| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Instituto de Investigaciones Biomédicas "Alberto Sols," Consejo Superior de Investigaciones Científicas/Universidad Autónoma de Madrid, 28029 Madrid, Spain
Address all correspondence and requests for reprints to: Mario Vallejo, M.D., Ph.D., Instituto de Investigaciones Biomedicas "Alberto Sols," Calle Arturo Duperier 4, 28029 Madrid, Spain. E-mail: mvallejo{at}iib.uam.es.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
-cells that produce glucagon, 
-cells that produce somatostatin, and cells that produce pancreatic polypeptide, all of these located in the islet periphery. The elucidation of the molecular mechanisms by which these cell types are generated constitutes a topic of intense study and has important implications for our knowledge of the etiopathological processes that lead to the occurrence of diabetes mellitus. During development, the pancreatic bud is formed from the primitive gut epithelium (1). The endocrine cells that compose the islets derive from a subpopulation of cells whose phenotypic fate is segregated from that of exocrine cells by the expression of the basic helix-loop-helix (bHLH) transcription factor neurogenin-3 (2). From this point, the differentiation of each islet cell type is regulated by the coordinated expression of lineage-specific sets of transcription factors. The expression of these sets is organized hierarchically, so that each type of fully differentiated islet cell ends up with a specific combination of transcription factors that regulate the expression of target genes characteristic of each one of the four cellular phenotypes and are, therefore, important for the specification of their function (3, 4).
The promoter of the insulin gene contains several DNA cis-regulatory elements that in ß-cells are coordinately occupied by transcription factors that form unique combinations typical of this type of pancreatic islet cells (5). In this manner, insulin gene expression is regulated with a high degree of cellular specificity that restricts its expression to only one type of pancreatic cells. Prominent among these elements due to their critical importance for insulin gene expression are the so called A- and E-boxes (6). A-boxes bind transcription factors of the homeodomain family, whereas E-boxes provide binding sites for bHLH proteins, such as E47 (also known as Pan1) and Beta2/NeuroD. Homeodomain and bHLH proteins synergistically enhance insulin transcription via direct protein-protein interactions (7, 8, 9, 10, 11, 12). Also important for the regulation of insulin gene expression in ß-cells are the paired-type homeodomain transcription factor Pax6 (13) and the basic leucine zipper protein MafA, which bind cis-regulatory elements known as C-boxes (14, 15, 16).
Notably, studies carried out mostly in rodents have demonstrated that many of the transcription factors that regulate insulin gene expression are also required for the proliferation of islet cell precursors and for the differentiation and survival of ß-cells (17, 18, 19). In addition, the discovery of ß-cell transcription factors in rodents has been highly predictive for the identification of some causative aspects of human diabetes (18, 20, 21, 22), because mutations in the genes that encode some of those transcription factors in humans are associated with different types of diabetes (18, 20, 23, 24, 25). Therefore, knowing the complete spectrum of transcription factors that operate in ß-cells has important implications for the design of protocols to program this type of cells as a putative tool to treat diabetes.
Alx3 is a paired-class aristaless-like homeoprotein expressed in different tissues during embryonic development (26). Although a partial cDNA clone encoding the Alx3 homeodomain was originally isolated from the hamster insulinoma cell line HIT-T15 (27), the expression of Alx3 in pancreatic islets or the target genes that it regulates have not been investigated. In the present study, we present evidence in support of a role for Alx3 on the regulation of insulin gene expression in pancreatic ß-cells.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
). In contrast, we did not detect Alx3 expression in BHK-21, COS-7, or HeLa cells (Figs. 1A and 4C). As a positive control, we confirmed the presence of Alx3 in RC2.E10 cells (Fig. 1A), a cell line derived from embryonic rat neuroepithelium, that express Alx3 constitutively (28). Nuclear localization of Alx3 was confirmed by immunocytochemistry in MIN6 and RIN-1027-B2 cells (Fig. 1, B–E).
|
|
). Double-label immunohistochemistry demonstrated that all ß-cells that synthesize insulin express Alx3, and that Alx3-positive cells located in the periphery of the islets correspond to glucagon- and somatostatin-expressing cells (Fig. 2). Thus, Alx3 appears to be present in all -, ß-, and -cells. In contrast, Alx3 immunoreactivity was not detected in the exocrine pancreas (Fig. 2).
|
). Promoter sequences from the phosphoenolpyruvate carboxykinase (PCK) gene, used as a control, were not amplified from the anti-Alx3 immunoprecipitated chromatin (Fig. 3A), indicating that Alx3 expressed in MIN6 ß-cells occupy the endogenous insulin gene, but not the PCK gene. In addition, we carried out ChIP assays on chromatin prepared from mouse pancreatic islets, and we found that a fragment of the insulin I gene was selectively amplified by PCR with the anti-Alx3 antiserum, but not with control preimmune serum (Fig. 3A), thus confirming that Alx3 is bound to the insulin gene promoter in intact islets in vivo.
|
). The A3/4 element forms a functional enhancer together with the adjacently located E2 element, which binds bHLH proteins (5). To determine whether Alx3 binds to any of those sites, we carried out EMSA with synthetic oligonucleotide probes using nuclear extracts from MIN6 cells.
When a probe corresponding to the A3/4 element was used, several DNA-protein complexes were found, whose sequence specificity was determined by competition with unlabeled A3/4 oligonucleotide added in excess to the binding reaction. Addition of an oligonucleotide of unrelated sequence failed to compete (Fig. 3B). To investigate whether any of the complexes observed contain Alx3, we carried out the protein extract-DNA binding reaction in the presence of the anti-Alx3 antiserum. The addition of this antiserum resulted in the disappearance of one of the upper complexes bound to the A3/4 probe, and in the appearance of a supershifted band, whereas the addition of preimmune serum did not disturb the band pattern (Fig. 3C). The specificity of the electrophoretic mobility supershift observed with the anti-Alx3 antiserum was further confirmed by using the antiserum 253 (29, 30), which specifically recognizes the homeodomain transcription factor Pdx1 (also known as IDX1, IPF1, and STF1), because the A3/4 enhancer is a known target for regulation by this homeoprotein (10, 31). Thus, when EMSAs were carried out in the presence of the anti-Pdx1 antiserum, we observed the disappearance of one of the lower complexes bound to the A3/4 probe and the appearance of a supershifted band. The upper complex affected by the anti-Alx3 antiserum was unaffected by the anti-Pdx1 antiserum (Fig. 3B).
When we used the A2 probe, only a single protein-DNA complex was identified. However, we found that this complex contains Pdx1, but not Alx3, as a major protein component, because it specifically disappeared in the presence of the anti-Pdx1 antiserum but was unaffected by the anti-Alx3 antiserum (Fig. 3D). Thus, these data demonstrate that Alx3 present in MIN6 ß cells is able to recognize the insulin promoter by binding specifically to the A3/4 element.
Alx3 Transactivates the Insulin Promoter
To investigate whether Alx3 exhibits insulin promoter transactivation activity, we cotransfected the ß-cell line MIN6 with the insulin promoter reporter plasmid INSCAT and with the Alx3 expression plasmid pcDNA-Alx3. INSCAT activity in transfected MIN6 cells was found to be relatively high, with values similar to those observed by the control plasmid RSVCAT. In addition, we found that cotransfection of INSCAT with pcDNA-Alx3 did not increase INSCAT activity in these cells (Fig. 4A). Similarly, we observed no increase in chloramphenicol acetyltransferase (CAT) activity when INSCAT was cotransfected in MIN6 cells with expression plasmids encoding known transactivators of the insulin gene such as the homeoprotein Pdx1, or the bHLH proteins E47/Pan1 (32) or Beta2/NeuroD (7) (Fig. 4A). Similar results were obtained when we used the ß-cell line INS-1 (data not shown). Because other homeodomain transcription factors expressed in ß-cells regulate insulin gene expression in combination with bHLH proteins synergistically (10, 11, 33), we investigated whether Alx3 could transactivate the insulin promoter in the presence of E47/Pan1 and Beta2/NeuroD. We found that cotransfection of INSCAT with expression vectors for E47/Pan1 and/or Beta2/NeuroD did not result in enhanced CAT activity in MIN6 cells. However, INSCAT activity increased 2- to 3-fold when an expression vector for Alx3 was combined with both E47/Pan1 and Beta2/NeuroD expression plasmids (Fig. 4A).
Because MIN6 cells exhibit features of a well-differentiated ß-cell phenotype (34), the relatively poor magnitude of this response was interpreted as reflecting the existence of near saturating levels of endogenous Alx3 and of other ß-cell-enriched transcription factors that regulate insulin gene expression (such as Pdx1 and Beta2/NeuroD), and to the possible existence of limiting amounts of transcriptional coactivators such as cAMP response element binding protein-binding protein (CBP)/p300, which is known to interact with them (35). In support of this notion, we found that INSCAT activity was increased 3-fold by cotransfection with an expression vector encoding the coactivator CBP (36) (Fig. 4B). We also observed that, although Alx3 had no effect on its own, it was able to stimulate INSCAT activity in the presence of CBP by approximately 2-fold relative to the activity elicited by CBP alone, and 5- to 6-fold relative to basal INSCAT activity (Fig. 4B).
Thus, to minimize interferences with endogenous Alx3 that may regulate insulin gene expression in ß-cells, we used HeLa cells in subsequent experiments, because they do not express ß-cell-specific transcription factors and we determined that they do not contain detectable levels of Alx3 (Fig. 4C). In HeLa cells, cotransfection of INSCAT with either Alx3 or Pdx1 expression vectors did not result in elevated CAT activity unless an E47/Pan1 expression vector was added, in which case CAT activity was very significantly elevated (Fig. 4D). These results are in agreement with previous studies indicating that homeodomain and bHLH proteins cooperate to regulate insulin gene expression (10, 11, 33), and suggest the existence of functional interactions between Alx3 and E47/Pan1.
Because the E2A3/4 enhancer provides a major site for synergistic interactions between bHLH and homeodomain proteins (5), we tested the relative transactivational activity of Alx3, alone or in combination with E47/Pan1 or Beta2/NeuroD, in directing transcription from this enhancer. For this purpose, we carried out transfections in HeLa cells using the reporter plasmid 5FF-CAT, which bears five multimerized copies of the rat insulin I E2A3/4 enhancer (37). We found that 5FF-CAT activity in transfected HeLa cells was not increased by Alx3 or E47/Pan1, but Beta2/NeuroD elicited a small increase that was slightly potentiated in the presence of E47/Pan1 (Fig. 4E). However, we observed that, although Alx3 lacks transcriptional transactivation activity on its own, it was able to induce transcription from 5FF-CAT in the presence of E47/Pan1 or NeuroD/Beta2, an effect that further increased significantly when both E47/Pan1 and NeuroD/Beta2 were present at the same time together with Alx3 (Fig. 4E). Thus, these experiments show that Alx3 exerts a synergistic activation of the insulin gene promoter in the presence of bHLH proteins expressed in ß-cells. Together with the results from the EMSA experiments described above, these experiments indicate that Alx3 directs transcription of the insulin gene by binding to the E2A3/4 enhancer.
Alx3 Interacts with E47 But Not with Beta2/NeuroD
The synergistic effects observed in the transfection assays prompted us to investigate the possible existence of direct protein-protein interactions between Alx3 and E47/Pan1 and/or Beta2/NeuroD, using glutathione S-transferase (GST) pull-down assays. We found that a GST-Alx3 fusion protein is able to interact with [35S]Met-labeled E47/Pan1, but not with [35S]Met-labeled Beta2/NeuroD (Fig. 5A). Because E47/Pan1 and Beta2/NeuroD are known to act as partners to form heterodimers, we used GST-E47 as a positive control to confirm the ability of Beta2/NeuroD to interact (Fig. 5A).
|
). Thus, it is possible that the functional interaction between Beta2/NeuroD and Alx3 is indirect using E47/Pan1 as a bridging protein interacting both with Beta2/NeuroD and with Alx3. To test this notion, we carried out additional GST pull-down experiments using a GST-Alx3 fusion protein and [35S]Met-labeled Beta2/NeuroD, in the presence or absence of unlabeled E47/Pan1 synthesized in vitro in a rabbit reticulocyte lysate system. The detection of a band corresponding to Beta2/NeuroD only when E47/Pan1 was present in the binding reaction (Fig. 5B) indicates the formation of a trimeric complex and confirms that Beta2/NeuroD interacts indirectly with Alx3 via direct contact with E47/Pan1.
The occurrence of direct interactions between Alx3 and E47/Pan1 in ß cells in vivo was tested by immunoprecipitation from MIN6 cell nuclei using the Alx3 antiserum, followed by Western immunoblot with a specific anti-E47/Pan1 monoclonal antibody. The detection of a band corresponding to E47/Pan1 in samples immunoprecipitated with the anti-Alx3 antiserum, but not in those immunoprecipitated with nonimmune rabbit serum (Fig. 5C), confirmed the Alx3-E47/Pan1 interaction in the nuclei of cells.
Besides the homeodomain, Alx3 contains three proline-rich domains (28), two of them (Pro1 and Pro2) located in the amino-terminal region and one (Pro3) located in the carboxyl-terminal region (Fig. 5D). To determine whether these domains are important for the interactions with E47/Pan1, we carried out GST pull-down assays using a GST-E47/Pan1 fusion protein and either full-length or truncated versions of [35S]Met-labeled Alx3 generated by deletion of residues spanning either the carboxyl or the amino terminus of Alx3. As shown in Fig. 5E, full-length Alx3 is able to interact with GST-E47/Pan1. Increasing deletions of either the amino (Alx357, Alx391, and Alx3143) or the carboxyl terminus of Alx3 (Alx31–228 and Alx31–279), but leaving intact the homeodomain, did not affect dimerization with the GST-E47/Pan1 fusion protein (Fig. 5E). Dimerization was also observed using labeled Alx3143–228, which corresponds to the Alx3 homeodomain (Fig. 5F). In addition, none of the labeled versions of Alx3 was observed to bind to control GST. Thus, these experiments indicate that the segment of Alx3 corresponding to the homeodomain is sufficient for heterodimerization with E47/Pan1.
The Homeodomain of Alx3 Promotes Binding of E47/Pan1 to the E2A3/4 Enhancer
To evaluate directly the binding of Alx3 and E47/Pan1 to the insulin promoter, we carried out EMSA. Because we have previously determined that deletions of segments of the amino or carboxyl terminus of Alx3 do not affect binding to DNA as long as the homeodomain remains intact (28), we used Alx3143 or Alx3143–228 synthesized in reticulocyte lysates to better resolve the protein-DNA complexes.
It is known that aristaless-like homeoproteins form dimers cooperatively upon binding to DNA consensus sequences known as P3 sites, which contain two inverted TAAT motifs separated by three nucleotides (28, 38, 39). In addition, Alx3 binds selectively as a monomer to specific TAAT-containing DNA sequences such as the GFAPT3 site found in the promoter of the glial fibrillary acidic protein gene (28).
We found that Alx3143 binds to the E2A3/4 probe forming two complexes (Fig. 6B), thus resembling the binding pattern generated by the cooperative dimerization observed upon binding to P3 sites (28). Indeed, a comparison with the binding obtained with a GFAPT3 probe demonstrated that the lower complex corresponds to the monomeric form of Alx3 bound to DNA (Fig. 6B).
|
When E47/Pan1 was incubated simultaneously with Alx3143, the two complexes corresponding to the monomeric and dimeric forms of Alx3 were observed unaltered (Fig. 6C). In addition, a band with lower electrophoretic mobility appeared, an effect that was not observed when Beta2/NeuroD was used instead of E47/Pan1 (Fig. 6C). A similar effect was found when we used Alx3143–228, which only spans the Alx3 homeodomain (Fig. 6D). These results indicate that Alx3 and E47/Pan1 can interact cooperatively forming a ternary complex on the DNA, and that the homeodomain of Alx3 is sufficient for the occurrence of this interaction.
The Homeodomain of Alx3 Is Sufficient to Promote E47/Pan1-Dependent Transactivation
In earlier studies, using promoter elements on which Alx3 transactivation activity does not depend on direct interactions with bHLH proteins, we found that Pro1 and Pro2 domains are necessary for transactivation, but Pro3 is dispensable, and that the region spanning the homeodomain is unable to transactivate on is own (28). Thus, we tested for the relative contributions of these domains to the transcriptional activity of Alx3 from the insulin promoter, which depends on interactions with E47/Pan1. For this purpose, we cotransfected HeLa cells with the 5FF-CAT reporter plasmid and expression plasmids encoding E47/Pan1 or truncated versions of Alx3 lacking the amino-terminal or the carboxyl-terminal domains.
As mentioned earlier, Alx3 is unable to transactivate the 5FF-CAT reporter unless E47/Pan1 is present (see Fig. 4E). In these experiments, we found that when residues 1–143 of Alx3 were deleted (Alx3143), eliminating the entire amino-terminal region to a position next to the homeodomain, the CAT activity elicited by both proteins was reduced to approximately 80% of that obtained with the full-length proteins, indicating that the Pro1 and Pro2 domains of Alx3 are not essential for transactivation in the presence of E47/Pan1 (Fig. 7A). To test whether the Pro3 domain located in the C-terminal region of Alx3 is essential for transactivation, we cotransfected an expression vector encoding a truncated version of Alx3 spanning residues 1–228 (Alx31–228), in which the Pro1 and Pro2 domains, as well as the homeodomain, remain intact. We found that Alx31–228 yields levels of CAT activity that are also about 80% of that obtained with full-length Alx3 in the presence of E47/Pan1, indicating that the Pro3 domain is not essential for activity when the other two domains are intact (Fig. 7A). Finally, we cotransfected an expression vector encoding Alx3143–228, a truncated Alx3 protein in which both the amino- and the carboxyl-terminal regions were deleted leaving intact the homeodomain. We found that the activity elicited by Alx3143–228 is about 20% of that elicited by full-length Alx3 (Fig. 7A).
|
and 7B). Because our previous work demonstrated that the isolated Alx3 homeodomain lacks transcriptional transactivation activity (28), these experiments suggest that the residual activity observed when Alx3143–228 and E47/Pan1 are present together is due to the activity of the E47/Pan1 interacting with the Alx3 homeodomain at the insulin promoter.
To test this notion, we cotransfected the 5FF-CAT reporter plasmid with expression vectors encoding the Alx3 homeodomain (Alx3143–228) and with expression vectors encoding truncated versions of E47/Pan1 corresponding to amino-terminal deletions to residues 91, 334, or 549, that lack one or two of the activation domains, but leave intact the bHLH domain (Fig. 7B). Cotransfection of these plasmids with the plasmid encoding Alx3143–228 resulted in the generation of background levels of CAT activity indistinguishable from those observed when Alx3143–228 was transfected alone. Thus, as predicted, stimulation of 5FF-CAT activity in the presence of the Alx3 homeodomain was dependent on the integrity of the transactivation domains of E47/Pan1.
In summary, all of these experiments taken together support the notion that the Alx3 homeodomain interacts with E47/Pan1 on the insulin promoter, and that full transcriptional transactivation is contributed by Alx3 proline-rich domains and E47/Pan1 transactivation domains, as well as by Beta2/NeuroD complexed with E47/Pan1.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
In the present study, we have identified the insulin gene as the first known target for regulation by Alx3. This notion is based on the expression of Alx3 in pancreatic islet cells, on DNA binding and ChIP assays showing specific interactions of Alx3 with the insulin promoter in ß-cells, and on functional transient transfection assays showing that Alx3 transactivates the insulin gene by interacting with bHLH proteins.
Alx3 Interacts with E47/Pan1 to Yield Synergistic Transactivation of Insulin Gene Transcription
In addition to Alx3, the homeodomain proteins Pdx1 and Lim1.1 have been shown to interact with the class A bHLH protein E47/Pan1 on the insulin E2A3/4 enhancer. In contrast, interactions with the class B bHLH protein Beta2/NeuroD are selective, because only Pdx1 appears to interact directly with Beta2/Neuro (10). Our data indicate that the synergistic transactivation observed between Alx3 and E47/Pan1 could be due to interactions at two different levels. On the one hand, direct protein-protein interactions with the Alx3 homeodomain bound to the E2A3/4 enhancer, bringing in addition Beta2/NeuroD to the transcriptionally active complex, and on the other hand functional interactions of the transactivation domains, perhaps by recruiting coactivator proteins that could not interact stably with each factor in isolation.
Alx3 in pancreatic ß-cells recognizes the A3/4 site in the promoter of the insulin gene. Although Alx3 binds the A3/4 site preferentially as a dimer, a less intense complex corresponding to the monomeric form was also observed in EMSA carried out with the synthetic version of the protein (see Fig. 6). The A3/4 element contains the sequence TAATCTAATTA, which resembles a P3 site to which paired class homeoproteins bind preferentially as dimers in a cooperative manner (28, 38, 39, 45). In addition, the palindromic motif TAATTA within this sequence serves as a preferred site for binding by Alx3 (28). Thus, these data suggest that Alx3 dimerizes cooperatively upon binding to the A3/4 site. In addition, the ability of Alx3 to interact directly with E47/Pan1 appears to favor the cooperative binding of this bHLH, which in turn would bring Beta2/NeuroD to form a functionally active multiprotein complex assembled on the E2A3/4 enhancer. Thus, these data reinforce the notion that synergistic activation of the insulin promoter occurs as a consequence of cooperative DNA binding that promotes recruitment of multiple activators (10).
Functional interactions at the level of the transactivation domains also appear to be important for synergistic enhancement of insulin transcription. We have previously determined that the integrity of the Pro1 and Pro2 domains of Alx3 is required for transcriptional transactivation from a generic P3 site, thus indicating that the Pro3 domain is unable to function on its own (28). However, in the present study, we have observed that when Alx3 is acting from the E2A3/4 element in the presence of E47/Pan1, the Pro3 domain is still able to synergize with this bHLH protein, whose activity is in turn dependent on the integrity of the transactivation AD1 domain. Interestingly, E47/Pan1 is able to stimulate transcription, although relatively weakly, when all the proline-rich domains of Alx3 have been deleted, further supporting the notion that the homeodomain of Alx3 is sufficient to recruit E47/Pan1 to a transcriptionally active complex assembled on the insulin promoter. The mechanism by which the transactivation domains of Alx3 and bHLH proteins synergize have not been explored in the present study, but previous work (35) and our own data (see Fig. 4B) suggest that interactions with coactivators such as CBP/p300 may be involved.
Finally, it is possible that additional mechanisms involving different types of proteins contribute to the synergistic transactivation of Alx3 and E47/Pan1. For example, recruitment of coactivators (37) or interactions with HMG-type proteins have been shown to be important in the case of the synergism between Pdx1 and E47/Pan1 (10). Also, it is possible that Alx3 can alter chromatin conformation, as has been shown to be the case for HNF1 (46), to favor the binding of other transcription factors or coactivators. The study of these possibilities remains the subject for future research.
Regulation of Insulin Gene Transcription by Alx3 in Pancreatic ß-Cells
One important concept emerging from genetic studies on aristaless-like genes is the existence of a relatively high degree of functional redundancy among them due to overlapping functions that only become evident when double or triple mutant mice are generated (38, 43, 44). It is possible that redundancy also exists between Alx3 and other proteins that regulate the expression of the insulin gene in ß-cells, because Alx3 appears to function like other homeodomain transcription factors that cooperate with bHLH proteins on the E2A3/4 elements of the insulin promoter, such as Pdx1 or the LIM-domain proteins Lmx1.1 and Lmx1.2 (10, 11, 31, 47). However, ongoing studies in our laboratory using Alx3 mutant mice (44) point to the existence of a mild glucose homeostasis phenotype, lending support to the notion that Alx3 participates in the control of ß-cell function.
Pdx1 has emerged as a major regulator of insulin gene transcription in both humans and rodents (11, 48, 49, 50). However, although clearly important for ß-cell function and survival, it appears that Pdx1 is not specifically required for insulin gene transcription, because insulin gene expression can occur in the absence of Pdx1 (17, 51, 52). These findings have been interpreted on the basis that other homeodomain transcription factors present in ß cells can compensate for the loss of Pdx1 by acting via a similar mechanism, i.e. interaction with bHLH proteins at the level of specific insulin promoter elements such as E2A3/4 (5). It is possible that Alx3 could substitute for Pdx1, as the LIM-homeodomain factors Lmx1.1 and Lmx1.2 may also do (5). Thus, loss of a single homeodomain transcription factor appears not sufficient to compromise insulin gene transcription, because this is probably the result of the coordinate activity of several of these factors that can compensate for one another. In this regard, studies aimed at determining functional interactions between Alx3 and Pdx1 and their relative contributions to regulation of insulin gene transcription are currently being carried out in our laboratory.
As mentioned earlier, some of the functions of Alx3 and other aristaless-related genes have been identified after generating mice with combined mutant alleles. On the other hand, it has been shown that combined heterozygous mutations in certain genes expressed in ß-cells that have no effect on pancreatic function as single mutants, such as Hnf1 or Hnf3ß, can lead to altered ß-cell function resulting in decreased insulin gene transcription or aggravation of diabetes caused by haploinsufficiency of Pdx1 (53). Thus, although not known at the present time, it is possible that mutations in Alx3 in combination with mutations in other diabetes-related genes may alter pancreatic ß-cell function as a consequence of decreased insulin synthesis or secretion. In this regard, it is important to note that Alx3 is functionally related to Beta2/NeuroD, Pdx1, and Hnf1, because it acts in combination with Beta2/NeuroD (this study) and regulates insulin gene transcription upon binding to the same promoter element used by Pdx1 and Hnf1. These three genes are causatively related to maturity onset diabetes of the young (3). Further studies will determine whether mutations in Alx3 are directly or indirectly related to the development of diabetes.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
CAT activity was measured by a solution assay (58) 48 h after transfection exactly as described (56). Values were normalized to those yielded by a Rous sarcoma virus enhancer reporter plasmid (RSVCAT) transfected in parallel in MIN6 cells, or to CMVCAT in HeLa cells. All of the values are expressed as mean ± SEM of at least three independent experiments carried out in duplicate.
Plasmids
INSCAT encodes a CAT reporter gene driven by a fragment of the rat insulin I gene promoter spanning nucleotides –410 to +34 (59). 5FF-CAT is a reporter plasmid bearing five multimerized copies of the rat insulin I E2A3/4 enhancer driving CAT expression (provided by Dr. Melissa Thomas, Massachusetts General Hospital, Boston, MA) (60). Construction of the expression plasmids containing full-length and deleted versions of Alx3 cloned into pcDNA3 has been described (28). The plasmid pcDNA-Pdx1 contains a full-length cDNA encoding rat IDX1 (61) cloned into the HindIII and NotI sites of pcDNA3. The plasmid pCMV-Beta2 (provided by Dr. J. Ferrer, Hospital Clinic, Barcelona, Spain) contains a full-length cDNA encoding Beta2/NeuroD (7). The E47/Pan1 expression vector used in CAT reporter assays corresponds to pCMV-Pan1 described by Nelson et al. (32).
For GST pull-down experiments and protein synthesis in reticulocyte lysates, we used GST-E47, pcDNA3-E47, and pcDNA3-Beta2 provided by Dr. Amparo Cano (Instituto de Investigaciones Biomédicas, Madrid, Spain) (62), as well as GST-Alx3 and pcDNA3-Alx3 (full-length and deleted versions) previously described (28).
For experiments designed to test deleted versions of E47/Pan1, we used the plasmid pZeoSV2 (Invitrogen Life Technologies, Carlsbad, CA) containing cDNAs encoding amino-terminal deletions to residues 91, 334, or 549. These were kindly provided by Dr. Amparo Cano (Instituto de Investigaciones Biomédicas, Madrid, Spain).
Western Immunoblot
Nuclear extracts from cells growing in 60-mm dishes were prepared as described (63), and proteins were resolved by SDS-PAGE and blotted onto a nitrocellulose membrane. Alx3 immunoreactivity was detected with a rabbit polyclonal primary antiserum (28) (1:5000 dilution), followed by incubation with a goat antirabbit peroxidase-conjugated secondary antibody (1:10,000 dilution) (Bio-Rad, Hercules, CA). Immunoreactive bands were visualized using an enhanced chemiluminescence detection system (ECL; Amersham Biosciences, Piscataway, NJ).
Immunocytochemistry
Immunocytochemistry on MIN-6 and RIN-1027-B2 cells was carried out basically as described (56). Cells plated into 35-mm tissue culture dishes were fixed in 4% paraformaldehyde in PBS for 5 min, washed in PBS, and permeabilized with methanol for 2 min at –20 C. After blocking with normal goat serum for 1 h, cells were incubated overnight with the anti-Alx3 antiserum (1:4000 dilution) at 4 C. Immunodetection was carried out with a secondary biotinylated goat antirabbit antiserum (Bio-Rad) using nickel-intensified immunoperoxidase staining with a Vectastain ABC kit (Vector Laboratories, Burlingame, CA).
Immunohistochemistry
Adult pancreata from Wistar rats (200–250 g body weight) were fixed by transcardial perfusion with 4% paraformaldehyde, removed, postfixed overnight, and cryoprotected in PBS containing 20% sucrose. Cryostat sections (20 µm) were cut and kept at –80 C until used. Sections were brought to room temperature, permeabilized with methanol for 2 min at –20 C, treated with 5% normal goat serum, and then incubated overnight with the anti-Alx3 (1:4000 dilution) antiserum at room temperature. Immunodetection was carried out with a secondary biotinylated goat antirabbit antiserum using nickel-intensified immunoperoxidase staining. For insulin staining, sections were incubated with guinea pig antihuman insulin (Linco Research, St. Charles, MO; 1:100 dilution), and immunodetection was carried out with a secondary biotinylated goat anti-guinea pig antiserum using immunoperoxidase staining. For glucagon staining, we used a guinea pig antibody (Linco Research; 1:100 dilution), detected with a secondary biotinylated goat anti-guinea pig antiserum and immunoperoxidase staining. For somatostatin staining, we used a mouse monoclonal antibody (Acris Antibodies, Hiddenhausen, Germany; 1:100 dilution), detected with a secondary biotinylated horse antimouse antiserum and immunoperoxidase staining.
For two-color dual antigen immunostaining, Alx3 was first detected using nickel-enhanced immunostaining, which yields a dark blue color. Sections were then washed, and insulin, glucagon, or somatostatin immunohistochemistry was carried as described above, using immunoperoxidase staining without nickel, which yields a brown color.
Experimental protocols involving Wistar rats for the preparation of sections for immunohistochemistry were approved by the institutional committee on research animal care, and meet the requirements of current Spanish and European Community legislation.
EMSAs
EMSAs were carried out with either nuclear extracts (63) of MIN-6 cells or with transcription factors synthesized in vitro using a rabbit reticulocyte lysate system (Promega, Madison, WI). Protease inhibitors (Complete protease inhibitor cocktail; Roche, Basel, Switzerland) were added to nuclear extracts, and protein concentrations were determined by the Bio-Rad protein assay. Synthetic complementary oligonucleotides with 5'-GATC overhangs were annealed and labeled by a fill-in reaction using [-32P]dATP and Klenow enzyme. Binding reactions were carried out in the presence of 20,000 cpm of radiolabeled probe (
6–10 fmol) in a total volume of 20 µl containing 2 µg poly(dI·dC), 20 mM HEPES (pH 7.9), 70 mM KCl, 2.5 mM MgCl2, 1 mM dithiothreitol, 0.3 mM EDTA, and 10% glycerol. The sequences of the oligonucleotides used are as follows (coding strand): A2, 5'-GATCCGAGCCCTTAATGGGCCAAA-3'; A3/4, 5'-GATCCTGTTAATAATCTAATTACCCTAGA-3'; E2A3/4, 5'-GATCCATCAGGCCATCTGGCCCCTTGTTAATAATCTAATTACCCTAGA-3'; GFAPT3, 5'-GATCCTTTGCCAATTAGTGTGACA-3'.
ChIP
ChIP assays were carried out basically as descried (64). Cross-linked chromatin from mouse islets was generously provided by Jorge Ferrer (Hospital Clinic, University of Barcelona, Barcelona, Spain). MIN6 cells were treated with 1% formaldehyde for 15 min at room temperature and the cross-linked protein-DNA complexes were isolated. After sonication, chromatin was incubated with anti-Alx3 antiserum or control normal rabbit serum. Antibody-protein-DNA complexes were isolated by incubation with protein A-Sepharose. To detect bound DNA, PCR was carried out using oligonucleotide primers that amplify a fragment of the mouse insulin gene I spanning nucleotides –360 to –12. PCR conditions were as follows: 95 C for 5 min, followed by 30 cycles of 94 C for 30 sec, 55 C for 30 sec, and 72 C for 30 sec, after which a 5-min incubation at 72 C followed. The sequence of the oligonucleotide primers are as follows: forward, 5'-ATATTAGGTCCCTAACAACTGCAGT-3'; and reverse, 5'-TACTGGATGCCCACCAGCTTTATAG-3'. To ensure specificity, the 3'-end of the forward oligonucleotide corresponds to a 5-nucleotide insertion absent in the insulin II gene. For amplification of the mouse insulin II gene, we used the following oligonucleotides: forward, 5'-CAAAGATACTAGGTCCCCAACTG-3'; and reverse, 5'-CCACTACCTTTATAGACCAAAGC-3'. PCR conditions were identical with those used for the insulin I sequences. In this case, the 3' end of the forward oligonucleotide corresponds to a 3-nucleotide insertion absent in the insulin I gene. In addition, specificity of the amplified fragments was checked by digestion with either AatII or PstI, both of which only cut the fragment amplified from the insulin II gene. As a control, we used promoter sequences from the PCK gene (nucleotides –434 to –96) as indicated by Cissell et al. (65). The sequence of the PCK oligonucleotide primers are as follows: forward, 5'-GAGTGACACCTCACAGCTGTGG-3'; and reverse, 5'-GGCAGGCCTTTGGATCATAGCC-3'. PCR conditions were as follows: 95 C for 2 min, followed by 28 cycles of 95 C for 30 sec, 61 C for 30 sec, and 72 C for 30 sec, after which a 5-min incubation at 72 C followed.
In all cases, PCR products were run on a 1% agarose gel, stained with ethidium bromide, and photographed.
GST Pull-Down Assays
Full-length or truncated versions of [35S]Met-labeled Alx3 were generated by in vitro translation using a rabbit reticulocyte lysate system (Promega) and cDNAs cloned into pcDNA3 (28). [35S]Met-labeled E47/Pan1 and Beta2/NeuroD were similarly generated using plasmids pcDNA3-E47 and pcDNA3-Beta2 (62). These labeled proteins were incubated with recombinant GST-Alx3, GST-Alx3143–228 (28), or GST-E47 (62) expressed in bacteria and bound to glutathione-Sepharose beads (Amersham Biosciences, Little Chalfont, UK). Incubations were carried out at 4 C for 1 h in buffer containing 20 mM sodium phosphate (pH 7.9), 150 mM KCl, 0.5 mM EDTA, 0.02% Triton, 1 mM phenylmethylsulfonyl fluoride, and 2 µg/ml aprotinin. After extensive washing, the bead-bound proteins were denatured, resolved by SDS-PAGE, and detected by autoradiography.
Immunoprecipitation
Nuclei from MIN6 cells (2 x 107) were prepared as described by Schreiber et al. (63) by incubation of cells in hypotonic buffer followed by vigorous agitation in the presence of 0.06% Igepal CA-630 (Sigma-Aldrich, St. Louis, MO). Nuclei were then lysed in buffer containing 50 mM Tris-HCl (pH 8.0), 150 mM NaCl, 5 mM EDTA, 0.5% Igepal CA-630, and protease inhibitor cocktail (Roche). Nuclear proteins were incubated overnight with either nonimmune rabbit serum or with anti-Alx3 antiserum at 4 C, and then protein A-Sepharose was added for another 2 h. After this, the samples were centrifuged, washed, and resolved by SDS-PAGE. Proteins were blotted onto a nitrocellulose membrane, and E47/Pan1 immunoreactivity was detected by Western immunoblot using a specific monoclonal antibody (BD Pharmingen, San Diego, CA; 1:500 dilution).
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
M.M. and M.V. have nothing to declare.
First Published Online July 6, 2006
Abbreviations: bHLH, Basic helix-loop-helix; CAT, chloramphenicol acetyltransferase; CBP, cAMP response element binding protein-binding protein; ChIP, chromatin immunoprecipitation; GST, glutathione S-transferase; PCK, phosphoenolpyruvate carboxykinase.
Received for publication November 24, 2005. Accepted for publication June 26, 2006.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONRESULTSDISCUSSIONMATERIALS AND METHODSREFERENCES |
|---|
function results in defective insulin gene transcription and impaired metabolism-secretion coupling in a pancreatic ß-cell line. EMBO J 17:6701–6713[CrossRef][Medline]/MODY3 gene in Japanese subjects with early- and late-onset NIDDM. Diabetes 46:1504–1508 directs nucleosomal hyperacetylation to its tissues specific transcriptional targets. Mol Cell Biol 21:3234–3243, and Hnf-3ß. Proc Natl Acad Sci USA 99:3818–3823
