| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
1 Molecular Endocrinology and Oncology Research Center, Laval University Medical Research Center (Centre Hospitalier Universitaire de Québec)
2 Department of Anatomy and Physiology, Faculty of Medicine, Laval University, Quebec City, Quebec, Canada
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
RheoSwitch; androgen-induced bZIP; cAMP response element-binding protein; microarray
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
One of the major functions of the ER is the folding and posttranslational modification of secreted and integral membrane proteins (5). The ER's ability to correctly process proteins is extremely sensitive to alterations in the intracellular environment. For example, protein processing requires adequate supplies of glucose, ATP, and calcium, such that deficiencies in any of these molecules can lead to the accumulation and aggregation of misfolded proteins in the ER lumen. The unfolded protein response (UPR) is a complex adaptive response that serves to prevent or reverse the accumulation of misfolded proteins by repressing protein biosynthesis and augmenting the protein-processing capacity of the ER (33). One of the key components of this response is ATF6 (activating transcription factor 6), a bZIP transcription factor that regulates the expression of ER chaperones. ATF6 is inserted in the ER membrane as a type II transmembrane protein such that its amino-terminal transcription factor domain resides in the cytoplasm, whereas its carboxy-terminal regulatory domain lies within the ER lumen. In a manner akin to SREBP, ATF6 is retained in the ER membrane by a protein that interacts with its regulatory domain, in this case the chaperone BiP/GRP78/HSPA5 (IgG heavy chain-binding protein; glucose-regulated protein, 78 kDa; heat shock 70 kDa protein 5). When misfolded proteins accumulate in the ER, BiP dissociates from ATF6, and ATF6 is transported to the Golgi apparatus where it is processed by the same proteases that process SREBP.
More recent work revealed that ATF6 is not the sole ER-associated bZIP transcription factor. The cAMP responsive element-binding protein-3 (CREB3) subfamily of bZIP proteins consists of five transcription factors that share the same domain structure as ATF6 and are also inserted in the ER membrane as type II transmembrane proteins (9, 20, 22, 23, 25, 32). The existence of multiple ER-associated bZIP proteins raises several important questions regarding their respective roles in cell physiology. While these proteins may serve partially redundant functions, it is conceivable that they could regulate unique subsets of genes in response to diverse physiological stimuli. Indeed, although the tissue distribution patterns of human CREB3 (LZIP/Luman) and the CREB3-like proteins OASIS, BBF2H7, CREBH, and AIbZIP partially overlap, there are some important differences. For instance, CREBH is exclusively expressed in hepatocytes, whereas AIbZIP is most highly expressed in prostate.
Experimentally, the UPR can be triggered by pharmacological compounds that interfere with different aspects of protein processing. Drugs that deplete ER calcium stores, such as thapsigargin, or that inhibit protein glycosylation, such as tunicamycin, have proven useful tools to probe the role of CREB3-related proteins in the UPR. These experiments have confirmed that CREB3-related proteins can be processed to their transcriptionally active forms in cells in which protein processing has been compromised (14, 15, 26, 31, 34). Moreover, it appears that one of these proteins (OASIS) cooperates with ATF6 in upregulating BiP expression in astrocytes (14). However, although these data are consistent with a role for CREB3-related proteins in the UPR, they do not exclude their implication in other biological pathways.
An alternative approach to defining the role of CREB3 proteins is to identify their transcriptional targets. A major advantage of this approach is that it does not require the use of drugs that profoundly disturb cell function. Having identified genes that are regulated by these transcription factors, one can then work backward to define the physiological context in which said genes might be upregulated. At present, the transcriptional targets of CREB3-related transcription factors are not well defined. Nevertheless, the few published attempts to identify these targets have proven to be quite enlightening. In mice, for instance, CREBH knockdown by short interfering RNA (siRNA) significantly reduced the expression of mRNAs that encode C-reactive protein (CRP) and serum amyloid P-component (SAP), thereby revealing an unsuspected link between CREBH and the acute phase response (34). Using cDNA microarrays, Liang et al. (19) discovered that CREB3 induces transcription of HERPUD1, an ER protein that is involved in the degradation of misfolded proteins.
Our laboratory is interested in androgen-induced bZIP (AIbZIP), an androgen-regulated bZIP protein that is most highly expressed in the luminal cells of the prostatic glandular epithelium (25). The primary function of these cells is to produce components of the seminal fluid that enable spermatozoa to fertilize ova. Thus the presence of high amounts of AIbZIP in luminal prostate cells suggests a link between AIbZIP and the secretory function of these cells. It is also important to consider that AIbZIP is overexpressed in prostate tumors (17, 25, 29), which suggests that some of the downstream targets of AIbZIP might contribute to prostate cancer cell survival. In humans, AIbZIP mRNA is also expressed, albeit at lower levels, in other organs such as the pancreas, liver, testis, ovary, kidney, and placenta, but the specific cell types that express AIbZIP have yet to be determined (6).
In an effort to clarify the role of AIbZIP in prostate cells, we first used a pharmacological approach to confirm that AIbZIP can be processed by RIP in LNCaP cells. We then used a genome-wide approach to identify AIbZIP-regulated genes in LNCaP cells. Interestingly, although some transcriptional targets of AIbZIP are implicated in protein processing, most of the genes that we identified are involved in other aspects of cell function, including transcriptional regulation, small molecule transport, signal transduction, and metabolism.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
Cell Lines That Conditionally Produce the Processed Form of AIbZIP
The plasmid used to generate cell lines that support RSL1-dependent expression has been described in detail elsewhere (16). Clone 6-6 was obtained by transfecting LNCaP cells with a plasmid (pZRD) that constitutively produces the RheoSwitch transcription factors, the green fluorescent protein hrGFP, and the Zeocin resistance gene (16). Cell lines that conditionally produce a recombinant form of nuclear AIbZIP (A290HA) were obtained by transfecting LNCaP cells with plasmid pZR290. pZR290 was derived from pZRD by adding an RSL1-responsive cassette that controls the production of the first 290 amino acids of AIbZIP (GenBank accession no. NP_570968) fused to the linker sequence SRGP and a carboxy-terminal HA epitope (YPYDVPDYASL).
Nuclear Extracts
Nuclear extracts were prepared as described previously (26, 28) with slight modifications. The cells were rinsed with ice-cold PBS two to three times, harvested, and centrifuged at 800 g for 10 min at 4°C. After centrifugation, the cell pellets were suspended in five packed cell volumes of hypotonic buffer (buffer H) and left on ice for 30 min. They were then passed through a 27.5-gauge needle 30 times and centrifuged at 1,000 g for 10 min at 4°C. The 1,000 g pellet was resuspended in one volume of extraction buffer (buffer E), rotated for 60 min at 4°C, and centrifuged at 30,000 g for 30 min at 4°C. The supernatant from this spin was used as the nuclear extract. Buffer H contained 10 mM HEPES-KOH (pH 7.6), 10 mM KCl, 1.5 mM MgCl2, 1 mM EDTA, 1 mM EGTA, 0.5 mM DTT, and a cocktail of protease inhibitors (Complete, Mini, Roche). Buffer E contained 20 mM HEPES-KOH (pH 7.6), 25% glycerol, and 0.5 M NaCl and the same concentrations of MgCl2, EDTA, EGTA, DTT, and protease inhibitors as in buffer H.
Immunoblotting
The full-length AIbZIP protein was detected using an affinity-purified rabbit polyclonal antibody (AB150) raised against a carboxy-terminal peptide of AIbZIP (25). The endogenous and recombinant forms of nuclear AIbZIP were detected using an affinity-purified rabbit polyclonal antibody (AB515) that was raised against a peptide of mouse AIbZIP (LPSHLPLTKAEERILKK) which is nearly identical (16 of 17 residues) to residues 204–220 of human AIbZIP. Primary antibodies against B23 (C-19) and BiP (H-129) were from Santa Cruz Biotechnology (Santa Cruz, CA). Peroxidase-conjugated secondary antibodies were from Jackson ImmunoResearch Laboratories (West Grove, PA), and the antigen-antibody complexes were revealed using a chemiluminescent detection system (Pierce, Rockford, IL).
RNA Interference
The Kill-1 siRNA used to silence AIbZIP expression was purchased from Qiagen (Valencia, CA). Kill-1 targets codons 335–341 of the AIbZIP coding sequence (GenBank accession no. NM_130898). The siRNA duplex consists of the sense strand r(GAAAUAUCCUGACCCACAA)dTdT and the antisense strand r(UUGUGGGUCAGGAUAUUUC)dTdG. Transfections were performed by mixing the siRNA and the RNAiFect transfection reagent in a 1:6 ratio.
Microarray Experiment and Data Analysis
Clone 7-11 cells (passage 42) were plated in 24 10-cm culture dishes. Sixteen dishes were cultured in medium containing RSL1 (500 nM), and eight dishes were cultured in medium containing an equivalent amount of the RSL1 vehicle (DMSO). Eight RSL1-treated dishes were harvested 6 h after addition of RSL1, and another eight dishes were harvested 24 h after addition of RSL1. At each time point, the dishes were pooled two by two to yield a total of 12 individual samples, i.e., 4 controls, 4 exposed to RSL1 for 6 h, and 4 exposed to RSL1 for 24 h. One sample of each group was processed to nuclear extracts to assess the levels of recombinant AIbZIP by immunoblotting. The three other samples of each group were processed separately to cRNA probes that were then hybridized to Affymetrix Human Genome U133 Plus 2.0 arrays (batch nos. 4003850 and 4003847).
The details of the microarray hybridization and the statistical methods used to identify differentially expressed genes were described previously (16). The data discussed in this publication have been deposited in the National Center for Biotechnology Information Gene Expression Omnibus (GEO; http://www.ncbi.nlm.nih.gov/geo/) and are accessible through GEO series accession number GSE7223 (1). The two-dimensional cluster map was constructed using the standard correlation similarity algorithm of GeneSpring 7.3. Ontology analyses of the genes that were modulated in RSL1-treated cells at 24 h were performed by uploading the corresponding probe sets to the web-based application database for annotation, visualization, and integrated discovery (DAVID) (7). For each data set (1,086 upregulated genes and 852 downregulated genes), two lists were generated using the default parameters of DAVID (Supplemental Tables 1–4; supplemental data are available at the online version of this article). The "functional annotation chart" highlights the most relevant Gene Ontology (GO) terms associated with each gene list, whereas "functional annotation clustering" classifies highly related genes into functionally related groups.
Probes for Northern Blot Analysis
The portions of transcripts corresponding to potential AIbZIP-responsive genes were amplified by RT-PCR from LNCaP cells and subcloned into plasmid pCR2.1 (Invitrogen). The sequences of the resulting plasmids were verified using an automated sequencer. The amplified fragments or portions thereof were excised from pCR2.1 by digestion (usually with EcoRI). The precise coordinates of each probe are available on request.
Candidate Gene Validation
Most of the Northern blots presented herein were performed using RNA isolated from clone 7-11 cells (passage nos. 46–51) that were exposed to 500 nM RSL1 (for 6 or 24 h), 2 µM A23187 (for 24 h), or vehicle. mRNA was purified from total RNA using the PolyATract system (Promega, Madison, WI). Some experiments were also performed using total RNA isolated from clones 6-6, 7-8, 7-11, 7-12, and 7-30, which were exposed to RSL1 (500 nM) or vehicle for 24 h. Total RNA (20 µg) or polyA(+) RNA (2 µg) samples were fractionated on 1% agarose-formaldehyde gels, transferred to nylon membranes, and hybridized to radiolabeled cDNA probes as described (16). Following hybridization with the candidate cDNA probe, selected blots were rehybridized (without stripping) with a radiolabeled cDNA probe that detects the constitutively expressed RheoReceptor-1 mRNA. Following autoradiography, the images were captured using a digital camera and imported into PhotoShop (Adobe, San Jose, CA). The images were not altered unless the background was very high, in which case the image contrast was adjusted accordingly.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
The unprocessed form of AIbZIP was detected using a polyclonal antibody raised against the regulatory domain of the protein (25). In whole cell extracts of LNCaP cells, this antibody detects three 
47.5-kDa polypeptides that correspond to differentially glycosylated forms of AIbZIP (Fig. 1) . In fact, human AIbZIP contains putative N-glycosylation sites, and we and others have established that mouse AIbZIP is glycosylated in vivo (8, 21). Treatment with brefeldin A (10 µg/ml) caused a rapid decrease in the amount of unprocessed AIbZIP, which was reduced to nil within 4 h (Fig. 1A). These results are consistent with AIbZIP being a substrate for Golgi proteases.
|
The compound tunicamycin, which causes ER stress by inhibiting N-glycosylation, has been shown to activate ATF6, CREB3, and OASIS (11, 14, 26). Tunicamycin (1 µM) produced a robust increase in the levels of the ER chaperone BiP, consistent with UPR induction, but it did not produce a notable decrease in the total amount of AIbZIP protein (Fig. 1B). Nevertheless, the relative abundance of the three AIbZIP polypeptides was transiently altered in tunicamycin-treated cells, which suggests that the drug interfered with AIbZIP glycosylation.
Calcium is a versatile signaling molecule that is stored primarily in the ER, where it is also required by chaperones involved in protein folding (3). We therefore examined whether AIbZIP is affected by drugs that deplete ER Ca2+ stores. Thapsigargin inhibits Ca2+ uptake by the sarco(endo)plasmic reticulum Ca2+-ATPase, whereas A23187 is a Ca2+ ionophore that sequesters Ca2+ in the cytoplasm. Both thapsigargin (1 µM) and A23187 (2 µM) induced ER stress in LNCaP cells, as indicated by the increase in BiP levels (Fig. 1, C and D). In addition, both drugs caused a progressive decrease in the amount of unprocessed AIbZIP, which was completely eliminated by 42 h. Caffeine (10 mM), which stimulates calcium release from the ER via ryanodine receptors, also downregulated AIbZIP (Fig. 1E). Caffeine produced a very slight increase in BiP protein levels.
AIbZIP is processed to its transcriptionally active form by drugs that alter ER calcium concentrations
The effects of A23187, thapsigargin, and caffeine on ER-associated AIbZIP suggested that AIbZIP was processed by RIP, but we could not exclude the possibility that this downregulation resulted from an effect on AIbZIP expression, translation, or protein stability. To verify that AIbZIP was indeed processed to its transcriptionally active form in stressed cells, we exposed LNCaP cells to stress-inducing drugs and analyzed nuclear extracts using a polyclonal antibody raised against the transcription factor domain of AIbZIP (Fig. 2A).
|
33-kDa polypeptide that was not detected in nuclear extracts of control cells (Fig. 2B, compare lanes 3 and 4). To verify that this polypeptide was derived from AIbZIP, we transfected LNCaP cells with Kill-1, an siRNA that targets the last coding exon of AIbZIP. Kill-1 reduced the amount of unprocessed AIbZIP by 95% within 32 h (Fig. 2B, compare lanes 1 and 2). Depletion of AIbZIP before addition of brefeldin A prevented the appearance of the 33-kDa nuclear polypeptide (Fig. 2B, lane 5), thereby confirming that this polypeptide corresponds to the processed form of AIbZIP. In preliminary experiments, we determined that the processed form of AIbZIP was detectable in nuclear extracts as early as 3 h following exposure to A23187 (data not shown). To compare the ability of stress-inducing drugs to induce AIbZIP processing, nuclear extracts were therefore prepared from cells exposed to the drugs for 3 h. Under these experimental conditions, A23187 was the most potent inducer of AIbZIP processing, and doses of A23187 ranging from 2 to 20 µM produced similar levels of processed AIbZIP. Caffeine also induced AIbZIP processing, whereas a lesser amount of processed protein was observed in thapsigargin-treated cells, and no processed AIbZIP was observed in tunicamycin-treated cells. Overall, the experiments with ER stress-inducing drugs indicate that alterations in calcium homeostasis induce AIbZIP activation.
Microarray-Based Screening for AIbZIP-Regulated Transcripts
To begin to understand the biological consequences of AIbZIP activation, we sought to identify genes that are regulated by the processed form of AIbZIP in LNCaP cells. For this purpose, we generated cell lines that conditionally produce a recombinant form of processed AIbZIP. The recombinant protein, hereafter referred to as A290HA, consists of the transcription factor domain of AIbZIP (amino acids 1–290) fused to a carboxy-terminal HA epitope.
Several independent LNCaP cell lines that conditionally produce A290HA were generated using the RheoSwitch expression system (Fig. 3A). This system employs a nonsteroidal analog of the insect hormone ecdysone (RSL1) to activate a heterodimer of chimeric nuclear receptors that regulates the production of A290HA via a GAL4-responsive promoter (24). We have extensively characterized the RheoSwitch system in LNCaP cells and found that RSL1 exerts minimal effects on the expression of endogenous genes (16).
|
2 h following the addition of RSL1 (data not shown). To capture as many A290HA-regulated transcripts as possible, we therefore prepared RNA from cells that were exposed to RSL1 for 6 and 24 h. As shown in Fig. 3B, A290HA was undetectable before addition of RSL1, whereas addition of RSL1 produced a robust increase in A290HA levels. RNA samples were processed to probes that were hybridized to Affymetrix Human Genome U133 Plus 2.0 microarrays containing 54,000 probe sets. Analysis of the microarray hybridization data produced a list of 2,701 probe sets that detected statistically significant (P < 0.001 vs. control cells) changes in mRNA abundance at 6 and/or 24 h following addition of RSL1. Of the 2,701 regulated probe sets, 1,296 detected increases in mRNA abundance, whereas 1,405 detected decreases in mRNA abundance. These 2,701 probe sets were then subjected to two-dimensional cluster analysis as depicted in Supplemental Fig. 1. This visual representation of the data confirms that the RSL1/A290HA-related changes in mRNA abundance were reproducible across multiple biological replicates.
To begin to understand the biological consequences of AIbZIP activation, we sought to determine whether the transcripts that were modulated at 24 h segregate to specific biological pathways or processes. The lists of probe sets that were significantly regulated at 24 h were uploaded to the DAVID database (7), which processed the microarray data to generate nonredundant lists of 1,086 upregulated genes and 852 downregulated genes. The GO annotations derived from the list of genes with increased expression identified several genes with either DNA-, RNA-, nucleotide-, or protein-binding activity, including genes that bind to unfolded proteins (Supplemental Table 1). Functional annotation clustering revealed that the list of A290HA-induced transcripts is most enriched for nuclear proteins (Supplemental Table 2). However, several other cellular functions (e.g., RNA and DNA metabolism) were highly enriched. On the other hand, the list of downregulated genes was most enriched for proteins involved in Golgi function and protein transport (Supplemental Tables 3 and 4). Together these analyses suggest that AIbZIP might be implicated in processes that have not been previously linked to other ER-associated bZIP transcription factors.
Because ER-bound transcription factors are generally thought to activate gene expression in response to physiological stress, we were primarily interested in the probe sets that detected increases in mRNA abundance. A large number of probe sets detected statistically significant increases at 6 h (227 probe sets) and 24 h (1,250 probe sets). Of the 227 probe sets that were upregulated at 6 h, 181 were also upregulated at 24 h. Although most of these probe sets reported relatively small (<3-fold) increases in mRNA levels, we identified 25 genes that were upregulated 3- to 5-fold at 6 h and 121 genes that were upregulated 3- to 17-fold at 24 h (Supplemental Table 5).
Validation of A290HA-Regulated Transcripts
For validation, we selected 49 genes that we considered of particular interest based on their novelty or on their known or suspected involvement in ER function, metabolic processes, transport, cell stress, transcription, and/or signal transduction (Table 1). Nine of the selected genes were upregulated at least threefold at both 6 and 24 h, one gene was upregulated at least threefold at 6 h only, and twenty-nine genes were upregulated at least threefold at 24 h only. The 10 other genes were upregulated 1.7- to 3-fold at 24 h. Representative probe set data for each of these genes are presented in Supplemental Table 6 None of the genes that were selected for validation was upregulated by RSL1 in a control RheoSwitch cell line (LNCaP clone 6-6) that does not produce A290HA (16).
|
|
|
Protein processing.
Although the microarrays did not identify BiP (heat shock 70 kDa protein 5; HSPA5) as a potential target of A290HA, we evaluated HSPA5 mRNA levels in A23187-treated and A290HA-expressing cells to monitor ER stress. As expected, HSPA5 mRNA levels increased in cells exposed to A23187. On the other hand, HSPA5 was unaffected by exposure to RSL1, which indicates that A290HA and its downstream targets did not induce ER stress.
Several genes implicated in protein maturation and sorting were upregulated in A290HA-producing cells. These include the chaperone regulators BAG3 and DNAJC12, the O-glycosylating enzyme GALNT3, the Golgi assembly protein GORASP2, and the ER sorting receptor KDELR3. Lesser increases in mRNA levels were observed for the chaperone CALU, the ER quality control receptor EDEM1, the lectin LMAN1, and KDELR2. The observed increase in EDEM1 mRNA levels confirms an earlier report showing that EDEM1 was upregulated by transient overexpression of mouse AIbZIP in HeLa cells (21). Several of these transcripts, DNAJC12 in particular, were markedly upregulated by A23187.
Sugar and lipid metabolism.
A number of genes implicated in lipid and sugar metabolism were upregulated in A290HA-expressing cells. These include GFPT1, a key enzyme of the hexosamine pathway; LIPA, a triglyceride and cholesterol ester lipase; and PGM1, which interconverts glucose-1-phosphate and glucose-6-phosphate, all of which were upregulated at 6 h. Appreciable increases in mRNAs encoding ACSL3, an enzyme implicated in lipid biosynthesis; INSIG1, a regulator of SREBPs; and UGT2B28, an enzyme that inactivates steroids, were observed at 24 h. GFPT1, INSIG1, and PGM1 were upregulated by A23187.
Channels and transporters.
Conditional expression of A290HA resulted in the upregulation of genes that encode membrane proteins that transport small molecules. Robust increases in mRNAs encoding the phosphate transporter ANKH, the water and glycerol channel AQP3, and the potassium channel KCNK1 were detected in cells exposed to RSL1 for 6 h. Substantial increases in mRNAs encoding the potassium channel CUZD1, the sulfate transporter SLC26A2, and the putative ion transporter TMCO3 were observed at 24 h. TMCO3 and, to a lesser extent, AQP3, CUZD1, and KCNK1 were upregulated by A23187.
Signal transduction.
Increased levels of mRNAs encoding proteins implicated in a variety of cell signaling pathways were observed in A290HA-producing cells. Substantial increases in mRNAs encoding a phospholipid-binding autocrine factor (ANXA2), a regulator of NF-
B signaling (IKIP), a secreted phospholipase (PLA2G2A), and a protein implicated in TGFß signaling (THBS1) were manifest at 6 and/or 24 h. Lesser increases in mRNAs encoding the Rho GTPase regulator CDC42EP3, the adenosine deaminase CECR1, and the tyrosine phosphatase DUSP1 were observed at 24 h. With the possible exception of DUSP1, none of these transcripts was markedly upregulated by A23187.
Miscellaneous genes.
Interestingly, the downstream targets of A290HA include three genes (ANG, DOM3Z, RNASE4) that are implicated in RNA metabolism. Increases in all three transcripts were detected 6 and 24 h following addition of RSL1. The other transcripts that were upregulated at 6 h encode GADD45B, a known stress-induced protein; OASL, a protein of unknown function that interacts with the transcriptional repressor MBD1; two transmembrane proteins of unknown function (SIDT2 and TMEM45A); and KIAA0774. Transcripts encoding a surface sialoglycoprotein (CD24), a lysyl oxidase (LOX), a myc- and stress-responsive gene (NDRG1), an androgen-regulated secreted serine protease (TMPRSS2), and a protein implicated in endocytosis (WDFY2) were upregulated at 24 h. ANG, DOM3Z, GADD45B, RNASE4, TMEM45A, and WDFY2 mRNAs were upregulated in cells exposed to A23187.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
We anticipated that A290HA might upregulate genes implicated in protein processing, because drugs that cause ER dysfunction induced AIbZIP activation. In fact, A290HA upregulated genes that encode ER proteins implicated in protein folding (BAG3, DNAJC12), sorting (KDELR3), and degradation (EDEM1). It is important to note that these genes are not known to be regulated by other ER-bound bZIP transcription factors. Conversely, A290HA did not upregulate genes that are known targets of ATF6 and OASIS (e.g., BiP) or of CREB3 (HERPUD1). Thus, while most ER-bound bZIP transcription factors appear to be implicated in the regulation of protein maturation in the ER, it appears that each CREB3-like protein fulfills specific functions.
In addition to identifying A290HA-regulated genes, the present study identified several transcripts that were not previously known to be upregulated during A23187-induced ER stress. These include genes that would be expected to function in the UPR, such as DNAJC12 and KDELR3, as well as genes that encode ribonucleases (ANG, RNASE4), transcription factors (CREB3, ID2), regulators of lipid and sugar metabolism (GFPT1, INSIG1), and proteins of uncertain function (DOM3Z, TMEM45A). These data provide new insight into the complexity of the physiological response to ER dysfunction.
In LNCaP cells exposed to A23187, the processed form of AIbZIP is detectable in nuclear extracts for at least 24 h following exposure to A23187 (data not shown). Interestingly, A23187, which induces AIbZIP processing, did not upregulate all of the A290HA-regulated transcripts. Both physiological and technical explanations can account for this observation. First, the UPR is a complex response that involves more than just AIbZIP activation. Consequently, it is conceivable that the A290HA-regulated transcripts that are irrelevant to the UPR were actually downregulated by A23187, as appears to be the case for TMPRSS2. Second, some transcripts may be differentially regulated by A23187 over time, and the single time point used in this experiment may have missed increases in mRNA abundance. In fact, in subsequent experiments, we found that the abundance of some transcripts peaked after a shorter (i.e., 8 h) exposure to A23187.
A number of transcripts that were upregulated by A290HA in LNCaP cells are also overexpressed in prostate cancer. Previous studies reported that the transcription factor ID2, the ribonuclease angiogenin (ANG), and the phospholipase PLA2G2A are more abundant in preneoplastic lesions and in adenocarcinoma compared with normal prostate (4, 10, 13). Among these proteins, PLA2G2A is especially intriguing from both a physiological and a pathophysiological perspective. The specific role of this secreted phospholipase is uncertain, but it is believed to play a role in fertilization, and it is known to exert bactericidal activity by virtue of its ability to destroy bacterial cell membranes (12). Furthermore, PLA2G2A releases arachidonic acid from cellular membranes, thereby providing the substrate for biosynthesis of eicosanoids, which are known to regulate cell proliferation.
We observed that AIbZIP is processed to its transcriptionally active form by drugs that lower ER Ca2+ concentrations and/or increase cytoplasmic Ca2+ concentrations. However, the amazing diversity of genes that were upregulated by A290HA in LNCaP cells argues that other physiological disruptions may also lead to AIbZIP processing. How might this occur? It is well established that the activation of ER-associated transcription factors is regulated by proteins, such as BiP (18), that interact with their regulatory domains. We postulate that AIbZIP might interact with different proteins or that multiple physiological pathways might converge on a single AIbZIP-binding protein. As of this report, however, the identities of the proteins that interact with the regulatory domains of AIbZIP and other CREB3-related proteins remain unknown.
The present study was designed to obtain a global portrait of the consequences of AIbZIP activation in LNCaP cells. In this regard, it is of interest to note that three of the validated targets of A290HA (i.e., CREB3, ETS2, and ID2) are transcription factors that could potentially contribute to some of the gene expression changes observed in A290HA-producing cells. However, it is unlikely that CREB3 contributed to the upregulation of the A290HA-regulated genes identified here, since it must be processed by RIP to become transcriptionally active. On the other hand, the fact that CREB3 was induced by both A290HA and A23187 is of considerable interest, since it suggests that AIbZIP and CREB3 are solicited during different phases of the UPR. As for ETS2 and ID2, their target genes in prostate cells are unknown, and additional work will be required to clarify their role in AIbZIP action.
In summary, our studies demonstrate that AIbZIP is a substrate for processing by Golgi proteases and that it is processed to its transcriptionally active form by drugs that deplete ER calcium stores. More importantly, the validated microarray data define a subset of genes, unique to AIbZIP, that function in protein processing in the ER. Moreover, the data reveal that AIbZIP also activates genes implicated in diverse biological pathways, many of which have not been extensively characterized (e.g., ANKH, IKIP, LOX, OASL). These novel discoveries provide insight into the regulation and activity of AIbZIP and will contribute to a better understanding of the physiological role of AIbZIP.
|
GRANTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Article published online before print. See web site for date of publication (http://physiolgenomics.physiology.org).
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONGRANTSREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
