| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Departments of Pathology and Microbiology and NYU Cancer Institute, New York University School of Medicine, New York, New York, USA
1Correspondence: Department of Pathology, New York University School of Medicine, 550 1st Ave. MSB548, New York NY 10016, USA. E-mail: del243{at}med.nyu.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: transcription factors • domain accretion • nuclear accumulation • DNA binding factor • nematodes • tyrosine phosphorylation • SH2 domain
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, 2
). The family members are related both structurally and functionally. Structurally, these proteins retain a related domain structure, including conserved coiled-coil, DNA binding, and SH2 domains, plus a site for regulated tyrosine phosphorylation and a less conserved transactivation domain. Functionally, these transcription factors exist as latent cytoplasmic proteins prior to activation by tyrosine phosphorylation in response to receptor engagement, usually mediated by members of the Janus kinase (Janus-activated kinase, or JAK) family (3)
. Reciprocal phosphotyrosine-SH2 interactions of phosphorylated proteins leads to multimer assembly, nuclear translocation, DNA binding at specific promoter/enhancer elements, and transcriptional regulation of target genes. The resulting gene expression pattern can have broad effects on development, cell growth, malignancy, and homeostasis (1)
.
The mammalian STAT family consists of seven members that are highly related to one another, are involved largely in responses to cytokines and growth factors, and exert diverse effects on a number of biological processes including immunity, hematopoiesis, inflammation, and development (1
, 4
5
6)
. Beyond mammals, STAT proteins have been characterized as ligand-dependent transcription factors in other vertebrates, in invertebrates, and in myxamoebae (7
8
9)
. For instance, the STAT protein in Drosophila sp. belongs to a classical cytokine-like JAK-STAT pathway, activated by the cytokine-like unpaired ligand operating on the domeless receptor to stimulate the hopscotch tyrosine kinase to phosphorylate the STAT92E protein (10)
. The pathway regulates numerous biological processes, including sex determination, stem cell renewal in the male germline, border cell migration and stalk cell development in oogenesis, embryonic segmentation, tracheal development, larval hematopoiesis, and ommatidial rotation (6
, 10
, 11)
.
A more divergent STAT pathway has been characterized in the slime mold, Dictyostelium discoideum. Several STAT proteins are expressed in these protists, and they regulate gene expression changes required for cell growth, differentiation, and chemotaxis (12)
. Although these proteins are highly diverged from their mammalian counterparts, sharing only 15–19% overall amino acid sequence similarity with human STATs, they nonetheless possess characteristic STAT domains and functions, including a DNA binding domain, an SH2 domain, a site for regulated tyrosine phosphorylation, and signal-induced nuclear accumulation. They serve largely as transcriptional repressors, and the absence of a JAK orthologue, cytokine receptor, or cytokine-like genes in Dictyostelium sp (13)
, along with the divergent nature of their extracellular small molecule activators, suggests they comprise a distinct branch of the STAT family.
We have found that the nematode Caenorhabditis elegans contains a STAT orthologue, STA-1 (for signal transducer and transcriptional activator), that regulates larval development in concert with the DAF-7/TGF-ß pathway (14)
. Here, we report gene expression and biochemical characterization of this STAT protein family member. With the exception of an apparently absent amino-terminal domain found in most STAT orthologues, this protein retains structural as well as biochemical features of mammalian STATs, including site-specific tyrosyl phosphorylation site, regulated nuclear translocation, and sequence-specific DNA binding. Promoter GFP studies suggest that STA-1 is widely expressed in pharynx, intestine, body muscle, and neuronal cells throughout most of its life cycle, but derivation of sta-1-null animals demonstrated that it is superfluous for viability. However, unique patterns of subcellular localization of STA-1 protein, particularly in the nervous system, suggest that it participates in regulated processes.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
; E. coli strain HB101 was used as a food source, and except where noted otherwise, worms were cultured at 20°C. The strains used in this study were obtained from the Caenorhabditis Genetics Center (Minneapolis, MN, USA) or were gifts from S. Clark (New York University School of Medicine, NY, USA).
Identification and cloning of sta-1
To identify a C. elegans STAT gene, the amino acid sequence of human STAT1 was used to search a C. elegans expressed sequence tag (EST) database (16)
by the TBLASTN algorithm (17)
. The two highest scoring matches, yk438c8 and yk494e9, which share the same 3'-tag sequences, showed significant sequence homology in the coiled-coil domain. Clone yk494e9, with a larger insert, was kindly provided by Y. Kohara (National Institute of Genetics, Japan), and was completely sequenced with a series of synthesized oligonucleotide primers. We obtained a full-length cDNA by using 5' RACE with the MarathonTM cDNA Amplification Kit (Clontech, CA, USA), according to manufacturer’s instructions, and designated this clone STA-1.
Isolation of mutant alleles
To generate worms defective in sta-1, a Tc1-mediated transposon insertion and subsequent deletion strategy (18)
were used, with modifications to avoid making a large frozen library and to increase screening sensitivity. Briefly, a library containing 200 independent 6 cm cultures of the mut-2 mutator strain MT3126 was established and kept at 15°C. DNA lysates were individually prepared and were not subject to any pooling. To identify Tc1 insertions at the sta-1 locus, screenings were performed using sets of nested sta-1-specific polymerase chain reaction (PCR) oligonucleotide primer pairs in combination with sets of nested Tc1-specific primer pairs. After identification of cultures with putative Tc1 insertions at the sta-1 locus, individual worms harboring this insertion were isolated by sib selection and were subsequently confirmed by single worm PCR. To isolate deletions that can accompany transposon excisions, we repeated PCR screenings as described above on subcultures of Tc1-insertion animals, except we used sets of nested sta-1-specific primers only, resulting in isolation of a deletion allele, qa5800. Both transposon insertion sites and genomic deletion sequences were determined by direct sequencing of PCR products amplified from genomic DNA. To purify the genetic background and to eliminate the mut-2 allele, qa5800 was extensively outcrossed to the wild-type Bristol N2 strain. A second sta-1 allele ok587 was isolated following a request to the C. elegans Gene Knockout Consortium (Oklahoma Medical Research Foundation, OK, USA) from a chemically mutagenized library. The genomic deletion of ok587 allele was determined by direct sequencing of PCR products and ok587 was extensively outcrossed to N2. Details of the screenings and primer sequences are available on request.
Generation of transgenic animals
To create Psta-1::GFP transgenic lines, the 5.6 kb intergenic region upstream of sta-1 plus sequences coding for the first 10 amino acid residues was amplified using ExpandTM Long Template PCR System (Boehringer Mannheim, Mannheim, Germany) and subcloned inframe into the BamHI/KpnI sites of pPD95.81 and pPD96.04 (kindly provided by A. Fire, Stanford, CA, USA). Cloning junctions were sequenced for verification. A 1.8 kb truncated version was created by cutting the 5.6 kb version with BamHI/ClaI and subsequent blunt-end ligation. The above constructs were injected into lin-15(n765) worms at 60 ng/µl, together with coinjectible marker lin-15(+) pL15.EK (kindly provided by S. Clark) at 30 ng/µl. Several independent transgenic lines carrying lin-15(+) extrachromosomal arrays were generated from each construct and scored for GFP expression pattern.
Generation of and use of anti-STA-1 antibodies
The putative transcriptional regulatory domain (corresponding to amino acids 568–703) at the carboxyl terminus of STA-1 was amplified by PCR from a cDNA construct using forward primer 5'-CGCGGATCCCTGCGTTTCTTCGAGTCC-3' and reverse primer 5'-GGAATTCAGGTAATTGGGGGGACTG-3'. The PCR product was subcloned inframe into pGEX-4T3 (Pharmacia, Uppsala, Sweden). E. coli strain XL-1 Blue was used for glutathione S-transferase (GST) fusion protein production. Purified fusion protein was used for immunization of two rabbits (Zymed, San Francisco, CA, USA). The antiserum was affinity-purified and used for Western blot analysis and immunofluorescence staining. Worm lysates were made by pelleting worms in M9 buffer, resuspending in SDS sample buffer, boiling for 10 min, and loading onto SDS-PAGE immediately. Coimmunoprecipitation and Western blot analysis was carried out using standard protocols.
Immunofluorescence staining and microscopy
Worms were fixed in 1% formaldehyde for 30 min at 4°C and stained with 1 µg/ml rabbit polyclonal STA-1 antiserum, as described previously (19)
. Nuclei were visualized by nucleic acid staining using a 1:1000 dilution of TO-PROTM-3 (Molecular Probes, Eugene, OR, USA). Immunostained worms were mounted in Vectashield mounting medium (Vector Laboratories, Burlingame, CA, USA). Confocal laser scanning microscopic images were collected on a Leica TCS-NT or Zeiss LSM-510. Regular epifluorescence microscopy was performed on a Nikon Eclipse TE300 equipped with DIC and a SPOTTM digital CCD camera (Diagnostic Instruments, Sterling Heights, MI, USA). Image processing and analysis were performed using Adobe Photoshop with ImageReady 7.0 (Adobe, San Jose, CA, USA) and the ImageJ software package (NIH, http://rsb.info.nih.gov/ij/).
Cell culture and reporter gene assays
Cos7 cells and human embryonic kidney 293T cells were maintained in Dulbecco’s modified Eagle medium (DMEM) supplemented with 10% calf serum and antibiotics and were transfected by standard methods. Electrophoretic mobility shift assays (EMSA) were performed with protein extracts prepared from transiently transfected HEK293T cells, as described (20)
. In brief, a double-stranded 32P-labeled DNA oligonucleotide containing a high-affinity STAT binding site (GAS) (21)
was incubated with 3 µg nuclear protein, fractionated on a nondenatured polyacrylamide gel, and autoradiographed.
Whole cell lysates from transfected cells were prepared in luciferase lysis buffer (Promega, MI, USA). Cos7 cells were cultured in 6-well dishes and were transiently transfected with either empty vector or Gal4-STAT fusion constructs, along with a luciferase-based minimal promoter construct containing 5x Gal4 UAS elements (gift of T. Hoey, Tularik, CA, USA). To normalize transfection efficiencies among samples, ß-galactosidase (ß-gal) activities derived from a cotransfected cytomeglovirus (CMV)-lacZ construct were measured as described (22)
. 16–18 h after transfection, cells were lysed in 100 µl, and 10 µl was used for luciferase assay (Promega) and 5 µl for ß-gal analysis (Tropix, Foster City, CA, USA).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
identified two candidate EST clones, yk438c8 and yk494e9, as potential STAT genes. The two clones share the same 3'-tag sequences, indicating that they were likely derived from the same gene. Sequence analysis of clone yk494e9 with the longer insert, supplemented with additional sequence obtained by 5'-RACE, revealed a cDNA encoding a protein with significant homologies to known STAT proteins (Fig. 1
).
|
The full-length cDNA was 2344 bp prior to a poly(A) stretch, and genomic comparison demonstrated that it was encoded by eight exons and a SL1 spliced leader (Fig. 1A
). The single open reading frame encoding an 80 kDa protein showed significant overall STAT homology, including SH2 and DNA binding domains (Fig. 1B
). In addition, a splice variant was detected that encoded an extra three amino acids after residue 603. Overall, the C. elegans protein showed 25% identity to Drosophila STAT, 23% to mammalian STAT5 and STAT1, and 18–22% to other STAT family members (Table 1
). Greater levels of identity were observed among individual domains, with the SH2 domain of the nematode protein showing 29% identity to that of other STAT proteins and the DNA binding domain showing 37% identity to that of STAT5b (although only 
22% to that of STAT1); the linker domain showed 36% identity to that of STAT1, but only 25% to that of STAT5b. Key residues and motifs were also highly conserved, such as the STAT-type SH2 domain signature motif, FLLRFS (aa 499–504), and a tyrosine phosphorylation motif GYIQ (aa 587–590) (23)
. These homologies led us to conclude that the nematode protein is a bona fide STAT family member, which we have designated STA-1. Other gene segments with homology to STAT were also recognized, but none fulfilled the criteria to be full-fledged STAT proteins (data not shown). sta-1 corresponds to the genetic locus Y51H4A.17 in WormBase (http://www.wormbase.org), and the sequence has been deposited in GenBank® as accession AY943817.
|
To examine the possible evolutionary relationships between STA-1 and some known STATs, including seven human STATs, three slime mold STATs, and one fruit fly STAT, corresponding protein sequences were aligned by the ClustalX program followed by the construction of a neighbor-joining phylogram (Fig. 1C
) (24)
. As the most ambiguous parts of the multiple alignment are usually concentrated around gaps, positions where any of the sequences have a gap were excluded from phylogenetic analysis. As might be expected, the three Dictyostelium STATs are more closely related to each other and form an outgroup of the STAT family. The main branch of the family consists of two diverged groups. C. elegans STA-1, together with Drosophila STAT and human STAT5 and STAT6, constitute one of the more ancient groups, whereas human STAT1, STAT2, STAT3, and STAT4 potentially arose from a more modern duplication event (Fig. 1C
).
STA-1 lacks the conserved STAT amino-terminal protein interaction domain
Sequence analysis indicated that STA-1 lacked the amino-terminal domain present in other STAT proteins (Fig. 1B
). Verification that our 2.3 kb cDNA represented a full-length clone was obtained by Northern blot of total RNA from mix-staged worms, which revealed a 2.3 kb transcript (data not shown). Moreover, genomic analysis of the sta-1 locus, which consisted of 8 exons spanning 7 kb (Fig. 1A
), did not identify any additional 5' exons. Sequences 6 kb upstream of sta-1 failed to reveal any region encoding amino-terminal domain-like sequences prior to the next predicted gene on the same DNA strand, y51h4a.15, an unrelated gene whose expression is supported by the presence of a matching EST clone (WormBase). Sequence comparison to the related nematode species C. briggsae (25)
identified CBP03337 as encoding a STA-1 homologue with the same domain structure. This homolgoue also lacked an amino-terminal domain (data not shown).
sta-1 is widely expressed during multiple developmental stages
To gain insight into sta-1 regulation and biology, we determined its pattern of expression. Genomic fragments containing putative 5' promoter elements and a portion of exon 1 were fused in frame with GFP and a nuclear localization sequence, and several independent transgenic lines were examined for expression. As shown in Fig. 2
, the sta-1 promoter::NLS::GFP reporter gene was widely expressed in a variety of cells and tissues, during most life stages. Specifically, it was expressed at higher levels in pharynx (Fig. 2A
) compared to expression levels in other tissues. GFP was also apparent in the entire intestine, marking all nuclei. Fluorescence intensities of GFP appeared weaker in the anterior intestine than in the posterior intestine, but this effect was due largely to the anterior part of the intestine being obscured by the gonad, where no reporter gene expression was observed. GFP expression was also observed in body muscles as well as in most of the nervous system.
|
Among neuronal cells, GFP expression was readily observed in head ganglia, particularly in the posterior ganglia, including the small dorsal ganglion, two lateral ganglia, and ventral ganglion (Fig. 2A
). Similarly, the sta-1 promoter was functional in the two lumbar ganglia, the dorsorectal ganglion of the tail (Fig. 2A
), and the ventral nerve cord (Fig. 2B
). Developmentally, reporter gene expression was first observed in enclosure stage embryos in a variety of cells (Fig. 2E
), and expression persisted throughout embryogenesis, the four larval stages, and the entire adult life span (data not shown). Neither cell- nor tissue-specific expression pattern changes were observed during development, nor was there any evident modulation of expression level, as judged by fluorescence intensity. Furthermore, GFP expression persisted throughout the dauer larval stage (Fig. 2C
), although GFP expression level decreased significantly, particularly in the pharynx. No significant changes were observed after food starvation or upon resumption of development subsequent to exiting the dauer stage (data not shown). Similar conclusions concerning expression pattern were supported by analysis of protein distribution by immunofluorescence (Fig. 3
and data not shown).
|
sta-1 mutant worms have normal development and morphology
To probe potential biological roles of STA-1, we determined the phenotype of sta-1 mutant animals. We generated a 2.1 kb genomic deletion mutant allele, termed qa5800, by PCR screening 106 animals mutagenized by a transposon-mediated procedure involving insertion and subsequent inaccurate deletion (18)
. An independent deletion mutant allele was requested through the C. elegans Gene Knockout Consortium at Oklahoma Medical Research Foundation, where ok587 was isolated by combined chemical and UV mutagenesis (26)
. Sequence analysis of the two mutants showed that they had each sustained distinct deletions (Fig. 1A
). qa5800 lacks 2039 bp, deleting exons 2, 3, 4 and parts of flanking introns 1 and 4. Splicing exon 1 to exon 5 would create an inframe STA-1 protein fragment lacking half of the coiled-coil domain and the majority of the DNA binding domain, resulting in a molecular mass of 56 kDa (Fig. 3A
). The second allele, ok587, has a deletion of 1290 bp, losing exon 2 and part of exon 3 along with half of intron 1 (Fig. 1A
). Splicing exon 1 to exon 4 would create a premature stop codon early in exon 4, predicting that sta-1(ok587), if transcribed and translated, would yield a small peptide containing only the first half of the coiled-coil domain (Fig. 3A
).
Western blot analysis using affinity-purified rabbit antibody (Ab) directed against the carboxyl terminus of STA-1 confirmed the sequence analysis of wild-type and mutant sta-1 alleles (Fig. 3B
). Wild-type Bristol N2 worms expressed an 85 kDa protein that was missing in extracts from qa5800 and ok587 mutant worms. Consistent with sequence predictions, sta-1(ok587) showed no proteins that reacted specifically with the antiserum, while sta-1(qa5800) expressed a 56 kDa cross-reactive protein consistent with the predicted truncated STA-1 protein. Truncated sta-1(qa5800) was expressed at lower abundance than wild-type protein (Fig. 3B-D
) and essentially was detected only in the pharynx in sta-1(qa5800) worms, the site of highest expression in wild-type worms.
Lack of protein expression and the nature of the deletion sustained by the ok587 allele indicated that it can be considered a null allele. The qa5800 allele is also likely to be a functional null, given that the protein fragment it encodes would be defective in carrying out the roles of its wild-type counterpart, due to the absence of functional coiled-coil and DNA binding domains and its reduced expression level. Phenotypic analysis of the two mutants has failed to uncover any significant biological differences between them (data not shown).
STA-1 subcellular localization in head neurons
A general characteristic of STAT proteins is that they exist latent in the cytoplasm and only accumulate in nuclei after phosphotyrosine-dependent activation, allowing them to drive transcription in a signal-dependent manner (1)
. Thus, subcellular localization of STAT proteins can be an indirect measure of activation. Immunofluorescence using affinity-purified rabbit serum that recognized endogenous STA-1 proteins (Fig. 3B-D
) showed STA-1 protein expression throughout the body of Bristol N2 animals and revealed a distinct immunostaining pattern in head ganglia where STA-1 accumulated in the nuclei of a small subset of neurons (14)
. To more precisely characterize STA-1 subcellular localization, we used detailed confocal immunohistochemistry. While many ganglion cells showed a uniform diffuse staining pattern, similar to patterns observed in other body regions, patterns of clear nuclear exclusion and of precise nuclear accumulation of STA-1 were observed in subsets of neurons. Examples of all three expression patterns are shown in Fig. 3E
, where uniform accumulation, nuclear exclusion, and nuclear accumulation were observed in three adjacent cells (Fig. 3E
, right panel). The immediate juxtaposition of these different patterns of expression, which were reproducibly detected in multiple individuals, suggest precise and selective regulation of STA-1 activity.
STA-1 function is regulated by tyrosine phosphorylation but DNA binding is not required for nuclear accumulation
All the STATs characterized to date are capable of being tyrosine phosphorylated (4)
. Therefore, we asked whether STA-1 could also be tyrosine phosphorylated. Since a nematode cell culture system is not amenable for transient transfection analysis (27)
and a nematode kinase capable of phosphorylating STA-1 in vivo has not been identified, a mammalian cell culture system was used. FLAG-tagged sta-1 protein was transiently transfected into HEK293T cells, together with Tpr-Met, a human tyrosine kinase capable of phosphorylating STATs (28)
. Immunoprecipitation of transfected STA-1 detected a specific tyrosine-phosphorylated protein when STA-1 and the kinase were coexpressed (Fig. 4
A, top panel, lane 5). Furthermore, when the tyrosine residue in the conserved predicted tyrosine phosphorylation motif GYIQ was mutated to phenylalanine, the phosphorylation event was abolished (Fig. 4A
, top panel, lane 6), even though the mutant protein was expressed at an equivalent level to the wild-type protein (Fig. 4A
, lower panel, compare lanes 5 and 6). Therefore, STA-1 can be tyrosine phosphorylated at a single specific tyrosine residue that is conserved with other STAT proteins. Similar site-specific tyrosine phosphorylation was observed when STA-1 was coexpressed with murine JAK1 (data not shown).
|
Another attribute of STAT activation is nuclear translocation after dimerization via reciprocal SH2-phosphotyrosine binding. To determine whether STA-1 undergoes phosphorylation-dependent nuclear translocation, cytoplasmic and nuclear extracts were fractionated from transfected HEK293T cells and analyzed for STA-1 (Fig. 4B
). When expressed alone, STA-1 was found predominantly in the cytoplasmic compartment (Fig. 4B
, lane 4, compare top panel and lower panel). However, coexpression of Tpr-Met yielding STA-1 specifically phosphorylated on tyrosine (Fig. 4A
) stimulated a 3- to 4-fold increased accumulation of STA-1 in the nucleus (Fig. 4B
, lane 9, compare top panel and lower panel), indicating phosphorylation-dependent nuclear translocation. In contrast, the Y-F mutant, which could not be tyrosine phosphorylated (Fig. 4A
), remained predominantly in the cytoplasm (Fig. 4B
, lane 5, compare top panel and lower panel) regardless of the absence (lane 5) or presence of an active kinase (lane 10), as indicated by the ratio of cytoplasmic and nuclear STA-1.
A STA-1 deletion mutant and an alternatively spliced deletion mutant with an extra three amino acids in the C-terminal domain, both encoded by the sta-1(qa5800) allele, also accumulated in the nucleus when coexpressed with kinase (Fig. 4B
, compare lanes 7 and 8 to lanes 2 and 3). Since these mutant proteins lack intact coiled-coil and DNA binding domains (Fig. 3A
), nuclear retention did not require specific DNA binding. Separation of nuclear translocation from DNA binding contrasts with some analyses of mammalian STAT proteins. For instance, growth hormone-induced STAT5B nuclear translocation required a functional DNA binding domain, since mutations that abolished DNA binding abrogated activation-dependent nuclear transport (29)
. Similarly, STAT1 nuclear accumulation has been tied to DNA binding (30
, 31)
, but DNA binding independent mechanisms have recently been reported as well (32)
.
STAT proteins bind enhancer elements in the promoters of target genes after activation and nuclear translocation. Vertebrate and fly STAT homodimers or heterodimers, with the exception of STAT2, all bind to a palindromic signal transducer and activator of transcription element, TTCNNNGAA (6)
, with each monomer recognizing one of the inverted repeats (33
, 34)
. In contrast, slime mold STATs bind distinct, nonpalindromic sequences related to IRF binding sites (35)
. Since STA-1 lacks the amino-terminal domain known to stabilize DNA binding of mammalian STAT proteins, we investigated the DNA binding capability of STA-1. EMSA were carried out using nuclear extracts from transfected cells expressing phosphorylated STA-1, and a double-stranded oligonucleotide containing a high-affinity mammalian STAT binding sequence (GAS) derived from the ICAM1 gene promoter was used as probe. As shown in Fig. 5
, a strong DNA binding activity was detected in cell extracts only when STA-1 and Tpr-Met were coexpressed (Fig. 5A
, lane 9, solid arrowhead). This protein-DNA complex could be supershifted by anti-FLAG Ab, indicating that it contained FLAG-tagged STA-1 (Fig. 5B
, compare lanes 4 and 9, solid arrowhead). Furthermore, DNA binding was abolished by mutation of the STA-1 tyrosine phosphorylation site (Fig. 5A
, lane 10). Consistent with the predicted lack of an intact DNA binding domain, STA-1 deletion mutants derived from the sta-1(qa5800) allele failed to bind the GAS probe (Fig. 5A
, lanes 7, 8), although the mutant proteins accumulated in the nucleus (Fig. 4B
, lanes 7, 8). Therefore, tyrosine phosphorylated STA-1 binds DNA with similar sequence recognition specificity as its mammalian counterparts but does not require DNA binding for nuclear accumulation. Affinity measurements suggested that STA-1 binds DNA with decreased avidity relative to mammalian STAT proteins (data not shown), consistent with the absence of the stabilizing amino-terminal domain. However, absence of this domain did not prevent nuclear accumulation.
|
STA-1 contains a C-terminal transcriptional activation domain
The carboxyl-terminal region containing transactivation potential is poorly conserved among STAT proteins. Not only do they differ in length, ranging from 50 amino acids in STAT1 to 250 in STAT6, they are sometimes missing entirely, such as splice variants of the mammalian STATs and the Dictyostelium sp. STATs. With the exception of Dd-STATs, which were demonstrated to be transcriptional repressors, all full-length STATs are transcriptional activators, requiring this carboxyl-terminal region (6)
, while truncated forms lacking this region may function as dominant-negative regulators that modulate transcriptional activities of the full-length proteins (4)
.
STA-1 contains a short carboxyl-terminal region of 130 amino acids, suggesting that it might encode a transcriptional activation domain. To investigate this possibility, the carboxyl terminus immediately following the SH2 domain was fused with the yeast Gal4 DNA binding domain (36)
. Coexpression of Gal4-STA-1 with a Gal4-responsive luciferase reporter plasmid resulted in transcriptional activation (Fig. 5C
). Although the level of activation was less than that conferred by the mammalian STAT3 transactivation domain, this experiment confirmed that C. elegans STA-1 is a potential transactivator, distinct from the Dictyostelium sp. proteins that act predominantly as repressors.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. It has been suggested that SH2 domains and their tyrosine-phosphorylated ligands arose as an important signaling system with the conversion to multicellularity, and STAT proteins may represent one of the earliest manifestations of this signal responsive cassette (37)
. In addition to the seven mammalian STATs, several vertebrate and invertebrate STATs have now been characterized, including one in Drosophila (38
,39)
, one in Anopheles sp. (40)
, at least four STAT proteins in Dictyostelium (35
, 41
, 42)
, and several in zebrafish (43)
. Here, we have described a novel C. elegans STAT protein, STA-1, which shares significant structural and functional homology with other STAT family members, adding to our understanding of STAT evolution.
Sequence analysis showed that C. elegans STA-1 contains the core features characteristic of the STAT family, with the exception of the amino-terminal multimerization domain (Fig. 1B
). We found that STA-1 was tyrosine phosphorylated at a single conserved phosphotyrosine motif (Fig. 4A
), translocated to the nucleus in response to phosphorylation (Fig. 4B
), and bound DNA in a phosphotyrosine-dependent manner with a sequence specificity similar to vertebrate STAT proteins (Fig. 5)
. In addition, STA-1 contained a transactivation domain (Fig. 5C
). Therefore, STA-1 fits all the criteria to be considered a bona fide STAT.
A major question remains concerning how STA-1 becomes activated and what is its function. In mammals, nonreceptor Janus tyrosine kinases (JAKs) are largely responsible for phosphorylation and activation of STATs in response to extracellular stimuli. However, similar to Dictyostelium sp., no JAK has been identified in nematodes (44)
, nor have cytokine or cytokine receptor-like genes been characterized (45)
. However, many protein tyrosine kinases exist in the C. elegans genome (44)
, and kinases other than JAK can activate STAT proteins in other species. In Dictyostelium sp., the activation of STATs by an unknown tyrosine kinase is mediated by a G-protein-coupled receptor (46)
. Likewise, STAT activation in mammals can occur through non-JAK enzymes (reviewed in ref 47
). Since there exist >90 protein tyrosine kinases in worms (44)
, numerous candidates exist for a tyrosine kinase that can activate STA-1.
STA-1 is expressed in many tissues, including pharynx, intestine, and the nervous system. However, nuclear accumulation indicative of constitutive activation of STA-1 was only detected in a limited set of sensory neurons, particularly amphid neurons, where we recently showed that it participated in regulating dauer development (14)
. The wide STA-1 expression pattern beyond amphid neurons suggests that it may have additional roles. Some of these roles may involve dauer formation, since many components required for the dauer pathway are widely expressed throughout nematode tissues, mediating the diverse developmental changes that are triggered during this response, although the decision itself is processed in head ganglion sensory neurons (48)
.
The majority of STA-1 expressing cells, such as intestinal, body muscle, and many neuronal cells, exhibited a uniform pattern of STA-1 distribution in both nuclei and cytoplasm (Fig. 3C, E
). This pattern is distinct from vertebrate cells, where STAT molecules reside primarily in the cytoplasmic compartment before activation, creating a pattern of nuclear exclusion (1)
. However, this steady-state pattern is the end result of a dynamic cytoplasmic-nuclear shuttling process and reflects the balance between rates of nuclear import and export (49)
. Possibly in many nematode tissues the equilibrium between these processes is shifted more toward nuclear accumulation than in vertebrate cells. However, the existence of cells with precisely cytoplasmic or nuclear STA-1 protein adjacent to one another suggests that STA-1 activity is stringently regulated.
Absence of the otherwise conserved amino-terminal domain suggests that STA-1 is distinct from most other family members and may be most related to the Dictyostelium sp. homologues, which also lack this domain. The amino-terminal domain has been ascribed numerous functions, including roles in multimerization, receptor-dependent activation, stabilization of DNA binding, interaction with regulatory proteins, and nuclear translocation (4
, 50
51
52
53
54)
. Drosophila sp. also express an amino-terminally truncated STAT isoform that appears to function as a repressor (55)
, similar in function to at least some of the Dictyostelium STATs. Since Dictyostelium sp. proteins may represent the most ancient members of the STAT family, it is tempting to speculate that the primordial protein lacked both an amino-terminal multimerization domain and a carboxyl-terminal transactivation domain, and functioned as a negative regulator of gene expression. By this scenario, domain accretion during evolution added new functionality, first by acquisition of transcriptional activation domains (present in C. elegans and all higher forms), conferring gene induction potential on a primordial negative regulatory protein, and later by acquisition of amino-terminal domains, allowing more flexibility in DNA binding and receptor-dependent activation.
The presence of STAT genes without cytokines, cytokine receptors, or JAK partners in slime mold and worms suggests that STAT family transcription factors arose early in metazoan evolution as a primitive switch regulated by tyrosine phosphorylation independent of the cytokine system. The existence of two distinct arms of the STAT family suggests that these genes arose by gene duplication, with the earliest duplication occurring before the divergence of insects and vertebrates (40
, 56
, 57)
. The C. elegans genome contains a second gene (F56E6.1) with a DNA binding and SH2 domain. However, this gene does not appear to qualify as a STAT family member since it lacks several significant features. For instance, the DNA binding homology domain lacks several key residues involved directly in DNA binding; it lacks the STAT SH2 domain signature FLLRFS, and no phospho-tyrosine motif is present in the carboxyl terminus. Our biochemical characterization of this protein did not support a role as a STAT protein (data not shown). Nonetheless, F56E6.1 has a clear homologue in C. briggsae, a nematode species separated from C. elegans by >100 million years (25)
, suggesting that the earliest STAT gene duplication event occurred even before the divergence of nematodes.
|
ACKNOWLEDGMENTS |
|---|
Received for publication March 7, 2006. Accepted for publication March 31, 2006.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, Stat1ß and truncated Stat1: interaction between NH2-terminal domains stabilizes binding of two dimers to tandem DNA sites. EMBO J. 15,5616-5626[Medline]
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||
