD could increase

the levels of the phosphorylated forms of eIF2



and the other ISR kinase PKR. However, the effects of AzaD

on ISR signaling were marginally abrogated in cells with genetic deletion of GCN2 and PKR, and evaluation

of protein target engagement by CETSA revealed no significant interaction between AzaD and ISR kinases.

Further investigation revealed that acute AzaD treatment negatively affected mTOR phosphorylation and

signaling. The analyses by CETSA and computational modeling indicated that mTOR may be a possible

protein target for AzaD. These findings indicate the potential for developing lamellarins as novel agents for

cancer treatment.

Keywords:

Lamellarin pyrrole alkaloids, Amino acid-sensing kinases, Cytotoxicity, Translational attenuation,

Cancer

ORIGINAL

UNEDITED

MANUSCRIPT

Downloaded from https://academic.oup.com/jb/advance-article/doi/10.1093/jb/mvae038/7658884 by UCSD-Philosophy user on 01 May 2024

Introduction

Cancer cells adjust their metabolic rate, nutrient consumption, and protein synthesis

via

kinase-controlled

cellular pathways to manipulate homeostasis and overcome microenvironmental stress (

1-3

). These metabolic

pathways not only benefit proliferation and evade the programmed cell death of cancer cells but also have

significant functions in the growth and survival of normal cells (

1, 2, 4

). As a part of cellular adaptation,

protein translation is initiated by phosphorylation of the eukaryotic translational initiation factor 2 (eIF2



)

alpha subunit at serine residue 51 (Ser-51), which is catalyzed by kinases, including general control

nonderepressible 2 (GCN2), protein kinase R (PKR), PKR-like ER kinase (PERK), and heme-regulated

inhibitor (HRI), in response to amino acid depletion, ER stress, double-stranded RNA, and heme deficiency,

respectively (

2, 5, 6

). Under severe and chronic stress conditions, the prolonged accumulation of

phosphorylated eIF2



can selectively activate the translation of uORF-containing mRNAs, including

transcription factor 4 (ATF4) and CCAAT enhancer binding protein (C/EBP)-homologous protein (CHOP), to

trigger the integrated stress response (ISR) and relieve stress or terminate damaged cells

via

apoptosis (

2, 6

Additionally, the homeostasis of protein synthesis and degradation can also be controlled at the initiation step

of protein translation by regulating the phosphorylation of kinases, specifically mechanistic target of

rapamycin (mTOR), which senses the levels of cellular energy and nutrients (such as ATP and amino acids)

and phosphorylates the target proteins involved in cap-dependent protein translation, including eukaryotic

initiation factor 4E binding protein 1 (4EBP1) and the ribosomal protein S6 kinase 1 (S6K1) (

). In addition, in

the presence of sufficient nutrients, mTOR can prevent lysosomal-mediated protein degradation or autophagy

by phosphorylating UNC-51-like autophagy activating kinase 1 (ULK1) at Ser-757 (

). Modulating cancer cell

growth

via

these important and elaborate pathways may assist in the discovery of novel anticancer therapeutics

and options for combating cancer chemotherapeutic resistance.

A group of pyrrole alkaloids isolated from marine organisms, lamellarins, possess a unique

pentacyclic framework of the pyrroloisoquinoline chromenone with a non-fused orthogonal aromatic ring at

C1 (Fig. 1A-C). Following the successful developments of their total syntheses, they have been demonstrated

to be cytotoxic toward several cancer cell lines and have been studied for their biological activities to establish

their structure-activity relationships (

8, 9

). Previous studies have shown that some lamellarins can inhibit

ORIGINAL

UNEDITED

MANUSCRIPT

Downloaded from https://academic.oup.com/jb/advance-article/doi/10.1093/jb/mvae038/7658884 by UCSD-Philosophy user on 01 May 2024

nuclear and mitochondrial topoisomerases, interfere with the mitochondrial electron transport chain (ETC),

and mediate apoptosis in caspase-dependent and caspase-independent manners (

8-10

). In addition, lamellarins

were shown to inhibit several kinases related to cancer cell growth and proliferation (

), although the roles of

these kinases in response to lamellarin exposure have not been determined at the

in vivo

or molecular level.

Although several cellular proteins have been proposed as putative targets for lamellarins, data on the molecular

interactions between lamellarins and the proposed protein targets are available only from computational

analyses (

9, 12

). Consequently, investigating how lamellarins affect cellular responses and the stress response

machinery of cancer cells may provide valuable information for the development of anticancer therapeutics

based on lamellarin pharmacophores.

Recently, novel evidence suggested that treatment with lamellarin D (LamD) in HeLa cells could

potentiate cellular stress, leading to disturbed mitochondrial potential, dysregulated mitochondrial dynamics,

and elevated autophagy flux (

). Intriguingly, similar results with greater magnitudes of those cellular

responses were also reported in cells treated with synthetic lamellarins with lactam ring substitution

(azalamellarins), however, the role of lactone-to-lactam conversion is still unknown at the molecular level (

12-

). Because of the change in autophagy flux in HeLa cells, lamellarin-mediated cytotoxic stress was

speculated to involve the interference of essential kinase-controlled metabolic pathways. In this study, we

evaluated the effects of azalamellarin D (AzaD) on the induction and function of signaling pathways associated

with nutrient and translational control, including the ISR and mTOR pathway. Information about the protein

targets of AzaD related to the ISR and mTOR signaling was also obtained. The effect of AzaD on metabolic

signaling and its interaction with target kinases may provide important insights for the development of novel

anticancer agents.

Materials and methods

Cell culture and chemical treatment

Human cervical cancer (HeLa) and human hepatocellular carcinoma (Huh-7) cells were cultured in Dulbecco’s

modified Eagle medium (DMEM) (HyClone, GE, USA) supplemented with 10% heat-inactivated fetal bovine

serum (FBS) (HyClone, GE, USA) in a humidified incubator at 37



C with 5% CO

. To perform the cell

ORIGINAL

UNEDITED

MANUSCRIPT

Downloaded from https://academic.oup.com/jb/advance-article/doi/10.1093/jb/mvae038/7658884 by UCSD-Philosophy user on 01 May 2024

treatment experiments, the cultured cells were maintained under these conditions for 48 h before exposure to

the designated compounds.

The chemical stocks of lamellarins and other compounds were prepared and diluted in

dimethylsulfoxide (DMSO) (Merck, Germany), and the evaluated concentrations were directly prepared in

DMEM containing FBS and antibiotics. Bortezomib or Bort (MedChemExpress, USA) was used to induce the

downstream effectors of ISR. To investigate mTOR signaling inhibition, the specific mTOR inhibitors used in

this study included torin2, torkinib, and rapamycin (MedChemExpress, USA).

Chemical synthesis

Lactam-containing lamellarins were synthesized as previously described (

). The compounds were verified

by NMR spectroscopy. The mass spectrometry data of azalamellarin



(AzaChi) and its derivatives showing

bulky groups on the N atom of the lactam ring are available in the supplemental materials.

Antibodies

The following mouse and rabbit primary antibodies were used for western blotting in this study and purchased

from Cell Signaling Technology, USA: anti-



-actin (#3700), anti-phospho-4EBP1 Thr37/46 (#2855), anti-Akt

(#4691), anti-phospho-Akt Thr-308 (#13038), anti-phospho-Akt Ser-473 (#4060), anti-cleaved Caspase-3

(#9664), anti-cleaved Caspase-9 (#52873), anti-CHOP (#2895), anti-COX IV (#4850), anti-cleaved PARP

(#5625), anti-eIF2



(#9722), anti-phospho-eIF2



Ser-51 (#3597), anti-LC3A/B (#12741), anti-mTOR

(#2983), anti-phospho-mTOR Ser-2448 (#5536), anti-PERK (#5683), anti-phospho-S6K1 Thr-389 (#9234),

anti-ULK1 (#8054), anti-phospho-ULK1 Ser-757 (14202), and anti-VDAC (#4661). Additional primary

antibodies, including anti-



-tubulin (32-2600), anti-GCN2 (Ab134053), anti-phospho-GCN2 Thr-899

(Ab75836), anti-PKR (32506), anti-phospho-PKR Thr-446 (Ab32036), and anti-ATF4 (PA5-13293), were

obtained from Abcam, UK, and Thermo Fisher Scientific, USA. The HRP-conjugated secondary antibodies

used in this study were goat anti-rabbit IgG (#7074) and horse anti-mouse IgG (#7076) from Cell Signaling

Technology, whereas Alexa Fluor® 488-linked goat anti-rabbit IgG (ab150077, Abcam) was used for the flow

cytometry detection of intracellular eIF2



-p.

ORIGINAL

UNEDITED

MANUSCRIPT

Downloaded from https://academic.oup.com/jb/advance-article/doi/10.1093/jb/mvae038/7658884 by UCSD-Philosophy user on 01 May 2024

Protein preparation and western blotting

Protein extracts prepared from whole-cell lysates were sonicated as previously described (

10, 15

). The protein

concentration was evaluated using the DC Protein Assay (Bio-Rad, USA) before adjusting to 1-2



L in 1×

SDS‒PAGE buffer.

The homogenized total protein lysates were then separated by SDS‒PAGE with 15–20



of total protein per lane. Proteins were transferred to a nitrocellulose membrane (Bio-Rad, USA) using a wet

blot transfer system.

The proteins on the blotted membranes were detected using appropriate antibodies. The

western blot signals were developed using Clarity ECL substrate (Bio-Rad, USA), captured using a G:box

chemiluminescent detector (Syngene, India), and analyzed using the ImageJ/Fiji image analysis program (NIH,

USA). The images of the protein bands were only processed with the level parameter in the ImageJ/Fiji

application before being used in the figures. Statistical analysis was performed using Prism 9 (GraphPad,

USA). All the data are presented as the mean ± standard deviation (SD), and a

value of less than 0.05

indicated statistical significance.

CRISPR‒Cas9 genome editing with Cas9 nickases for GCN2 and PKR knockout

GCN2

and

PKR

gene knockout (KO) HeLa cells were generated by double-nickase CRISPR‒Cas9 gene

editing with Cas9 nickase as described in previous literature (

). Oligonucleotides targeting

GCN2

PKR

gene loci (Supplementary Table S1) were cloned and inserted into the

BbsI

site of pSpCAS9n-2A-puro

(PX462), a gift from Feng Zhang (Addgene plasmid # 62987; http://n2t.net/addgene:62987;

RRID:Addgene_62987). The recombinant plasmids were verified by automated DNA sequencing before

transfection of HeLa cells with Lipofectamine LTX (Invitrogen, USA). Successfully transfected cells were

selected in DMEM containing 1 μg/mL puromycin, FBS, and antibiotics for 72 h. Puromycin-resistant cells

were cultured in a 100-mm culture plate, and the isolated clones were subsequently picked up by the agarose

overlay method with blunted pipette tips. GCN2 and PKR knockout were finally confirmed by western

blotting.

Analysis by flow cytometry

ORIGINAL

UNEDITED

MANUSCRIPT

Downloaded from https://academic.oup.com/jb/advance-article/doi/10.1093/jb/mvae038/7658884 by UCSD-Philosophy user on 01 May 2024

Flow cytometry analysis in this study was performed using the Amnis

FlowSight

imaging flow cytometry

system (Luminex Corporation, USA). To analyze the intracellular level of eIF2



-p, HeLa cells were seeded in

a 6-well plate at 6



cells per well. Upon completion of cell treatment, the cells were detached by trypsin,

fixed in 2% paraformaldehyde (PFA) in 1x PBS for 10 min, and permeabilized in ice-cold methanol for 10

min. The cells were immediately washed three times in 1x PBS containing 1% BSA and incubated with a

phospho-specific antibody for eIF2



-p Ser-51 at a 1:1000 dilution in 1x PBS containing 1% BSA for 45 min at

RT. The cells were washed twice with 1× PBS containing 1% BSA and then incubated with Alexa Fluor®

488-linked goat anti-rabbit IgG at a 1:1000 dilution and 300 nM DAPI for 45 min at RT. The stained cells

were washed twice in 1× PBS containing 1% BSA, resuspended in 1× PBS, and analyzed.

Cell viability assay

To evaluate the cytotoxicity of lamellarin treatment, the viability of the treated cells was assessed by the ability

of the cells to reduce the tetrazolium dye MTT reagent (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium

bromide). Cells were seeded at 2



cells per well into a 96-well plate. Upon completion of treatment, the

media were replaced with media containing 2.5



L MTT reagent, and the cells were incubated at 37



C for

3 h. The reactions were supplemented with 150



L of MTT solvent (4 mM HCl, 0.1% NP-40 in isopropanol).

The optical density absorbance at 590 nm (OD

590

) was measured and the cell viability was determined.

Cellular thermal shift assay (CETSA)

A CETSA was performed to evaluate the interaction between AzaD and target proteins (

). Briefly, HeLa

cells were treated with AzaD for 20 min and harvested for solubilization in ice-cold NP-40 lysis buffer (0.4%

NP-40, 50 mM Tris-HCl (pH 7.5), 150 mM NaCl, and 0.05% SDS) supplemented with 1 mM PMSF and 1x

cOmplete

™

Protease Inhibitor Cocktail (Roche, Switzerland). The lysates were isolated by centrifugation at

20,000 ×g and 4



C for 15 min. The clarified lysates were aliquoted for heating at the indicated temperatures

for 5 min, and the insoluble fractions were removed by centrifugation at 20,000 ×g and 4 °C for 20 min. The

ORIGINAL

UNEDITED

MANUSCRIPT

Downloaded from https://academic.oup.com/jb/advance-article/doi/10.1093/jb/mvae038/7658884 by UCSD-Philosophy user on 01 May 2024

soluble protein fractions were diluted with 2× SDS sample loading buffer prior to western blot analysis with

the indicated antibodies.

Molecular docking

The energy of LamD, AzaD, and known mTOR ligands (torin2 and resveratrol) was minimized using the

MM2 force field in Scigress software. The protein structure of mTOR with torkinib as the cocrystallized ligand

(PDB: 4JT5) was obtained from the Protein Data Bank and prepared by removing the cocrystallized ligand,

removing water molecules, adding hydrogen atoms, and adjusting the protonation stage according to pH 7.4

using Discovery Studio software. The docking analysis was performed by GOLD software (version 5.8.1), and

the docking scores, represented as GoldScore fitness, were used to rank different binding modes. The

GoldScore fitness is based on a molecular mechanics–like function in which factors such as the hydrogen

bonding energy, van der Waals energy, metal interaction, and ligand torsion strain have been considered (

Molecular docking was carried out with the center of the binding pocket set according to the position of the

cocrystallized ligand in the pocket (30 Å radius, x=-18.199276, y=-33.082009, z=-55.057455). The calculation

was performed with 50 docking runs and 100% search efficiency. The higher the docking score is, the higher

the probability that the compound binds to mTOR is. Discovery Studio software was used to visualize the 2D

and 3D binding mode results. The distance and dihedral angle (DHA) of H-bonds between ligands and key

amino acid residues of the mTOR binding pocket were determined.

Results

HeLa cervical cancer cells treated with azalamellarin D (AzaD) exhibit cytotoxic stress associated with

multiple cellular events, including interference with autophagy, induction of apoptosis, and phosphorylation

of the translational factor, eIF2



Previous findings showed that modification of the lactone-to-lactam modification of the lamellarin framework

could enhance its physicochemical and biological properties (

12-14

). Azalamellarin D (AzaD) was previously

developed to replace the lactone ring of lamellarin D (LamD) with a lactam moiety (

) (Fig. 1A-B). The MTT

assay results showed that AzaD treatment for 18 h effectively inhibited HeLa cell viability of more than 60%

ORIGINAL

UNEDITED

MANUSCRIPT

Downloaded from https://academic.oup.com/jb/advance-article/doi/10.1093/jb/mvae038/7658884 by UCSD-Philosophy user on 01 May 2024

at concentrations of 1–5



M, compared to 35% and 60% inhibition by LamD treatment at 1 and 5



respectively (Fig. 1D). However, the lactone ring is not the only structure that determines the cytotoxicity of

lamellarin. The bonding between C5-C6 at the D-ring, among other positions, is also involved in mediating

biological activities (

8, 9

). Lamellarin



(LamChi), which possesses a single bond instead of a double bond in

the D ring, was less effective at inhibiting HeLa cell viability (~40% inhibition at 5



M) than AzaD or LamD

(Fig. 1C-D). Previous findings have also indicated that LamD-mediated cytotoxicity is correlated with the

induction of intrinsic or mitochondrial apoptosis (

8, 9

). The HeLa cell lysates treated with LamD showed an

accumulation of active executioner caspase-3 (cleaved caspase-3 - cCasp-3) and the cCasp-3 substrate nuclear

poly (ADP) ribose polymerase-1 (cPARP), which can be detected in cells undergoing mitochondrial apoptosis

(Fig. 1E-F and Supplementary Fig. S1). However, AzaD-treated lysates revealed a much greater accumulation

of cCasp-3 and cPARP, whereas in the same experiment, treatment with LamChi weakly upregulated these

proteins (Fig. 1E- F and Supplementary Fig. S1). Therefore, lactone-to-lactam modification in AzaD has

substantial effects on the biological activities of lamellarins.

A previous study revealed that lamellarins can cause cytotoxicity

via

accelerated autophagy flux (

The ratio of microtubule-associated protein 1 light chain 3 (LC3) from soluble LC3-I to

phosphatidylethanolamine (PE)-conjugated LC3-II in HeLa cells showed that LamD, AzaD, and LamChi

enhanced autophagy compared to DMSO (Fig. 1G and Supplementary Fig. S1). Autophagy is a catabolic

process controlled by protein kinases in response to nutrients and energy, and autophagic flux induction is

observed under the suppression of protein synthesis (

18, 19

). Moreover, our past data showed that

mitochondria are strongly affected by cytotoxic lamellarins (

), and several recent articles revealed that the

proteostasis of mitochondria could be modulated by stress stimuli in the cytosol

via

the integrated stress

response (ISR), a pathway that attenuates global translation (

6, 20

). We then determined the phosphorylation

status of the



subunit of eukaryotic translational initiation factor 2 (eIF2



) at Ser-51, a hallmark of ISR

pathway activation. In the HeLa cells treated with AzaD for 6 h, we found that eIF2



phosphorylation (eIF2



p) was more than 10-fold greater than that in DMSO-treated cells, and LamD-treated cells exhibited

approximately 3-fold induction, whereas LamChi treatment did not really affect the level of eIF2



-p (Fig. 1H

ORIGINAL

UNEDITED

MANUSCRIPT

Downloaded from https://academic.oup.com/jb/advance-article/doi/10.1093/jb/mvae038/7658884 by UCSD-Philosophy user on 01 May 2024

and Supplementary Fig. S1). Therefore, these data suggest that AzaD can mediate potent cytotoxicity by

altering intracellular events such as apoptosis, autophagy, and protein translation.

The lactam moiety and a double bond at the D-ring in the lamellarin structure are essential for the

biological activities of AzaD.

We further investigated which chemical determinants of the lamellarin structure are involved in its ability to

promote the phosphorylation of eIF2



. A new lamellarin compound with a lactam ring and a single bond at the

D-ring was synthesized, namely, azalamellarin



(AzaChi) (Fig. 2A), and its structure was similar to that of

LamChi. The cell lysates treated with AzaChi showed an approximately 2-fold increase in eIF2



-p levels

beginning at 6 h after treatment compared with the basal level at 0 h (Fig. 2B and Supplementary Fig. S2). In

contrast, the parallel data from the AzaD treatment group exhibited a 4-fold change (Fig. 2B and

Supplementary Fig. S2). Interestingly, the ability of AzaChi to inhibit HeLa cell viability was comparable to

that of AzaD; cell viability was inhibited by ~60% at 5



M for the 18-h treatment (Fig. 2C). Therefore, the

degree of unsaturation at C5-C6 of the D-ring is involved in the ability of lamellarins to promote eIF2



-p

levels. We further assessed the effect of lactam ring-mediated cytotoxicity through the induction of eIF2



-p by

synthesizing lactam-containing derivatives that have bulky groups on the N atom of the lactam ring, creating

AzaD-P (PMB), AzaD-C3 (

-propyl), and AzaD-C4 (

-butyl) (Fig. 2D). The results indicated that the HeLa

cells treated with AzaD-P, AzaD-C3, and AzaD-C4 for 6 h did not increase eIF2



-p levels, promote LC3

conversion, or enhance the accumulation of cleaved caspase-9 (cCasp-9) and cPARP, unlike the cells treated

with AzaD

(Fig. 2E–F and Supplementary Fig. S3). Moreover, these derivatives did not decrease the viability

of treated HeLa cells (Fig. 2C). Therefore, the lactam moiety and the double bond at the D-ring of the AzaD

structure are involved in the chemical interactions that mediate its cytotoxicity and ability to phosphorylate

eIF2



AzaD-mediated cytotoxicity induced ISR signaling through the sequential activation of GCN2 and PKR.

ORIGINAL

UNEDITED

MANUSCRIPT

Downloaded from https://academic.oup.com/jb/advance-article/doi/10.1093/jb/mvae038/7658884 by UCSD-Philosophy user on 01 May 2024

Our data indicated that HeLa cells treated with AzaD exhibited high levels of LC3-II and eIF2



-p, which was

a result of a process controlled by protein kinases that regulate cellular metabolism, nutrients, energy, and

protein translation (

1, 6, 18

). We further investigated how eIF2



was phosphorylated under AzaD-induced

cytotoxicity in HeLa cells to confirm that AzaD could specifically affect cellular protein translation. The

detection of the intracellular level of eIF2



-p using flow cytometry showed that the presence of eIF2



-p after

AzaD treatment was detected as soon as 2 h, and the level increased until 8 h (Fig. 3A). The data were also

confirmed by western blot analysis, and a similar pattern was observed (Fig. 3B and Supplementary Fig. S4).

The presence of eIF2



-p indicates the attenuation of global protein translation in response to stress stimuli (

). Therefore, we detected whether the levels of some kinases controlling eIF2



-p, including GCN2, PKR, and

PERK, changed in response to AzaD-mediated cytotoxicity. The detection of PKR phosphorylation at Thr-446,

which represents the activation of PKR, revealed that PKR-p was predominantly induced by approximately 2-

fold at 4–8 h compared with 0 h (Fig. 3C and Supplementary Fig. S4), and the induction pattern of PKR-p was

similar to that of eIF2



-p (Fig. 3A-C). In the same sample set, the level of GCN2 phosphorylation at Thr-899

was ~4 times greater at 2 h after AzaD challenge, and it gradually declined to the basal level at 8 h (Fig. 3D

and Supplementary Fig. S4). In addition, PERK phosphorylation was unchanged in the presence of AzaD

(Supplementary Fig. S5), as a shift in PERK band migration from phosphorylation could be conveniently

detected under stress conditions in the ER (

). We further detected GCN2-p and PKR-p during the acute

phase of AzaD treatment, and the data indicated that approximately 5-fold greater amounts of GCN2-p were

detected 15 min after AzaD treatment and lasted until 120 min, while eIF2



-p and PKR-p were rather

unaffected during that period (Supplementary Fig. S6). Therefore, the data suggest that both ISR kinases,

GCN2 and PKR, are sequentially activated in response to AzaD-induced cytotoxic stress. To investigate

whether these cellular events were conserved in human cancer cells, we performed the same experiment using

hepatocellular carcinoma Huh-7 cells. AzaD-treated Huh-7 cells showed sequential activation of GCN2-p and

PKR-p, a pattern similar to that observed in HeLa cells (Supplementary Fig. S7).

An increased level of eIF2



-p allows the selective translation of mRNAs containing uORFs, which,

upon stress, can be translated to protein effectors for cell recovery or cell death, such as ATF4 and CHOP (

ORIGINAL

UNEDITED

MANUSCRIPT

Downloaded from https://academic.oup.com/jb/advance-article/doi/10.1093/jb/mvae038/7658884 by UCSD-Philosophy user on 01 May 2024

). We then explored how AzaD treatment affects this downstream ISR signaling pathway. The lysates

prepared from the AzaD-treated HeLa cells showed elevated levels of ATF4 (1.5-1.8-fold) at 4-24 h after

treatment (Fig. 3E, Lanes 1-5), however for the level of CHOP in the same lysates were not elevated. Because

of the transient stability of CHOP, we questioned whether CHOP could be a substrate for rapid degradation by

the proteasome under AzaD-induced cytotoxicity. To address this issue, we cotreated HeLa cells with 5



AzaD and 15



M bortezomib, a proteasome inhibitor. The presence of bortezomib after AzaD treatment did

not affect the extent of AzaD-mediated ATF4 induction. However, CHOP could still not be detected under

these conditions (Fig. 3E, Lane 1 compared to 6-9). In contrast, in cells treated with only bortezomib, CHOP

was substantially induced for during 4-24 h after treatment (Fig. 3E, Lane 1 compared to 10-13). Therefore,

these data indicate that AzaD treatment selectively increases ATF4 levels but does not affect CHOP

accumulation. We also performed a parallel experiment in which HeLa cells were challenged with a

combination of LamD and bortezomib. Surprisingly, LamD treatment alone resulted in a detectable level of

CHOP, whereas bortezomib cotreatment promoted the accumulation of ATF4 and CHOP (Supplementary Fig.

S8). These data implicate the specific effect of AzaD on ISR signaling.

Depletion of GCN2 and PKR marginally affected ISR signaling upon AzaD treatment, and neither kinase

showed ligand engagement with AzaD according to CETSA.

To further identify whether the ISR kinases GCN2 and PKR are the molecular targets of AzaD, we

investigated the ISR in cells lacking these kinases, expecting that ISR signaling would be discretely affected

by AzaD toxicity. Therefore, we established knockout (KO) cell lines using a CRISPR‒Cas9 with Cas9

nickase approach in which the

GCN2

and

PKR

genomic loci were disrupted in HeLa cells. Cells lacking these

specific ISR kinases were successfully selected and isolated (Supplementary Fig. S9). We evaluated AzaD

cytotoxicity in GCN2-KO cells within a 48-h period and determined the phosphorylation of eIF2



and PKR.

The AzaD-treated GCN2-KO cell lysates showed that eIF2



-p was maintained at approximately 1.5-fold the

baseline level for 8-48 h. However, the eIF2



-p level in parental HeLa cells declined after 24 h (~0.5-fold) and

then completely returned to the baseline level at 48 h (Fig. 4A). Moreover, the lack of GCN2 did not affect

ATF4 induction or PKR phosphorylation mediated by AzaD cytotoxicity.

ORIGINAL

UNEDITED

MANUSCRIPT

Downloaded from https://academic.oup.com/jb/advance-article/doi/10.1093/jb/mvae038/7658884 by UCSD-Philosophy user on 01 May 2024

We also evaluated AzaD-induced cytotoxicity in PKR-KO cells. However, the results indicated that

compared with parental cells, AzaD-treated PKR-KO cells exhibited a slight effect on the expression of eIF2



p and increased expression of GCN2-p. Furthermore, the cells lacking PKR exhibited a pattern of ATF4

activation similar to that of the wildtype parental cells (Fig. 4B). Collectively, these data suggest that eIF2



phosphorylation and ATF4 induction in GCN2-KO and PKR-KO cells under AzaD treatment are still ongoing.

We subsequently determined whether AzaD can directly interact with GCN2 and PKR using a

CETSA (

6). To assess drug-target interactions, a CETSA can be used to determine the change in the thermal

stability of the protein target upon interaction with the compound or ligand. After HeLa cells were treated with



M AzaD for 30 min, the treated cell lysates were subjected to western blotting, which indicated that

compared with DMSO, the GCN2 protein was slightly protected at 46.6-50.8



C (Fig. 4C and Supplementary

Fig. S10). Using the same lysates, the denaturation kinetics of PKR did not differ between AzaD-treated and

DMSO-treated cells (Fig. 4D and Supplementary Fig. S10). However, these results did not fully support GCN2

or PKR as direct molecular targets for AzaD.

Acute treatment with AzaD caused abrupt alterations in the phosphorylation of mTOR, another protein

kinase that controls cellular nutrients, metabolism, and protein translation.

The data from the previous sections indicated that GCN2 and PKR may not be molecular targets of AzaD. In

addition, our data suggested that AzaD treatment interferes with GCN2 phosphorylation during the acute

phase. Because of the reciprocal regulation between GCN2 and mTOR (

21, 22

), we further investigated the

phosphorylation status of mTOR in GCN2-KO cells treated with AzaD to identify the target of AzaD. The data

indicated that the lack of GCN2 did not affect the mTOR phosphorylation (mTOR-p) ratio at Ser-2448 or the

phosphorylation of the mTOR substrates S6K1-p at Thr-389 or 4EBP1-p at Thr-37/46 (

), which was mediated

by AzaD treatment (Fig. 5A). We then explored the acute effect of AzaD treatment on changes in the

phosphorylation status of the mTOR protein and mTOR substrates, as these proteins respond rapidly to

changes in nutrient consumption, protein synthesis, and protein degradation

via

autophagy (

), and some of

these pathways are also affected by AzaD activity. Consistently, wildtype HeLa cell lysates treated with 5



AzaD showed increased levels of GCN2-p beginning 15 min after treatment (Supplementary Fig. S11). In the

ORIGINAL

UNEDITED

MANUSCRIPT

Downloaded from https://academic.oup.com/jb/advance-article/doi/10.1093/jb/mvae038/7658884 by UCSD-Philosophy user on 01 May 2024

same lysates, the detection of mTOR phosphorylation at Ser-2448, which reflects mTOR activity, showed that

the mTOR-p ratio gradually decreased, as did the phosphorylation of S6K1-p at Thr-389 and 4EBP1-p at Thr-

37/46 (

) (Fig. 5B-E and Supplementary Fig. S11). mTOR can also phosphorylate ULK1

via

the Ser-757

residue, which inhibits autophagy initiation (

). The data also revealed that the AzaD-treated lysates

suppressed Ser-757 ULK1-p, similar to what was observed for mTOR-p, while an increase in the LC3 ratio

was detected as soon as 15 min and lasted until 120 min (Fig. 5E and Supplementary Fig. S11). Hence, AzaD

treatment causes cytotoxic stress, affecting mTOR phosphorylation and signaling, and the AzaD compound

may directly interfere with the function of mTOR kinase.

We also compared the ability of AzaD to suppress mTOR-p with that of known mTOR inhibitors,

including torin2, torkinib, and rapamycin. The data indicated that compared to DMSO treatment, treatment

with 1 and 5



M AzaD increased the mTOR-p levels in HeLa cells by 0.54- and 0.42-fold, respectively (Fig.

6A). In the same sample set, cells treated with 0.25



M torin2, 0.25



M torkinib, or 0.4



M rapamycin showed

a 0.21-, 0.66-, or 0.37-fold increase in mTOR-p levels, respectively, compared with the control (Fig. 6A-6B).

S6K1-p was undetectable after treatment with torin2 and rapamycin, whereas 1 and 5



M AzaD promoted a

0.97- and 0.61-fold greater levels of S6K1-p, respectively, compared to those in the control group (Fig. 6A and

6C). 4EBP1-p was undetectable only after torin2 treatment, while 5



M AzaD had a moderate effect (0.61-

fold) on the 4EBP1-p level (Fig. 6A and 6D). Therefore, the data suggest that treatment with a high

concentration of AzaD can suppress the phosphorylation of mTOR and mTOR substrates in a manner similar

to that of known mTOR inhibitors, whereas a low concentration of AzaD can interfere with only mTOR-p and

cannot effectively suppress the phosphorylation of mTOR substrates.

Because S6K1 and 4EBP1 are well-known substrates of mTOR complex 1 (mTORC1), we questioned

whether interference by AzaD is specific to a certain form of the mTOR complex. We then investigated the

changes in the phosphorylation of the substrate of mTORC2, Akt. The phosphorylation of Akt at Thr-308 and

Ser-473 in cell lysates treated with 5



M AzaD was detected. The data indicated that Thr-308 and Ser-473

Akt-p were transiently activated 15 min after AzaD challenge, after which the phosphorylation levels abruptly

decreased (Fig. 6E-6G). These data indicate that AzaD may cause nonspecific suppression of mTOR

complexes.

ORIGINAL

UNEDITED

MANUSCRIPT

Downloaded from https://academic.oup.com/jb/advance-article/doi/10.1093/jb/mvae038/7658884 by UCSD-Philosophy user on 01 May 2024

The CETSA and computational analyses revealed that mTOR is a possible molecular target of AzaD.

To verify that mTOR is a target protein of AzaD, we studied the interaction of AzaD with the mTOR protein.

The HeLa cell lysates treated with 10



M AzaD were evaluated by CETSA, and the results showed that the

mTOR protein was thermally stable between 36.7-43.5



C upon AzaD treatment compared with DMSO- and

LamD-treated conditions (Fig. 7A and 7B). The results from three independent experiments showed that the

AzaD-treated mTOR protein was significantly stabilized at 40.9



C (Fig. 7B). Hence, these data imply that

mTOR could also be a direct protein target for AzaD.

To verify this, we performed

in silico

analysis to explore the chemical bonding interactions between

AzaD and mTOR. As previous studies have shown the inhibitory effect of several lamellarins on several

kinases and computationally analyzed AzaD to identify the GSK3



kinase domain (

11, 12

), we hypothesized

that the ATP-binding pocket of mTOR was the target site of AzaD. We performed computational docking

analysis of AzaD and LamD with the mTOR ATP-binding pocket and compared them with known mTOR

inhibitors: torkinib, torin2, and resveratrol (

23, 24

). The results of the analysis indicated that the AzaD

molecule fit well in the pocket site with the most favorable docking score (73.98 Goldscore), followed by

LamD (69.88), torin2 (65.04), resveratrol (56.42), and torkinib (54.85) (Fig. 7C). AzaD was predicted to

interact with the mTOR ATP binding pocket through various interactions, including van der Waals

interactions, hydrogen bonding, and pi-interactions (Fig. 7D and 7E). The phenolic hydrogens of AzaD formed

two hydrogen bonds with Val-2240 and Asp-2357 of the mTOR ATP cleft. Furthermore, another hydrogen

bond was found between the amide hydrogen of AzaD and Asp-2195 on the mTOR structure (Fig. 7D and 7E).

Similarly, LamD binds to mTOR

via

two hydrogen bonds that interact with Ser-2165 and Asp-2195

(Supplementary Fig. S12). Previous studies reported that Asp-2195, Val-2240, and Asp-2357 are critical for

interactions with known mTOR inhibitors (

24-26

). Pi-pi stacking was also found in the complex of AzaD and

mTOR, in which electron clouds of the C5-C6 double bond of AzaD bound to the aromatic ring of Tyr-2225.

Additionally, a pi-pi T-shaped interaction was observed between the aromatic rings of AzaD and Trp-2239. In

the LamD-mTOR complex, stabilization through pi-stacking was also observed through pi orbitals between the

lactone ring of LamD and the aromatic ring of Trp-2239. Interestingly, the hydrogen bond distance between

AzaD and Asp-2195 was shorter than that between LamD and the same amino acid residue (Fig. 7F). In

ORIGINAL

UNEDITED

MANUSCRIPT

Downloaded from https://academic.oup.com/jb/advance-article/doi/10.1093/jb/mvae038/7658884 by UCSD-Philosophy user on 01 May 2024

addition, LamD possesses a hydrogen bond with Asp-2195 with a dihedral angle closer to 180

than AzaD

(Fig. 7F). Comparing the orientation of AzaD and LamD in the pocket site, ligand superimposition revealed

that both AzaD and LamD turned the fused aromatic rings, which are flat conjugated systems, toward the

innermost region of the mTOR pocket site, where AzaD was inserted deeper in the pocket cleft than LamD

(Fig. 7G). Therefore, our data imply that AzaD might interact with the mTOR kinase cleft and have specific

molecular interactions and orientations inside the mTOR ATP binding cleft that are different from those of

LamD due to the lactam moiety.

Discussion

LamD can trigger cellular stress pathways in HeLa cells, leading to the perturbation of mitochondrial

homeostasis and elevated autophagy flux (

). Although how LamD treatment promotes autophagy has not yet

been elucidated, the information on lamellarin-mediated LC3 conversion led to this investigation, in which the

stress could somehow be from the interference of cellular proteostasis. To address the modulation of

proteostasis by lamellarin activity, we detected a change in the phosphorylation of eIF2



at Ser-51, which has

been reported to occur upon exposure to various stress stimuli and inhibits cellular protein translation (

2, 6

Our data presented herein indicated that HeLa cells exhibited elevated levels of phosphorylated eIF2



during

exposure to LamD. However, a more significant accumulation of eIF2



-p was found in the cells treated with

AzaD. The presence of eIF2



-p was also correlated with the cytotoxicity of the lamellarin analogs tested. A

further detailed study indicated the importance of an amide N atom in the lactam ring of the pentacyclic

structure and the 5,6-unsaturated bonding at the D ring of AzaD (Figs. 1 and 2). Several past studies have

shown that the conversion of lactones to lactams in the pentacyclic structure of lamellarins could enhance their

physical properties and cytotoxicity, although the underlying molecular mechanism has not been revealed (

12-

). In addition, some previous reports have suggested that the nonalkylated lactam structure is required for its

cytotoxicity (

9, 13

). Consistently, the results of the present study indicated a correlation between AzaD activity

(LC3 conversion and eIF2



-p accumulation) and the free lactam structure (Figs. 1 and 2). Furthermore, this

study’s data suggest that lactam substitution in the pentacyclic structure also affects biological activities and

cellular responses more than lactone analogs of lamellarins.

ORIGINAL

UNEDITED

MANUSCRIPT

Downloaded from https://academic.oup.com/jb/advance-article/doi/10.1093/jb/mvae038/7658884 by UCSD-Philosophy user on 01 May 2024

Because of the effects of AzaD and LamD on changes in LC3 and eIF2



-p during the early phase of

the treatment, it is speculated that these compounds could interfere with pivotal cellular events that govern

rapid responses to the cellular environment. Lamellarins were previously reported to possess kinase inhibitory

activity against several cellular kinases (

11, 27

). We hypothesized that acute treatment with AzaD or LamD

could alter the activity of kinases that control protein homeostasis, leading to cellular responses in which LC3

conversion and eIF2



phosphorylation are activated. To support this notion, our data revealed that acute AzaD

treatment caused nearly immediate activation of GCN2, a kinase that senses amino acid depletion and

catalyzes eIF2



phosphorylation for ISR signaling (Fig. 3). In the same period, suppression of mTOR and

mTOR substrates, which control cellular protein translation and autophagy, was also observed (Fig. 5).

Subsequently, the AzaD-mediated increase in eIF2



-p was detected as soon as 2 h after treatment (Fig. 3).

Therefore, GCN2 activation and interference with mTOR occur before eIF2



-p accumulation, suggesting that

acute treatment with AzaD affects the activity of kinases that regulate protein homeostasis in cells.

During the extended period of AzaD cytotoxicity, stressed cells exhibit activation of the other eIF2



kinase, PKR, which senses the presence of dsRNA derived from virus infection and damaged mitochondria

during the apoptotic phase under stress (

2, 6

). Therefore, the activation of PKR at 6-8 h during AzaD treatment

could be due to apoptosis, as mitochondria are disrupted beyond the cell survival threshold. In support of this

idea, AzaD treatment can mediate toxicity, disturbing mitochondrial dynamics and potential (

). The

accumulation of eIF2



-p inhibits protein translation but selectively promotes the translation of uORF-

containing mRNAs, including ATF4 and CHOP, to mediate pro-survival or pro-death signaling

via

the ISR (

). In this study, AzaD-treated cells exhibited increased expression of ATF4 without CHOP induction. To

overcome the sensitivity limitation of western blot analysis, we cotreated cells with bortezomib cotreatment to

stabilize ATF4 and CHOP; however, the results confirmed that CHOP in AzaD-treated cells was undetectable

in our experimental setting (Fig. 3E). Interestingly, with similar experimental procedures, the chronically

LamD-treated cells showed a detectable level of CHOP (Supplementary Fig. S8). These data indicate a

discrepancy in cellular responses between AzaD and LamD treatment, implicating different genetic programs

needed to cope with stressful situations mediated by these two compounds.

ORIGINAL

UNEDITED

MANUSCRIPT

Downloaded from https://academic.oup.com/jb/advance-article/doi/10.1093/jb/mvae038/7658884 by UCSD-Philosophy user on 01 May 2024

Past studies have proposed several putative protein targets for lamellarins; however, the interactions

with the target proteins and the cellular effects upon target protein depletion during lamellarin-mediated

toxicity have not been comprehensively reported. Our data herein show that AzaD-treated cells exhibited

activation of the ISR kinases GCN2 and PKR and ISR signaling. However, our data also indicate that neither

GCN2 nor PKR depletion could completely abrogate the activation and accumulation of phosphorylated eIF2



by AzaD treatment, and the signaling of downstream ISR effectors, including ATF4, in these cells was rather

intact. Further analysis by ligand and target engagement (CETSA) between AzaD and GCN2 (or PKR) showed

no significant interaction between AzaD and the ISR kinases. As AzaD has monumental effects on the

inhibition of protein translation, several studies have indicated reciprocal control between GCN2 and mTOR,

another pivotal kinase for protein synthesis (

). We then investigated mTOR phosphorylation and

signaling, and the data clearly showed that AzaD toxicity could also affect mTOR phosphorylation and

signaling (Figs. 5 and 6). To determine whether mTOR is a putative molecular target of AzaD, we evaluated

the interaction between AzaD and mTOR using CETSA, and the results indicated that AzaD could protect

mTOR from heat denaturation (Fig. 7A and 7B).

Subsequent

in silico

analysis also revealed that AzaD could occupy the active site pocket of mTOR

with a favorable Goldscore compared with LamD and known mTOR inhibitors

Our predicted model

suggested that AzaD was positioned in the mTOR binding pocket to favor H bond formation between the

hydrogen of the amide N atom of AzaD and Asp-2195, two additional hydrogen bonds, and two pi-pi

interactions. However, LamD did not have an ester moiety involved in any H-bond interactions but rather

positioned the pi electron clouds of the lactone ring to form pi-pi stacking and arranged the phenolic groups to

form two H-bonds with the amino acid residues of mTOR. This highlights an essential aspect of the lactam

ring in AzaD. Moreover, the lactam ring of the azalamellarins is positioned toward the innermost area of the

mTOR ATP cleft. Thus, the presence of bulky groups may prevent the azalamellarin molecule from orienting

into the mTOR ATP cleft, in accordance with the lack of translational inhibition activity of AzaD-P, AzaD-C3,

and AzaD-C4 (Fig. 2E). Tyr-2225 has been identified as the critical residue that facilitates the expansion of the

inner hydrophobic pocket, as observed by the binding mode of torkinib and mTOR (

). In the present study,

the pi-pi interactions between the pi electron clouds of the C5-C6 double bond in the AzaD lactam ring and the

ORIGINAL

UNEDITED

MANUSCRIPT

Downloaded from https://academic.oup.com/jb/advance-article/doi/10.1093/jb/mvae038/7658884 by UCSD-Philosophy user on 01 May 2024

aromatic ring of Tyr-2225 were observed. The discrepancy in the strength of the activities of AzaD and

AzaChi may be explained by the differences in the type of chemical bonds between C5 and C6 of

azalamellarins (Fig. 2). Furthermore, both AzaD and LamD aligned the polycyclic aromatic group toward the

inner ATP cleft, in which AzaD was positioned deeper in the mTOR pocket than LamD, indicating a stronger

interaction with the key amino acid residues in the pocket site (Fig. 7G).

The data presented in this study suggest that mTOR may be a putative target for AzaD, and that the

interaction between AzaD and mTOR may interfere with mTOR activity, causing altered cellular proteostasis

and ISR activation. Overall, this investigation demonstrated the novel biological activity of AzaD in cellular

pathways that control protein homeostasis. The information gained from this study is essential for developing

this compound into a novel pharmacophore for anticancer therapeutics.

Funding

This work was supported by the Chulabhorn Graduate Institute research grant (611-AB03), Thailand Research

Fund (MRG6108191), Thailand Science Research and Innovation (TSRI)-Chulabhorn Graduate Institute,

Chulabhorn Royal Academy (FFB640035 Project code 50187, FRB650039/0240 Project code 165423,

FRB660044/0240 Project code 180860, and FRB660044/0240 Project code 180874 to PS and PP). This study

was also supported by TSRI-Chulabhorn Research Institute (Grant No. 36824/4274394 and 36827/4274409) to

PP and RS.

Author contributions

TM, BC, NP, RS, and PS performed the formal analysis and investigation. KK and PP conducted chemical

synthesis and verification. PG, SR, and PP provided resources for the experiments. RS generated molecular

docking data, which were interpreted and analyzed by RS, PP, and PS. PP and PS played a role in the

conceptualization, methodology, and funding acquisition. SR, PG, and PP provided supervision related to the

experimental procedures, chemical synthesis, and manuscript writing. The original draft was prepared with

TM, PP, and PS. The manuscript was reviewed and edited by TM, SR, PG, RS, PP, and PS. All authors have

read and approved the final version of the manuscript.

ORIGINAL

UNEDITED

MANUSCRIPT

Downloaded from https://academic.oup.com/jb/advance-article/doi/10.1093/jb/mvae038/7658884 by UCSD-Philosophy user on 01 May 2024

Availability of data and materials

All relevant data and materials in this study are available from the authors upon reasonable request.

Conflict of Interests

The authors declare no conflicting interests associated with this study.

Supplementary data

Supplementary data are available.

Acknowledgments

PS would like to thank the following researchers from the Chulabhorn Research Institute and the Chulabhorn

Graduate Institute; Dr. Piyajit Watcharasit (Laboratory of Pharmacology); Dr. Pattama Singhirunnusorn, Dr.

Benchamart Khumkrong (Laboratory of Chemical Carcinogenesis); Dr. Nopporn Thasana (Laboratory of

Medicinal Chemistry); Dr. Montri Yasawong (Program in Environmental Toxicology); and Dr. Satapanawat

Sittihan and Dr. Watthanachai Jumpathong (Program in Chemical Sciences) for kindly sharing equipment,

chemicals, and suggestions for this manuscript. PS and RS would like to thank Dr. Chatchakorn Eurtivong

(Faculty of Pharmacy, Mahidol University) for the software sharing.

This research work was supported in part by a grant from the Center of Excellence on Environmental

Health and Toxicology (EHT), OPS, Ministry of Higher Education, Science, Research and Innovation.

References

(1)

Saxton, R.A., and Sabatini, D.M. (2017) mTOR Signaling in Growth, Metabolism, and Disease.

Cell

168

960-976

(2)

Houston, R., Sekine, S., and Sekine, Y. (2020) The coupling of translational control and stress responses.

Biochem

168

, 93-102

(3)

Towers, C.G., Wodetzki, D., and Thorburn, A. (2020) Autophagy and cancer: Modulation of cell death

pathways and cancer cell adaptations.

The Journal of cell biology

219

(4)

Koromilas, A.E. (2019) M(en)TORship lessons on life and death by the integrated stress response.

Biochim

Biophys Acta Gen Subj

1863

, 644-649

ORIGINAL

UNEDITED

MANUSCRIPT

Downloaded from https://academic.oup.com/jb/advance-article/doi/10.1093/jb/mvae038/7658884 by UCSD-Philosophy user on 01 May 2024

(5)

Sasaki, K., and Yoshida, H. (2015) Organelle autoregulation-stress responses in the ER, Golgi,

mitochondria and lysosome.

J Biochem

157

, 185-195

(6)

Pakos-Zebrucka, K., Koryga, I., Mnich, K., Ljujic, M., Samali, A., and Gorman, A.M. (2016) The

integrated stress response.

EMBO Rep

, 1374-1395

(7)

Kim, J., Kundu, M., Viollet, B., and Guan, K.L. (2011) AMPK and mTOR regulate autophagy through

direct phosphorylation of Ulk1.

Nature cell biology

, 132-141

(8)

Chittchang, M., and Theppawong, A. (2019) Chapter 12 - An Overview of the Multifaceted Lessons

Learned from Marine-Derived Bioactive Lamellarin Natural Products. in

Studies in Natural Products

Chemistry

(Atta ur, R., ed.)^eds.), pp. 411-460, Elsevier,

(9)

Fukuda, T., Ishibashi, F., and Iwao, M. (2020) Lamellarin alkaloids: Isolation, synthesis, and biological

activity.

Alkaloids Chem Biol

, 1-112

(10)

Sopha, P., Phutubtim, N., Chantrathonkul, B., Ploypradith, P., Ruchirawat, S., and Chittchang, M. (2021)

Roles of autophagy in relation to mitochondrial stress responses of HeLa cells to lamellarin cytotoxicity.

Toxicology

462

, 152963

(11)

Baunbaek, D., Trinkler, N., Ferandin, Y., Lozach, O., Ploypradith, P., Rucirawat, S., Ishibashi, F., Iwao,

M., and Meijer, L. (2008) Anticancer alkaloid lamellarins inhibit protein kinases.

Mar Drugs

, 514-527

(12)

Theppawong, A., Ploypradith, P., Chuawong, P., Ruchirawat, S., and Chittchang, M. (2015) Facile and

Divergent Synthesis of Lamellarins and Lactam-Containing Derivatives with Improved Drug Likeness and

Biological Activities.

Chem Asian J

, 2631-2650

(13)

Boonya-Udtayan, S., Yotapan, N., Woo, C., Bruns, C.J., Ruchirawat, S., and Thasana, N. (2010) Synthesis

and biological activities of azalamellarins.

Chem Asian J

, 2113-2123

(14)

Klumthong, K., Chalermsub, P., Sopha, P., Ruchirawat, S., and Ploypradith, P. (2021) An Expeditious

Modular Hybrid Strategy for the Diversity-Oriented Synthesis of Lamellarins/Azalamellarins with

Anticancer Cytotoxicity.

J Org Chem

, 14883-14902

(15)

Sopha, P., Ren, H.Y., Grove, D.E., and Cyr, D.M. (2017) Endoplasmic reticulum stress-induced

degradation of DNAJB12 stimulates BOK accumulation and primes cancer cells for apoptosis.

The Journal

of biological chemistry

292

, 11792-11803

(16)

Martinez Molina, D., Jafari, R., Ignatushchenko, M., Seki, T., Larsson, E.A., Dan, C., Sreekumar, L., Cao,

Y., and Nordlund, P. (2013) Monitoring drug target engagement in cells and tissues using the cellular

thermal shift assay.

Science (New York, N.Y.)

341

, 84-87

(17)

Verdonk, M.L., Cole, J.C., Hartshorn, M.J., Murray, C.W., and Taylor, R.D. (2003) Improved protein-

ligand docking using GOLD.

Proteins

, 609-623

(18)

Carroll, B., Korolchuk, V.I., and Sarkar, S. (2015) Amino acids and autophagy: cross-talk and co-operation

to control cellular homeostasis.

Amino Acids

, 2065-2088

(19)

Zhao, J., Zhai, B., Gygi, S.P., and Goldberg, A.L. (2015) mTOR inhibition activates overall protein

degradation by the ubiquitin proteasome system as well as by autophagy.

Proceedings of the National

Academy of Sciences of the United States of America

112

, 15790-15797

ORIGINAL

UNEDITED

MANUSCRIPT

Downloaded from https://academic.oup.com/jb/advance-article/doi/10.1093/jb/mvae038/7658884 by UCSD-Philosophy user on 01 May 2024

(20)

Topf, U., Uszczynska-Ratajczak, B., and Chacinska, A. (2019) Mitochondrial stress-dependent regulation

of cellular protein synthesis.

Journal of cell science

132

, jcs226258

(21)

Ye, J., Palm, W., Peng, M., King, B., Lindsten, T., Li, M.O., Koumenis, C., and Thompson, C.B. (2015)

GCN2 sustains mTORC1 suppression upon amino acid deprivation by inducing Sestrin2.

Genes &

development

, 2331-2336

(22)

Wengrod, J., Wang, D., Weiss, S., Zhong, H., Osman, I., and Gardner, L.B. (2015) Phosphorylation of

eIF2α triggered by mTORC1 inhibition and PP6C activation is required for autophagy and is aberrant in

PP6C-mutated melanoma.

Sci Signal

, ra27

(23)

Park, D., Jeong, H., Lee, M.N., Koh, A., Kwon, O., Yang, Y.R., Noh, J., Suh, P.G., Park, H., and Ryu, S.H.

(2016) Resveratrol induces autophagy by directly inhibiting mTOR through ATP competition.

Sci Rep

21772

(24)

Yang, H., Rudge, D.G., Koos, J.D., Vaidialingam, B., Yang, H.J., and Pavletich, N.P. (2013) mTOR kinase

structure, mechanism and regulation.

Nature

497

, 217-223

(25)

Parate, S., Kumar, V., Lee, G., Rampogu, S., Hong, J.C., and Lee, K.W. (2021) Marine-Derived Natural

Products as ATP-Competitive mTOR Kinase Inhibitors for Cancer Therapeutics.

Pharmaceuticals (Basel)

, 282

(26)

Xie, C., Chen, X., Zheng, M., Liu, X., Wang, H., and Lou, L. (2017) Pharmacologic characterization of

SHR8443, a novel dual inhibitor of phosphatidylinositol 3-kinase and mammalian target of rapamycin.

Oncotarget

, 107977-107990

(27)

Bailly, C. (2015) Anticancer properties of lamellarins.

Mar Drugs

, 1105-1123

ORIGINAL

UNEDITED

MANUSCRIPT

Downloaded from https://academic.oup.com/jb/advance-article/doi/10.1093/jb/mvae038/7658884 by UCSD-Philosophy user on 01 May 2024

Figure captions

Fig. 1. Azalamellarin D (AzaD) is cytotoxic and induces changes in proteins involved in apoptosis,

autophagy, and protein translation.

(

A) Structure of azalamellarin D (AzaD), a synthetic lamellarin containing a double bond (purple highlight) in

the D-ring adjacent to the lactam ring (N-atom highlighted in the blue circle). The derivatives of AzaD in this

study included LamD (B) and LamChi (C), which are natural products and consist of lactone moieties (yellow

circles) with a double (purple highlight) or single bond (gray highlight) in the D-ring, respectively. (D) HeLa

cell viability after 18 h of treatment with titrated concentrations (1, 2, and 5



M) of AzaD, LamD, and

LamChi. Statistical significance was analyzed by two-way ANOVA; a

value less than 0.05 indicated

statistical significance. (E) – (H) The effects of LamD, AzaD, and LamChi on alterations in apoptotic proteins

(cCasp-3 and cPARP), the autophagy marker LC3, and phosphorylation at serine 51 (Ser-51) of eIF2



, an

indication of global protein translational attenuation. HeLa cells were cultured for at least 48 h before 5



LamD treatment for 6 h. The treated cells were collected and subjected to SDS‒PAGE and western blot

analysis. The steady-state levels of the indicated proteins were determined, recorded, and analyzed for

densitometric data for cCasp-3 (E), cPARP (F), the LC3-II ratio (G), and the eIF2



-p ratio (H). The data from

the four independent experiments were analyzed by two-way ANOVA, in which a

value less than 0.05 was

considered to indicate statistical significance.

Fig. 2. The lactam ring of AzaD revealed its biological activities and cytotoxicity.

(A) Structure of azalamellarin



(AzaChi), which shows a single bond (gray highlight) at the D ring adjacent to

the lactone moiety (the N-atom is highlighted in the blue circle). (B) Effect of AzaChi on the eIF2



-p ratio

over an 8-h duration. Quantitative data analysis for the band density of the eIF2



-p ratio analyzed by western

blotting using HeLa lysates treated with 5



M AzaD compared with those treated with AzaChi. The data are

presented as the mean



SD from four individual experiments. A

value less than 0.05 was considered to

indicate statistical significance according to two-way ANOVA. (C) The impact of AzaD and its derivatives

with a protected N atom on HeLa cell viability was determined by MTT assay. The data are presented as the

mean



SD from two different individual experiments. A

value less than 0.05 was considered to indicate

ORIGINAL

UNEDITED

MANUSCRIPT

Downloaded from https://academic.oup.com/jb/advance-article/doi/10.1093/jb/mvae038/7658884 by UCSD-Philosophy user on 01 May 2024

statistical significance according to two-way ANOVA. (D) The general structure of AzaD derivatives

containing a protected N atom. The functional groups for the

-group (pink circle) at the N-atom of the lactam

ring are indicated. (E) and (F) The effects of AzaD, AzaD-P, AzaD-C3, and AzaD-C4 on the phosphorylation

of eIF2



(E) and the induction of the autophagic protein LC3-II ratio (F).

Fig. 3. AzaD-mediated cytotoxicity in HeLa cells caused sequential activation of the ISR kinases GCN2

and PKR, which control cellular protein translation.

(A) Analysis of the intracellular level of eIF2



-p in viable HeLa cells by flow cytometry. The data are

presented as the mean



standard deviation (SD) of three independent experiments (n = 6). Each color

represents an independent experiment, a triangle represents the mean, and a circle represents the raw data from

the mean. Statistical analysis was determined by two-way ANOVA at a

value ≤ 0.05, which was considered

to indicate significance. (B) – (D) The phosphorylation of PKR, GCN2, and eIF2



during the 8-h time course.

The image shows representative western blot data for seven independent experiments of HeLa cell lysates

treated with 5



M AzaD for 2-h intervals. The densitometry data from multiple independent experiments for

changes in the eIF2



-p, PKR-p, and GCN2-p ratios are plotted in (B), (C), and (D), respectively. (E)

Immunoblot analysis of the downstream ISR effectors ATF4 and CHOP in HeLa cells treated with AzaD.

Cotreatment with the proteasome inhibitor bortezomib (Bort) stabilized CHOP protein levels. The image

shows the representative results of two different individual experiments.

Fig. 4. CETSA revealed that cells with depletion of GCN2 or PKR exhibited ISR activation upon AzaD

treatment and that neither GCN2 nor PKR could engage with AzaD.

(A) Impact of AzaD-mediated ISR induction in GCN2-KO cells compared with that in the corresponding

parental HeLa cells. The image shows the representative results of three different individual experiments. (B)

Effect of AzaD on ISR activation in PKR-KO cells compared with that in parental HeLa cells. The image

shows representative results for three different individual experiments. (C) Analysis of the interaction between

AzaD and GCN2 by cellular thermal shift assay (CETSA) in intact HeLa cells. The density data from three

independent experiments are plotted. (D) CETSA of AzaD and PKR interactions in intact HeLa cells.

ORIGINAL

UNEDITED

MANUSCRIPT

Downloaded from https://academic.oup.com/jb/advance-article/doi/10.1093/jb/mvae038/7658884 by UCSD-Philosophy user on 01 May 2024

Fig. 5. Acute AzaD treatment also targeted mTOR kinase, which controls cellular nutrient uptake,

metabolism, and translation in HeLa cells.

(A) Analysis of the phosphorylation of mTOR and mTOR substrates in HeLa GCN2-KO cells. (B) – (E) Effect

of 5



M AzaD on the phosphorylation of mTOR and mTOR substrates (S6K1, 4EBP1, and ULK1) in wildtype

HeLa cells.

Fig. 6. AzaD treatment suppressed signaling mediated through mTORC1 and mTORC2 in HeLa cells.

(A) Effects on mTORC1 signaling suppression mediated by the mTOR inhibitor and AzaD treatments. HeLa

cell lysates treated with torin2, torkinib, rapamycin, and AzaD were blotted and detected with the indicated

antibodies. The mTOR-p ratio and S6K1-p and 4EBP1-p from four individual experiments were recorded and

plotted in (B), (C), and (D), respectively. (E) The impact of AzaD treatment on the phosphorylation of Akt, the

substrate of mTORC2, in HeLa cells. Quantitative data analyses for the Thr-308 and Ser-473 Akt-p ratios are

plotted in (F) and (G). The data are presented as the mean



SD from multiple different individual

experiments. A

value less than 0.05 was considered to indicate statistical significance according to two-way

ANOVA.

Fig. 7. Biochemical and computational analyses revealed a possible interaction between AzaD and

mTOR.

(A) CETSA of mTOR engagement by AzaD and LamD in intact HeLa cells. (B) Densitometric analysis of (A).

* indicates statistical significance at a

value less than 0.05. (C) Docking scores of AzaD, LamD, and known

mTOR kinase inhibitors with the mTOR kinase domain (PDB ID: 4JT5). (D) Computational analyses of the

predicted binding interactions between AzaD and mTOR. (E) The 2D binding mode in the mTOR kinase

domain is displayed for the AzaD interaction. The dotted lines and colors in the 2D pose interactions represent

different intermolecular interactions. (F) The distance and dihedral angle (DHA) of H-bonds between key

amino acid residues of the mTOR binding pocket (PDB ID: 4JT5) and AzaD (or LamD). (G) Superimposition

of AzaD (green) and LamD (yellow) within the ATP-binding pocket of the mTOR kinase domain.

ORIGINAL

UNEDITED

MANUSCRIPT

Downloaded from https://academic.oup.com/jb/advance-article/doi/10.1093/jb/mvae038/7658884 by UCSD-Philosophy user on 01 May 2024