Divergent Proteome Reactivity Influences Arm-Selective Activation of Pharmacological Endoplasmic Reticulum Proteostasis Regulators

SUMMARY Pharmacological activation of the activating transcription factor 6 (ATF6) arm of the Unfolded Protein Response (UPR) has proven useful for ameliorating proteostasis deficiencies in a variety of etiologically diverse diseases. Previous high-throughput screening efforts identified the small molecule AA147 as a potent and selective ATF6 activating compound that operates through a mechanism involving metabolic activation of its 2-amino-p-cresol substructure affording a quinone methide, which then covalently modifies a subset of ER protein disulfide isomerases (PDIs). Intriguingly, another compound identified in this screen, AA132, also contains a 2-amino-p-cresol moiety; however, this compound showed less transcriptional selectivity, instead globally activating all three arms of the UPR. Here, we show that AA132 activates global UPR signaling through a mechanism analogous to that of AA147, involving metabolic activation and covalent PDI modification. Chemoproteomic-enabled analyses show that AA132 covalently modifies PDIs to a greater extent than AA147. Paradoxically, activated AA132 reacts slower with PDIs, indicating it is less reactive than activated AA147. This suggests that the higher labeling of PDIs observed with activated AA132 can be attributed to its lower reactivity, which allows this activated compound to persist longer in the cellular environment prior to quenching by endogenous nucleophiles. Collectively, these results suggest that AA132 globally activates the UPR through increased engagement of ER PDIs. Consistent with this, reducing the cellular concentration of AA132 decreases PDI modifications and allows for selective ATF6 activation. Our results highlight the relationship between metabolically activatable-electrophile stability, ER proteome reactivity, and the transcriptional response observed with the enaminone chemotype of ER proteostasis regulators, enabling continued development of next-generation ATF6 activating compounds.


INTRODUCTION
The unfolded protein response (UPR) is an endoplasmic reticulum (ER) stress-responsive signaling pathway that corrects imbalances in ER protein homeostasis (proteostasis) caused by an ER stress. [1][2][3] The UPR functions by simultaneously expanding ER folding capacity and restricting ER protein flux to restore ER proteostasis through both transcriptional and non-transcriptional responses. 1,3,4 Activation of the UPR occurs downstream of three resident ER stress sensors: PKR-like Endoplasmic Reticulum Kinase (PERK), Inositol-requiring enzyme 1 (IRE1), and Activating Transcription Factor 6 (ATF6). Transcription factors regulated downstream of ATF6 and IRE1, ATF6f and XBP1s respectively, induce expression of numerous protective genes that remodel biological pathways involved in cellular metabolism, redox regulation, and ER proteostasis. 1,3,4 This regulation functions to alleviate the ER stress and adapt cellular physiology to pathologic ER insults. While persistent activation of the IRE1 and PERK arms of the UPR have been associated with pathologic consequences, constitutive ATF6 activation to physiologically relevant levels has not generally been found to be detrimental in mammalian cell culture or mouse models. [5][6][7][8][9][10] As such, genetic and pharmacological activation of ATF6 signaling has proven beneficial in mitigating pathological conditions resulting from proteostasis imbalances in numerous disease models. [11][12][13][14][15][16] This suggests that ATF6 is an attractive therapeutic target to intervene in etiologically diverse diseases. 4,5,17,18 We previously conducted a cell-based high-throughput screen to identify selective activators of ATF6 signaling. 19 From this screen, N-(2-hydroxy-5-methylphenyl)-3-phenylpropanamide (AA147; Fig   1A) emerged as a selective ATF6 activator that is able to promote adaptive ATF6 activity to mitigate pathologies associated with etiologically diverse disorders. 11,[19][20][21][22][23][24][25][26] We found that AA147 functions as a prodrug, wherein oxidative conversion of the 2-amino-p-cresol substructure by ER-resident oxidases (e.g., cytochrome P450s) leads to an electrophilic quinone methide and selective covalent engagement of a subset of ER proteins primarily consisting of protein disulfide isomerases (PDIs). 27 PDIs maintain ATF6 in disulfide-bonded structures that limit its activation. [28][29][30][31] This suggested that AA147-dependent modification of a subset of PDIs could lead to reduction, monomerization, and subsequent trafficking

AA132 activates UPR signaling through a mechanism involving metabolic activation and covalent protein modification
We previously reported that AA147 selectively activates ATF6 signaling through a process involving metabolic activation of its 2-amino-p-cresol substructure affording a reactive quinone methide that covalently modifies ER-localized PDIs. 27 AA132 also contains a 2-amino-p-cresol moiety, suggesting that this compound could activate UPR signaling through an analogous mechanism (Fig. 1A,B).
Next, we sought to determine the sensitivity of AA132-mediated UPR activation to cotreatment with resveratrol, which blocks metabolic activation, or β-mercaptoethanol (BME), which reacts with the AA132-derived quinone methide prior to protein labeling (Fig 1B). We initially confirmed that AA132 activated luciferase-based reporters of ATF6 signaling (ERSE-FLuc) 19 , IRE1 signaling (XBP1-RLuc) 17 , and PERK signaling (ATF4-FLuc), while AA147 only significantly activated the ATF6-selective ERSE-FLuc reporter (Fig. S1A-C). This confirms previous results showing that AA132 activates all three signaling arms of the UPR. 19 Cotreatment with either resveratrol or BME inhibited AA132-dependent activation of all three UPR reporters (Fig. 1D,E, Fig. S1D-G). These results support a model wherein AA132 activates global UPR signaling through a mechanism involving AA132 ER metabolic activation generating a quinone methide followed by covalent protein modification (Fig. 1B) -a mechanism strictly analogous to that observed for selective AA147-dependent ATF6 activation. 27
We next treated HEK293T cells with AA132 yne for 4h and visualized cellular protein labeling by CuAAC conjugation to a fluorescent azide-cyanine tag followed by SDS-PAGE and in-gel fluorescence scanning (Fig 2D). 34 Cotreatment with four-fold excess AA147 or AA132 reduces protein labeling by AA132 yne , indicating that these compounds target similar subsets of the cellular proteome ( Fig 2D). As observed with AA147 yne (Fig. 1C), excess AA132 showed stronger competition with AA132 yne labeling than excess AA147. These results establish AA132 yne as an efficient probe to monitor protein conjugation afforded by oxidation of AA132 to the putative quinone methide reactive species. AA132 yne in HEK293T cells (Table S2). Despite labeling similar proteins (Fig. 4A), we found that AA132 yne showed higher overall labeling of multiple targets, as compared to AA147 yne at the same concentration (10 µM) (Fig. 4B). This includes greater labeling for numerous PDIs linked to ATF6 activation, including PDIA1, PDIA3, PDIA4, PDIA6, and TXNDC5 (Fig. 4A,B). Similar results were observed in other cells types, including liver-derived HepG2 cells (Fig. S4A,B, Table S2). We confirmed the increased labeling of PDIA3 and PDIA4 in HEK293T cells by biotin conjugation, streptavidin enrichment, and quantitative immunoblotting (Fig. S4C). These results suggest that increased modification of PDIs could define the differential transcriptional signaling observed between AA147 and AA132.

AA132 yne shows slower labeling kinetics as compared to AA147 yne
Given the relative structural similarity between AA147 and AA132, we sought to identify the basis for the divergence in proteome labeling and subsequent transcriptional selectivity. We initially predicted that the small structural differences between AA147 and AA132 could lead to changes in the stability of target proteins upon small molecule conjugation. However, we did not see significant differences in resistance to heat denaturation for PDIA1 labeled with AA147 or AA132 (Fig. S5A,B).
We next tested whether differences in the kinetics of conjugate formation by AA147 or AA132 could mediate their differential proteome labeling. We performed an in-gel fluorescence time course experiment in HEK293T cells monitoring AA147 yne or AA132 yne conjugate formation. Intriguingly, despite AA132 yne showing higher overall PDI labeling after 6h of treatment (Fig. 4A,B), AA132 yne showed slower kinetics of proteome labeling, as compared to AA147 yne (Fig. 5A-C). The relative extent of PDI labeling by AA132 and AA147 becomes equal after four hours, suggesting that AA132 has a longer operational half live in cellulo (Fig. 5C). Similar results were observed in other cell types, including ALMC2 plasma cells (Fig. S5C,D). This suggests that AA132 yne labels proteins at a slower rate than AA147 yne . If correct, this effect is predicted to make AA132 yne protein labeling more susceptible to inhibition afforded by exogenous nucleophiles such as BME. Consistent with this, cotreatment with BME showed greater inhibition of AA132 yne labeling, as compared to AA147 yne labeling ( Fig. S5E,F). Together, these results indicate that AA132 yne forms protein conjugates slower, as compared to AA147 yne , despite ultimately reaching higher levels. This suggests that AA132 yne forms a more stable, longer-lived electrophilic species in its mechanism of protein conjugation.

Dose-dependent regulation of AA132 transcriptional selectivity.
AA132 and AA147 induce distinct transcriptional profiles in HEK293T cells after 6h treatment. AA147 selectively activates the ATF6 arm of the UPR, while AA132 activates all three arms of the UPR. 19 Our results indicate that the distinct transcriptional profiles induced by these two compounds could be attributed to differences in PDI labeling, with AA132 modifying PDIs to a greater extent than AA147 (Fig. 4B). This would suggest that decreasing AA132-dependent PDI modification should result in increased transcriptional selectivity for ATF6 activation. To test this, we monitored mRNA expression by RNAseq in HEK293T cells treated with increasing concentrations of AA132 from 0.1 -30 µM for 6 h (Table S3). We confirmed dose-dependent protein modification in HEK293T cells treated with increasing concentrations of AA132 (Fig. S6A). As expected, the number of differentially expressed genes (DEGs) increased with increasing AA132 concentrations (Fig. 6A). Gene ontology (GO) analysis of significantly induced genes in AA132-treated cells showed selective increases in terms associated with ER function, ER stress, and the UPR (Fig. 6B,C, Fig. S6B-D, Table S4). These results confirm the genome wide transcriptional specificity of AA132 for UPR activation reported previously. 19 Next, we sought to define the transcriptional selectivity of increasing concentrations of AA132 for activation of the three arms of the UPR. Monitoring expression of select target genes of ATF6 (HSPA5, PDIA4, PDIA6, MANF), IRE1/XBP1s (DNAJB9, SSR3, SEC61A, DERL2), and PERK/ATF4 (DDIT3, PPP1R15A, ASNS, TRIB3) shows dose-dependent increases of these genes ( Fig. 6D-F).
Interestingly, the EC50 for ATF6-target genes (3.2 µM) is less than that observed for IRE1/XBP1s and . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted January 17, 2023. ; https://doi.org/10.1101/2023.01.16.524237 doi: bioRxiv preprint PERK/ATF4 target genes (8.8 and 8.9 µM, respectively). This suggests that ATF6 signaling is induced at lower concentrations of AA132, as compared to IRE1/XBP1s and PERK/ATF4. Further, normalizing the expression of 15-20 target genes regulated by ATF6, IRE1/XBP1s, or PERK/ATF4 observed following 6h treatment with AA132 to that observed with the global ER stressor thapsigargin (Tg) shows that ATF6 target genes are globally and significantly induced starting at 3 µM, while IRE1/XBP1s and PERK/ATF4 target genesets are significantly induced starting at 10 and 15 µM, respectively ( Fig. 6G- Table S5). 37 Interestingly, the activation of these genesets observed at 3 µM AA132 match the selective ATF6 activation observed in HEK293T cells treated with AA147 (10 µM, 6 h; Fig. 6G-I, right).
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

DISCUSSION
Here, we sought to define the molecular basis for the differential transcriptional selectivity of AA147 and AA132 with regards to arm-selective UPR activation. Like AA147, we show that AA132 achieves pharmacologic UPR modulation through a mechanism involving metabolic oxidation of its 2-amino-pcresol moiety yielding a putative quinone methide that subsequently covalently modifies ER-localized proteins, most notably multiple PDIs. 27 Despite sharing numerous protein targets, comparative chemoproteomic experiments revealed that AA132 labels ER proteins, including PDIs, to a greater extent than AA147 at treatment times exceeding four hours. This divergence in proteomic reactivity provides a mechanism to explain the differential selectivity for UPR activation observed upon AA132 or AA147 treatment, with AA132 activating all three arms of the UPR and AA147 selectively activating ATF6 UPR signaling. 19 Our results indicate that this greater PDI labeling is attributed to increased stability of the metabolically activated AA132 compound, increasing its diffusion within the cell and subsequent accessibility to a larger pool of ER proteins. The increased labeling of ER proteins afforded by AA132 provides a mechanism to explain the global activation of UPR signaling pathways observed with this compound, as higher labeling would be predicted to further disrupt PDI activity and subsequently increase UPR signaling, possibly by producing misfolded secretory proteins. In contrast, the more modest PDI inhibition induced by AA147 would only be sufficient to promote ATF6 reduction, monomerization, and trafficking, leading to selective activation of ATF6 transcriptional activity.
Consistent with this, lower concentrations of AA132 decrease PDI modifications and lead to selective ATF6 activation, mimicking the transcriptional selectivity observed with AA147. Together, these results highlight the connection between electrophilic reactivity, apparent quinone methide half-life, and transcriptional selectivity for pharmacologic ATF6 activators, establishing new opportunities to develop next generation compounds of this class with improved activity and selectivity.
While we and others have demonstrated the importance of PDI family members in dictating ATF6 activity, the PDIs also have critical roles in regulating the IRE1 and PERK arms of the UPR.
PDIA6 interacts with luminal domains of both PERK and IRE1 to prevent hyperactivation of these UPR . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made  [28][29][30][31] Previous results indicate that AA147 activates ATF6 signaling through selective modification of only a subset of specific ER PDIs. For example, AA147 was shown to modify only ~20% of PDIA1 within the ER. 27 This suggests that the ability for AA147 to selectively activate ATF6 signaling lies in its unique capacity to modify a small population of specific PDIs. Consistent with this, genetic depletion of PDIs including PDIA1, PDIA3, PDIA4, and PDIA5 limit AA147-dependent ATF6 activation. 27 Our results suggest that the increased modification of PDIs afforded by AA132 underlies its ability to globally activate UPR signaling. We show that AA132 modifies multiple PDIs to a greater extent than AA147, including PDIA1 and PDIA6. Interestingly, highly selective covalent PDIA1 inhibitors that modify PDIA1 to greater extents than AA147 can activate IRE1/XBP1s signaling. 40,41 This suggests that increased targeting of PDIA1 by AA132 may contribute to the IRE1/XBP1s activation observed under these conditions. Similarly, increased PDIA6 modification could mitigate the repression of IRE1 and PERK hyperactivity afforded by this PDI 38 , increasing activity of these two UPR signaling pathways. To support this model, we show that decreasing AA132-dependent modifications of PDIs, including of PDIA1 and PDIA6, decreases activation of IRE1 and PERK transcriptional signaling and allows selective ATF6 transcriptional activity. This suggests that the transcriptional activity of AA132 for global UPR activity can be viewed as a graded disruption in the tightly controlled regulatory roles of ER PDIs involved in regulating ER proteostasis and UPR signaling. While we cannot rule out different protein targets contributing to the activation of different arms of the UPR, our data are consistent with a more pronounced on-target mechanism involving increased PDI modification being responsible for the global UPR activating capacity of AA132.
AA132 and AA147 have close structural similarity, with both compounds containing the 2-aminop-cresol substructure critical for compound activity. However, there exists a profound difference in both . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted January 17, 2023. ; https://doi.org/10.1101/2023.01.16.524237 doi: bioRxiv preprint transcriptional activity and equilibrium proteome labeling between AA147 and AA132. A possible modulatory role of the linker region on the activity of the metabolically derived electrophile helps reconcile these opposing results. AA132 contains an enaminone linkage, while AA147 contains an amide linkage. Increased electrophile stability or decreased rate of metabolic activation mediated by the amide to enaminone substitution could explain the observed divergence in biological effects. For instance, activation of AA132 to the requisite electrophilic species may require additional isomerization steps, possibly involving the enaminone moiety, in an alternate metabolic activation pathway. 42,43 The rate of tautomerization of quinone imine/quinone methide moieties, by nonenzymatic or enzymatic means, may be modulated by the linker region. 44 In addition, π-conjugation present in the enaminone class of metabolically activatable proteostasis regulators may stabilize the resulting electrophilic species, and thus enable a larger sphere of activity. 45 Importantly, the PDIs are not amongst the most reactive ER-resident cysteines, and thus increased labeling is not likely to be a simple function of electrophile reactivity towards hyperactive cysteines. 46 Nevertheless, other soft electrophilic species similar to those derived from AA147 or AA132, such as quinone-based electrophiles, have preferential reactivity towards the PDIs. [47][48][49][50] Our kinetic analyses observed in context with the greater reduction in AA132 yne conjugation seen with BME cotreatment suggest that increased stability, rather than increased reactivity, may underlie our observation of greater PDI engagement with the enaminone class of ER proteostasis regulators. Finally, the identification of unique ER AA132 yne targets not seen with AA147 yne could be a result of a larger diffusion radius from site of AA132 yne electrophilic metabolite generation. Alternatively, subtle reactivity differences between AA132 yne -and AA147 yne -derived electrophiles may differentially facilitate conjugation with previously annotated reactive cysteines on some of these targets. [51][52][53] Our chemoproteomic results provide insight into the narrow window of PDI conjugation necessary for selective ATF6 activation mediated by metabolically activatable ER proteostasis regulators. In this view, the activity of a few of the PDIs can be viewed as a rheostat controlling a spectrum of UPR transcriptional activities. From a therapeutic perspective, concomitant activation of . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted January 17, 2023. ; https://doi.org/10.1101/2023.01.16.524237 doi: bioRxiv preprint ATF6 and IRE1 signaling leads to a uniquely remodeled ER proteostasis environment preferable for some disease contexts, as compared to individual activation of individual UPR pathways. 54,55 Thus, the extent of global UPR activation observed with AA132 may be of some value in certain disease contexts. Regardless, a continued understanding of the factors driving selectivity of the transcriptional response induced by ER proteostasis regulators is essential for the development of improved ATF6 activators for treatment of etiologically diverse diseases.
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made  (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted January 17, 2023. ;  . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted January 17, 2023. ; Fold Change of AA132 yneenriched proteins relative to vehicle (X-axis) and Log2 Fold Change of competition ratio (AA132 yne + Veh / AA132 yne + AA132) (Y-axis). Dotted lines indicate significantly enriched proteins (>3-fold; x-axis) and proteins with a significant competition ratio (>1.5 fold; y-axis). Red circles identify AA132 yne targets with PDI GO annotation; blue circles identify additional AA132 yne targets identified across two separate experiments. Grey circles identify AA132 yne targets identified in one experiment. Data included in Table S1. C. Venn Diagram showing unique targets of AA147 yne (red) or AA132 yne (blue). Underlined proteins have the GO annotation GO:0044432 "Endoplasmic Reticulum Part".
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made   Table S2. B. Bar graph of enrichment ratio of select PDIs by indicated the compound relative to DMSO from data shown in Fig. S4A (N = 4 biological replicates). ****p < 0.001 from multiple unpaired t-test. C. Representative immunoblot and quantification of PDIA3 and PDIA4 recovery in streptavidin enrichments from HEK293T cells treated with AA147 yne (10 µM; 6 h) or AA132 yne (10 µM; 6 h) and then conjugated to biotin. Fraction enrichment was calculated by dividing the signal in enriched samples by the input signal. *p < 0.05 from unpaired t test for N=3 replicates.
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made  Representative SDS-PAGE gel of Cy5-conjugated proteins from HEK293T cells treated for the indicated time point with AA132 yne (10 µM). C. Quantification of gels described in Fig 5A and Fig 5B. Y-axis labeling relative intensity represents PDIA4 intensity normalized to labeling at 6h. Error bars represent S.E.M (N = 4 Biological Replicates). *p < 0.05, **p < 0.01.
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted January 17, 2023.    Table S3. B,C. Top-10 GO terms for significantly induced genes (fold change>1.3, p<0.05) identified by RNAseq in HEK293T cells treated with 3 µM (B) or 6 µM (C) AA132 for 6 h. RNAseq data is included in Table S3. Full GO analysis is included in Table S4. D-F. Fold change, relative to vehicle, for select ATF6 target genes (D), IRE1/XBP1s target genes (E), or PERK/ATF4 target genes (F) from RNAseq data of HEK293T cells treated with increasing concentrations of AA132 for 6 h. The average EC50 for the four target genes representing each UPR pathway are shown. G-I. Fold change, relative to vehicle, for genesets of 15-20 genes regulated downstream of ATF6 (G), IRE1/XBP1s (H), or PERK/ATF4 (I) from RNAseq of HEK293T cells treated with increasing concentrations of AA132 for 6 h. The fold change expression of individuals genes was normalized to that observed with the global ER stressor thapsigargin (Tg), as reported previously. 19,37 The impact of AA147 (10 µM; 6 h) on these genesets is shown on the right. The expression of UPR genesets is shown in Table S5. *p < 0.05, **p < 0.01, ***p < 0.005, ****p < 0.001 for one-way ANOVA.
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted January 17, 2023.  Table S3 Full GO analysis is included in Table S4. . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made  SUPPLEMENTARY TABLE LEGENDS   Table S1. Excel spreadsheet showing the enrichment and competition ratio for proteins identified as targets of AA132 yne . Related to Fig. 3. Table S2. Excel spreadsheet showing fold enrichment for AA132 yne /AA147 yne of proteins identified in proteomics experiments performed in HEK293T or HepG2 cells. Related to Fig. 4 and Fig. S4. . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted January 17, 2023. incubated at 37°C overnight. The following day, cells were treated with 25 µL of compound-containing media to give final concentration as described before incubating for 18 hr at 37°C. The plates were equilibrated to room temperature, then either 125 μL of Firefly luciferase assay reagent-1 (ERSE.FLuc and ATF4.FLuc) or Renilla luciferase assay reagent-1 (XBP1s.RLuc) (Targeting Systems) were added to each well. Samples were dark adapted for 10 min to stabilize signals. Luminescence was then measured in an Infinite F200 PRO plate reader (Tecan) and corrected for background signal (integration time 250 ms). All measurements were performed in biologic triplicate.
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted January 17, 2023. ; https://doi.org/10.1101/2023.01.16.524237 doi: bioRxiv preprint

Quantitative RT-PCR
The relative mRNA expression levels of target genes were measured using quantitative RT-PCR. Cells were treated as described at 37°C, harvested by trypsinization, washed with Dulbecco's phosphatebuffered saline (GIBCO), and then RNA was extracted using the QuickRNA Miniprep Kit (Zymo). qPCR reactions were performed on cDNA prepared from 500 ng of total cellular RNA using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems). PowerSYBR Green PCR Master Mix (Applied Biosystems), cDNA, and appropriate primers purchased from Integrated DNA Technologies (see Table   below

Immunoblotting of AA132 yne or AA147 yne conjugated proteins
HEK293T cells grown to 80-90% confluency in 10 cm plates were treated with 10 µM indicated compound (AA147 yne , AA132 yne , or Veh) for 6h at 37 ˚C. The cells were washed with PBS before harvesting with tryspin, pelleting (500 g, 5 min), and washed with PBS (1 mL). Cells pellets were resuspended in radioimmunoprecipitation assay (RIPA) buffer before sonication with a probe tip sonicator to lyse the cells (15 sec, 3 sec on/2 off, 30% amplitude). The lysates were cleared via centrifugation and the concentration of protein adjusted to 4 mg / mL using the BCA assay. 2 g protein (500 µL) was taken and reacted with a mixture of diazo biotin azide (100 µM), copper (II) sulfate (800 µM), BTTAA (1.6 mMM), sodium ascorbate (5.0 mM) for 90 min at 30 ˚C with shaking (600 rpm). The . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted January 17, 2023. ; https://doi.org/10.1101/2023.01.16.524237 doi: bioRxiv preprint reaction was quenched with the sequential addition of cold methanol (4x volume), chloroform (1x volume), and DPBS (4x volume) to precipitate proteins. Proteins were pelleted by centrifugation (4,700 g, 10 min, 4 ˚C). The supernatant was discarded, and the pellets dried under air for 5 min. Protein pellets were resuspended in 6M urea in PBS (500 µL) with brief sonication. 25 µL of sample was taken as the "input" for Western blot analysis. Sample diluted with 5.5 mL DPBS (.2% SDS) and streptavidin agarose resin (100 µL, washed 3 x 1 mL with PBS) was added to each sample before incubation for

ALMC2 cells grown to concentration of 2 million cells/mL and 15 mL cell suspension incubated in T75
flask with indicated compound (10 µM, 2h, 37° C). After treatment, cells were pelleted by centrifugation (3 min, 300 g) and washed with PBS before resuspension in PBS at 30 million cells/mL. 100 µL cell suspension was added to 0.2 mL PCR tube before heat treatment at indicated temperature for 3 min and room temperature incubation for 3 min. Samples snap frozen at -80 C were lysed by sequential freeze-thaw cycles and centrifuged (15,000 g, 10 min) to pellet insoluble material. Soluble fractions were boiled for 5 min in Laemmli buffer with 100 mM DTT before loading onto SDS-PAGE gels. Proteins were transferred from gel slabs to PVDF membranes and blotted using rabbit anti-PDIA1 antibody (1:1000) Protein Tech) and visualized on the Odyssey Infrared Imaging System (Li-Cor Biosciences).
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted January 17, 2023. ; https://doi.org/10.1101/2023.01.16.524237 doi: bioRxiv preprint
Cells pellets were resuspended in radioimmunoprecipitation assay (RIPA) buffer before sonication with a probe tip sonicator to lyse the cells (15 sec, 3 sec on/2 off, 30% amplitude). For each sample, 1 g lysate (500 µL) were reacted with click reagents to give final concentrations as follows: 100 µM of diazo biotin-azide (Click Chemistry Tools, Scottsdale, AZ), 800 µM copper (II) sulfate, 1.6 mM BTTAA ligand and acidified with 5% formic acid. Acetonitrile was evaporated on a SpeedVac and debris was removed . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted January 17, 2023. ; https://doi.org/10.1101/2023.01.16.524237 doi: bioRxiv preprint by centrifugation for 30 min at 18,000×g. MuDPIT (Multi-Dimensional Protein Identification Technology) microcolumns were prepared as described previously 79. LCMS/MS analysis was performed using a Q Exactive mass spectrometer equipped with an EASY nLC 1000 (Thermo Fisher). MuDPIT experiments were performed by 5 min sequential injections of 0, 20, 50, 80, 100% buffer C (500 mM ammonium acetate in buffer A) and a final step of 90% buffer C / 10% buffer B (20% water, 80% acetonitrile, 0.1% fomic acid, v/v/v) and each step followed by a gradient from buffer A (95% water, 5% acetonitrile, 0.1% formic acid) to buffer B. Electrospray ionization was performed directly from the analytical column by applying a voltage of 2.5 kV with an inlet capillary temperature of 275°C. Datadependent acquisition of MS/MS spectra was performed with the following settings: eluted peptides were scanned from 400 to 1800 m/z with a resolution of 30,000 and the mass spectrometer in a data dependent acquisition mode. The top ten peaks for each full scan were fragmented by HCD using a normalized collision energy of 30%, a 100 ms activation time, a resolution of 7500, and scanned from 100 to 1800 m/z. Dynamic exclusion parameters were 1 repeat count, 30 ms repeat duration, 500 exclusion list size, 120 s exclusion duration, and exclusion width between 0.51 and 1.51. Peptide identification and protein quantification was performed using the Integrated Proteomics Pipeline Suite (IP2, Integrated Proteomics Applications, Inc., San Diego, CA) as described previously.

RNA-seq analysis
Cells were lysed and total RNA collected using the Quick-RNA Miniprep kit from Zymo Research (R1055) according to manufacturer's instructions. RNA concentration was then quantified by NanoDrop. Whole transcriptome RNA was then prepared and sequenced by BGI Americas on the BGI Proprietary platform, which provided paired-end 50 bp reads at 20 million reads per sample. Each condition was performed in triplicate. RNAseq reads were aligned using DNAstar Lasergene SeqManPro to the Homo_sapiens-GRCh38.p7 human genome reference assembly, and assembly data were imported into ArrayStar 12.2 with QSeq (DNAStar Inc.) to quantify the gene expression levels and normalization to reads per kilobase per million. Differential expression analysis was assessed using . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

Statistical Analysis
Unless otherwise noted, the data were tested for significance using One-Way ANOVA with a post-hoc Dunnett's test.

General Synthetic Procedures
All compounds and reagents were purchased from Sigma-Aldrich, Acros, Alfa Aesar, Combi-blocks, and EMD Millipore unless otherwise noted and were used without further purification. Thin layer chromatography with Merck silica plates (60-F254), using UV light as the visualizing agent, was used to monitor reaction progress. Flash column chromatography was carried out using a Teledyne Isco Combiflash Nextgen 300+ machine using Luknova SuperSep columns (SiO2,25 µm) with ethyl acetate and hexanes as eluents. 1 H NMR spectra were recorded on a Varian INOVA-400 400MHz spectrometer. Chemical shifts are reported in δ units (ppm) relative to residual solvent peak. Coupling constants (J) are reported in hertz (Hz). Characterization data are reported as follows: chemical shift, multiplicity (s=singlet, d=doublet, t=triplet, q=quartet, br=broad, m=multiplet), coupling constants, number of protons, mass to charge ratio. The compound's identity was confirmed via high-resolution mass spectrometry.

Synthesis of AA132
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted January 17, 2023. ; https://doi.org/10.1101/2023.01.16.524237 doi: bioRxiv preprint To a flame dried flask added 1-Decyl-3-methylimidazolium chloride (3.52 g, 13.6 mmol, 3.2 eq) and aluminum chloride (1.02 g, 7.65 mmol, 1.8 eq) and stirred for 16 hours under Argon to make the chloroaluminate fluid. To this flask at 0 °C, added 4-fluorobenzoyl chloride (S1, 500 µL, 4.25 mmol, 1 eq) slowly and the mixture became syrupy. The mixture was warmed to 75 ° C and acetylene gas generated in a separate flask from calcium carbide was bubbled through for 2 hours to afford the ßchlorovinyl ketone. The crude was added to ice water and extracted with ether which was washed once with brine and concentrated to give a yellow oil (S2), which was used for the next step without further purification.
To a microwave vial with S2 (141 mg, 0.75 mmol, 1 eq), we added 2-amino-p-cresol (S3, 91 mg, 0.75 mmol, 1 eq) and p-toluenesulfonic acid (28.5 mg, 0.15 mmol,0.2 eq). Solids were dissolved in 3 mL EtOH and stirred under argon for 16h at 70 °C. Reaction diluted in EtOAc and washed sequentially with water and 1M HCl before drying over MgSO4. Product purified by column chromatography (SiO2, 4:1 Hex:EtOAc) to give AA132 as a yellow solid (98.2 mg, 48% yield). . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

Synthesis of AA132 yne
To 4-Iodobenzoic acid (S3, 738 mg, 3 mmol, 1 eq) dissolved in 15 mL DCM added oxalyl chloride (321 µL, 3.75 mmol, 1.25 eq) dropwise at 0 °C before addition of several drops of DMF. Reaction allowed to warm to room temperature and stir for 3h before TLC indicated complete conversion to the acyl chloride.
Reaction stirred under argon at 40 °C overnight before washing with saturated NaHCO3, and brine before drying over MgSO4 and concentrating. The crude residue was purified by flash column chromatography (SiO2, 9:1 Hex/EtOAc) to give an orange solid (S4, 638 mg, 71 % yield). . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made To a microwave vial charged with S4 (59 mg, 0.2 mmol, 1 eq), we added 2-amino-p-cresol (S3; 23.6 mg, 0.2 mmol, 1 eq). The reaction vessel was sealed before addition of 1 mL anhydrous methanol and heated at 70 ° C for 16 h. Reaction diluted in EtOAc, washed with 1M HCl and brine, and dried over MgSO4. The solvent was removed under reduced pressure. The crude residue purified by flash column chromatography (SiO2, 9:1 Hex/EtOAc) to afford the product S5 as a yellow powder (43 mg, 62% yield).  (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted January 17, 2023. ; https://doi.org/10.1101/2023.01.16.524237 doi: bioRxiv preprint