Tissue Distribution of AU-rich mRNA-binding Proteins Involved in Regulation of mRNA Decay*

Short lived cytokine and proto-oncogene mRNAs are destabilized by an A+U-rich element (ARE) in the 3′-untranslated region. Several regulatory proteins bind to AREs in cytokine and proto-oncogene mRNAs, participate in inhibiting or promoting their rapid degradation of ARE mRNAs, and influence cytokine expression and cellular transformation in experimental models. The tissue distribution and cellular localization of the different AU-rich binding proteins (AUBPs), however, have not been uniformly characterized in the mouse, a model for ARE mRNA decay. We therefore carried out immunoblot and immunohistochemical analyses of the different AUBPs using the same mouse tissues. We show that HuR protein, a major AUBP that stabilizes the ARE mRNAs, is most strongly expressed in the thymus, spleen (predominantly in lymphocytic cells), intestine, and testes. AUF1 protein, a negative regulator of ARE mRNA stability, displayed strong expression in thymus and spleen cells within lymphocytic cells, moderate expression in the epithelial linings of lungs, gonadal tissues, and nuclei of most neurons in the brain, and little expression in the other tissues. Tristetraprolin, a negative regulator of ARE mRNA stability, displayed a largely non-overlapping tissue distribution with AUF1 and was predominantly expressed in the liver and testis. KH-type splicing regulatory protein, a presumptive negative regulator of ARE mRNA stability, was distributed widely in murine organs. These results indicate that HuR and AUF1, which functionally oppose each other, have generally similar distributions, suggesting that the balance between HuR and AUF1 is likely important in control of short lived mRNA degradation, lymphocyte development, and/or cytokine production, and possibly in certain aspects of neurological function.

Cytokine mRNAs in mammalian cells are typically targeted for rapid degradation by an ARE 1 present in the 3Ј-untranslated region (reviewed in Ref. 1). The mechanism and process by which the ARE promotes rapid degradation of mRNAs are not well understood. AREs generally consist of repeating pen-tamers of the sequence AUUUA, which promotes the rapid cytoplasmic degradation of cytokine and proto-oncogene mRNAs (1)(2)(3). The accelerated decay of ARE mRNAs likely proceeds by rapid deadenylation followed by rapid degradation of the mRNA body (4 -9). Although a number of proteins can be cross-linked in vitro to the ARE, only several have been shown to have regulatory functions for ARE mRNAs. Two AUBPs (AUF1 and TTP) promote rapid decay of ARE mRNAs, one (HuR) inhibits mRNA turnover, and one (KH-type splicingregulatory protein (KSRP)) is suspected in promoting ARE mRNA degradation. The protein TTP promotes rapid decay of tumor necrosis factor (TNF) and granulocyte macrophage colony-stimulating factor mRNAs (10,11). TTP overexpression further facilitates the rapid decay of TNF and granulocyte macrophage colony-stimulating factor mRNAs (12), whereas in TTP knock-out mice these mRNAs are stabilized (11,13). The protein family known as AUF1 or hnRNP D also binds the ARE and is associated with rapid decay of ARE mRNAs, as originally shown in an in vitro ARE mRNA decay system (14). AUF1 comprises four isoforms produced by differential splicing of a single transcript (15). The four isoforms consist of a 37-kDa core protein (p37), a 40-kDa protein (p40) with an N-terminal 19-amino acid insertion of exon 2, a 42-kDa protein (p42) with a C-terminal 49-amino acid insertion of exon 7, and a 45-kDa protein (p45) containing both exon 2 and 7 insertions (reviewed in Ref. 1). The different protein isoforms possess different ARE-RNA binding characteristics (16), ubiquitination, and stabilities (17). The p37 followed by the p40 isoform is most closely associated with promotion of ARE mRNA degradation in a number of studies (18 -24).
The Hu family of proteins (HuA/R, HuB, HuC, HuD) are the only proteins shown to date to stabilize ARE mRNAs. HuR and HuB inhibit ARE mRNA turnover when ectopically overexpressed (25)(26)(27), and antisense RNA knockdown of endogenous HuR expression decreases the half-lives of certain ARE-containing mRNAs (28 -30). Other studies (30 -32) have identified HuR as an important stabilizer of short lived AU-rich mRNAs in a variety of settings and in different cell types in culture. The protein KSRP has been implicated in ARE mRNA decay, but it has not been shown to directly promote rapid mRNA turnover (33). KSRP has been found to copurify with the mammalian exosome, a large multiprotein complex containing 3Ј-5Ј-exoribonucleases and other proteins involved in mRNA degradation (33).
In summary, two or three AUBPs promote ARE mRNA destabilization (TTP, AUF1, and possibly KSRP), yet only one family (Hu proteins) is known to promote ARE mRNA stabilization. In mammalian cell lines, HuR is widely expressed (34) but at different levels (22,25). It is possible that destabilization of ARE mRNAs involves multiple redundant protein functions, or it is restricted to different tissues by different proteins. Additionally, the fact that HuR, a major expressed member of the Hu protein family, stabilizes ARE mRNAs suggests that it is expressed in most tissues to provide for regulation in opposition to that of the destabilizing AUBPs, that in some tissues destabilizing AUBPs are likely unopposed in function, or that other inhibitors of ARE mRNA turnover remain to be identified. As might be expected for proteins that regulate protooncogene and cytokine mRNA stability, both HuR and AUF1 expression levels have been implicated in carcinogenesis. Overexpression of HuR has been associated with malignant brain and lung tumorigenesis (35,36) and cellular proliferation (37). AUF1 levels have also been associated with tumorigenesis in a mouse model (38). There has been only a limited analysis of the adult mouse tissue distribution of the different regulators of ARE mRNA stability, and no study has examined the distribution of all AUBPs simultaneously in mice, in which much of the research on cytokine and proto-oncogene mRNA stability takes place. Here we report the tissue distribution of the major known regulators of ARE mRNA stability. These results suggest that HuR and AUF1 are found at the highest levels primarily in the same tissues (thymus and spleen, followed by intestine), whereas TTP is distributed in a largely non-overlapping manner with HuR. KSRP, which may play a role in ARE mRNA destabilization, was distributed widely.

MATERIALS AND METHODS
Antibodies-Anti-hnRNP D (5B9) monoclonal antibodies were kindly provided by Dr. G. Dreyfuss (University of Pennsylvania, Philadelphia, PA). Rabbit polyclonal antibodies to recombinant p37 AUF1 were developed in our laboratory to bacterially produced and purified p37 protein, affinity purified and concentrated (PcAb 995). Antibodies to HuR (3A2 monoclonal antibody, Santa Cruz), TTP (H-120 polyclonal antibody, Santa Cruz), and ␤-tubulin monoclonal antibody (Sigma) were purchased commercially. A polyclonal antibody against KSRP (C2742) was a gift from Dr. D. Black (Howard Hughes Medical Institute/University of California, Los Angeles, CA).
Cell Culture and Northern Blot Analysis-Chinese hamster ovary and HeLa cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum or bovine calf serum, respectively. Mouse embryonic stem (ES) cells were cultured on gelatincoated plates and maintained in Dulbecco's modified Eagle's medium with high glucose supplemented with 15% fetal bovine serum, ␤-mercaptoethanol (100 M), non-essential amino acids (100 M), 1 mM sodium pyruvate, 2 mM L-glutamine, and 1000 units/ml leukemia inhibitory factor. Total RNA was isolated from ES cells by using TRIzol reagent (Invitrogen). 20 g of total RNA was resolved in agarose/ formaldehyde gels and transferred to nylon membrane. The blot was hybridized to 32 P-labeled probes prepared against the 220-bp murine AUF1 exon 1 and exposed to film.
Animal Studies-Experiments were carried out in accordance with National Institutes of Health guidelines for animal treatment, housing, and euthanasia. Young adult C57BL/6 male and female mice (6 weeks old) were sacrificed by CO 2 asphyxiation. The organs were removed rapidly by dissection, cut into thin slices, homogenized, and snap-frozen in liquid nitrogen for later use or fixed in neutral-buffered formaldehyde. Homogenization was carried out using 5 volumes of homogenization buffer (50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 1 mM EDTA, 1% Nonidet P-40, 1ϫ complete protease inhibitor mixture (Roche)). After incubation on ice for 15 min, samples were clarified by centrifugation at 12,000 rpm at 4°C for 10 min. Then the soluble supernatant was reserved. Protein concentrations were determined by Bradford assay (Bio-Rad).
Protein Analysis-For Western immunoblot analysis, equal amounts of protein extracts from murine tissues or cells were denatured by the addition of 2ϫ SDS-PAGE buffer (100 mM Tris-HCl, pH 6.8, 10% ␤-mercaptoethanol, 4% SDS, bromphenol blue, 20% glycerol) and heating to 100°C for 5 min. Equal amounts of cell lysates were resolved by SDS-PAGE and transferred to nitrocellulose membrane, and immunoblotting was performed using primary antibody as indicated above. Immunoblots were developed using the ECL system.
Immunohistochemistry Analysis-For immunohistochemistry, surgically excised organs were fixed in 10% buffered formalin, dehydrated through a series of ethanol followed by xylene clearing, and embedded in paraffin. Sections were cut to a 5-m thickness and heated overnight in at 56°C. Sections were deparaffinized in xylene and rehydrated through a series of alcohol. Immunohistochemical staining was performed using a Vectastain ABC kit (Vector Laboratories) The sections were subjected to microwave antigen retrieval using a sodium citrate buffer. Endogenous peroxidase activity was blocked by a 10-min incubation in 3% H 2 O 2 in phosphate-buffered saline followed by rinsing in phosphate-buffered saline. Nonspecific binding of the secondary antibody was blocked by incubating in 15% normal goat serum for 1 h at room temperature. Sections were subsequently incubated overnight at 4°C in anti-AUF1 polyclonal antibody (PcAb 995), which was diluted at 1:200 with phosphate-buffered saline containing 5% normal goat serum. Following incubation with biotinylated secondary antibody and ABC reagent, staining was visualized using a 3,3Ј-diaminobenzidine substrate kit (Vector Laboratories). The sections were then counterstained with hematoxylin, cleared, and mounted with Permount medium (Sigma).

RESULTS AND DISCUSSION
Validation of AUF1 Antibodies-Previous studies have not collectively compared the major AUBP regulators of ARE mRNA stability in mice in a variety of tissues and organs. In addition, our immunoblot studies indicate that the antibody (5B9) often used to detect hnRNP D/AUF1 strongly cross-reacts with another protein of similar molecular mass to the p45 AUF1 isoform (Fig. 1B), confounding previous immunohistochemical and immunoblot studies. We therefore first developed and affinity purified a highly specific polyclonal rabbit antibody (PcAb 995) to AUF1 that does not cross-react with other polypeptides within the range of 35-50 kDa. To demonstrate the specificity of this antibody, protein extracts from wild type and AUF1 knock-out ES cells were used. The details of the development of the AUF1 gene knock-out cells will be reported elsewhere. Northern mRNA analysis confirmed that the homozygous AUF1 knock-out ES cells (Ϫ/Ϫ) do not express AUF1 mRNA (Fig. 1A) compared with wild type (ϩ/ϩ) and heterozy-FIG. 1. Validation of AUF1 antibody. A comparison was made of wild type, heterozygous, and homozygous AUF1 mutant murine ES cells for reactivity to AUF1 antibodies. A, Northern blot mRNA hybridization analysis of AUF1 mRNA from wild type, heterozygous, and homozygous AUF1Ϫ/Ϫ ES cells. mRNAs were hybridized to 32 P-labeled probes prepared against the first exon of AUF1 cDNA. B, equal amounts of whole cell protein extracts from ES, Chinese hamster ovary (CHO), and HeLa cells were resolved by SDS-12% PAGE transferred to nitrocellulose membranes, probed by immunoblot analysis using either 5B9 monoclonal antibody (middle panel) or PcAb995 polyclonal antibody (left panel) to AUF1, and visualized using the enhanced chemiluminescence system. An additional protein band of ϳ46 -47 kDa was detected by the 5B9 antibodies even in homozygous AUF1Ϫ/Ϫ cells. Right panel, equal amounts of whole cell extracts from Chinese hamster ovary and HeLa cells were resolved by higher resolution SDS-10% PAGE and subjected to 5B9 antibody immunoblot analysis, which revealed a partial doublet at 45-46 kDa. gous (ϩ/Ϫ) ES cells. Using the same ES cells, anti-hnRNP D/AUF1 monoclonal antibody (5B9), which has been widely used in the literature, was found to cross-react with a slightly larger polypeptide. In 12% SDS-PAGE, an unidentified polypeptide was recognized by the 5B9 but not the PcAb 995 antibodies; this polypeptide migrates very closely with, but is clearly distinct from, the largest isoform of AUF1, as it remained present even in homozygous AUF1 knock-out ES cells (Fig. 1B, compare left with middle panel). This cross-reactivity could lead to misinterpretation of the unknown protein as the p45 AUF1 isoform and, accordingly, the actual p45 AUF1 protein as the p42 AUF1 isoform. The high affinity polyclonal anti-AUF1 antibody recognized four bands in wild type ES cells and in Chinese hamster ovary cells (the p40 and p42 AUF1 bands co-migrate) but not the cross-reactive polypeptide detected by the 5B9 monoclonal AUF1 antibody. Higher resolution of AUF1 proteins in SDS-10% PAGE suggests that the unknown protein that cross-reacts with the 5B9 antibody has a molecular mass of 46 -47 kDa (Fig. 1B, right panel). These data therefore validate the use of the PcAb 995 antibody for this study. Antibodies were validated previously to KSRP (39) and HuR, which has shown some cross-reactivity to HuD and HelN1 (HuB) but not HuC (22).
Tissue Distribution of AUBP Proteins-Six-week-old male and female C57BL/6 mice (young adults) were sacrificed, and organs and tissues were rapidly dissected and either snapfrozen in liquid nitrogen, fixed in formalin, and embedded in paraffin or immediately processed for SDS-PAGE. For electrophoresis, whole cell extracts were prepared by mechanical disruption in homogenization buffer and clarified by centrifugation, and equal amounts of protein were resolved by SDS-PAGE. Proteins were transferred to nitrocellulose membrane, and immunoblot analysis was performed using primary antibodies as indicated and ECL for detection.
Tissue Expression of HuR Protein-There are four members of the Hu family of RNA-binding proteins that participate in RNA processing events and ARE mRNA stability (reviewed in Ref. 40). HuB (HelN1), HuC, and HuD are developmentally regulated and tissue-restricted in expression, predominantly in the brain (40). HuR (also known as HuA) is expressed widely but at different levels in mammalian cell lines (22,25,40). However, a limited number of human and mouse tissues have been studied for HuR protein expression (22,25,(41)(42)(43), largely by utilizing mRNA expression (41, 42, 44 -46). Studies have not investigated fully HuR protein expression levels in a variety of tissues from the adult mouse. Soluble whole cell protein extracts were prepared from male and female mouse tissues; organs were denatured and resolved by SDS-PAGE, and HuR was identified by immunoblot analysis using a monoclonal antibody to HuR (3A2), which most strongly recognizes the HuR member of the Hu family but also cross-reacts with HuB and HuD (22). Normalization to equal protein levels and the lack of protein degradation were reflected in similar total polypeptide levels (within 2-fold) by Ponceau S staining, permitting direct comparisons between tissues (Fig. 2C). Immunoblot staining of poly(A)-binding protein in tissues (Fig. 2D) demonstrated more specifically the lack of degradation of proteins and that similar (within 2-fold) levels were present in equal amounts of tissue protein extracts. Similar results were found for immunoblot analysis of tubulin proteins as well (data not shown). Thus, it is highly unlikely that differential protein degradation could occur in these rapidly frozen tissues. In both male and female young adult mice HuR was predominantly expressed in the intestine, thymus, and spleen, followed by the liver (Fig. 2, A and B). The significant expression of HuR in proliferating lymphocytic cells (43) could play a role in the high levels observed here. HuR was almost undetectable in the brain, skeletal muscle, kidney, lung, and heart (Fig. 2, A and  B), arguing against strong cross-reactivity of the 3A2 antibody with HuB, -C, or -D. In female mice, a low level of HuR expression was detected in the uterus and little in the ovary, whereas male mice expressed HuR strongly in the testes. These results are consistent with several reports demonstrating strong HuR expression in malignant murine lung and brain carcinomas but not in normal tissue (35,36). A previous study (47) of HuR protein tissue expression in the adult mouse reported strong expression in the spleen, thymus, and testes as found here but additionally showed strong expression in the brain, lung, and intestine. Thus, our studies concur with those of Gouble and Morello (47), apart from the abundance of HuR in the brain. In this regard, a previous study (35) reported strong HuR expression in brain tumors but not normal brain using a validated antibody to HuR, similar to our results. Because studies demonstrated that the adult mouse brain predominantly expresses HuC and -D forms of the Hu protein (48), we suspect that reports of high levels of HuR protein in the brain more likely reflect cross-reactivity to HuC and -D neuronal forms. The analysis of HuR in a limited number of human tissues, using a commercial source of tissue, showed strong expression in the A comparison was made of HuR protein levels in tissues from 6-week-old male and female C57BL/6 mice. Tissues were rapidly dissected from sacrificed mice and snap-frozen or processed immediately for SDS-PAGE and immunoblot analysis. Equal amounts of whole cell protein extract from female (A) and male (B) mouse tissues were resolved by SDS-PAGE and subjected to immunoblot analysis for HuR using antibody 3A2 and ECL. C, the same blot used in B was stained by Ponceau S to demonstrate similar protein loading levels. D, equal amounts of the protein used above were resolved by SDS-PAGE, and immunoblot detection was carried out using antibodies specific for poly(A)-binding protein.
spleen and testes as found here in mouse tissue (the thymus was not investigated) but differed in finding strong expression in the brain, kidney, and lung (22), which might reflect species differences or differences in the amount of protein analyzed. It is also interesting to note that a survey of reported HuR protein and mRNA expression levels in tissues suggests a possible disparity, potentially implicating translational control or differential mRNA versus protein stabilities in HuR regulation.
Tissue and Isoform Expression of AUF1 Proteins-AUF1 consists of four isoforms (p37, p40, p42, and p45) derived from the same mRNA by differential splicing (1). The p37 and p40 isoforms are most closely associated with the promotion of ARE mRNA decay (14,19,21,23,24,49). The studies carried out demonstrating tissue-and isoform-specific distribution of AUF1 previously were limited in scope (45,47) and may be compromised by detectable cross-reactivity to other proteins of similar molecular weight to the larger AUF1 isoforms. The same mouse tissues that were analyzed for HuR were examined for AUF1 protein expression using an affinity-purified polyclonal antibody (PcAb 995), validated as shown in Fig. 1. Additional studies from multiple independent tissue isolates from other mice confirm the results presented below. Both male and female mice demonstrated a very strong expression of AUF1 in the spleen and thymus (Fig. 3). As is typical with AUF1, the p40 and p42 isoforms could not be separated readily, although presumptively there were no discernable differences in isoform expression patterns. AUF1 was also represented at moderate levels in the brain and reproductive organs including the testis, ovary, and uterus, with weak expression in both intestine and lung. AUF1 expression in all other tissues was extremely weak or undetectable. However, the expression pattern of the four AUF1 isoforms was not uniform among these organs. In the brain, testis, and uterus, p45 and p42/p40 were predominant with only a low level of p37 evident. This pattern was reversed in the lung and ovary in which p45 was undetectable, and p37 and p42/p40 were the major isoforms detected. When compared in parallel, there is a striking isoform-specific expression of AUF1 in different organs, suggesting that expression or accumulation of individual isoforms is differentially regulated and possibly adapted to the function of a specific organ or tissue. A previous analysis of AUF1 mRNA expression in adult mouse tissues showed strong expression in all organs except the brain (45, 47), suggesting that differential expression of the AUF1 isoforms is related to different patterns of alternate exon 2 and 7 splicing or differential protein stabilities. From our work, it is apparent that HuR and AUF1 proteins significantly overlap in tissue-related gene expression in the spleen, thymus, and testis, the only differences being a high level expression of HuR in intestine and a low level of AUF1.
Tissue Expression of TTP Protein-TTP is an immediate early protein transiently expressed in response to TNF and other stimuli through a process that involves feedback inhibition (50). TTPϪ/Ϫ mice develop systemic inflammatory syndrome mediated by overexpression of TNF through failure to rapidly degrade its mRNA, predominantly in macrophages. The tissue distribution of TTP has been examined in mice primarily at the mRNA expression level, despite the fact that there is evidence for auto-regulation of TTP expression. TTP mRNA is reportedly expressed in developing oocytes and regenerating liver, intestine, lung, spleen, thymus, and macrophages (13). Analysis of TTP protein expression in a variety of adult mouse tissues and organs has not been reported. Surprisingly, examination of 6-week-old male and female mice for TTP protein expression (Fig. 4) demonstrated strong expression only in the livers of male and female mice and in the testes in male mice, with weaker expression in ovaries of female mice. Despite the reported expression of TTP mRNA in a variety of mouse tissues, TTP protein expression showed a remarkable degree of tissue-restricted expression. Given the significant expression of TTP protein in macrophages, the strong expression in liver is likely restricted to Kupfer cells and macrophages rather than hepatocytes. These data indicate that TTP is largely expressed at the protein level in a tissue-restricted fashion, which does not significantly overlap with the expression of AUF1 and HuR. Moreover, these data suggest that translation of TTP mRNA may be tissue-restricted.
Tissue Expression Profile of KSRP Protein-KSRP is an hnRNP protein that binds tightly to U-rich residues (39,51). The interest in KSRP with respect to ARE mRNA turnover stems from its co-purification with the mammalian exosome (33), a multisubunit particle composed of RNases that are involved in cytoplasmic 3Ј-5Ј-mRNA degradation. Whether KSRP actually plays a role in ARE mRNA decay in vivo is unknown, but HuR and AUF1 were also co-purified with the exosome (33). The tissue distribution of the KSRP protein in the mouse has not been reported previously. Given its potential role in promotion of ARE mRNA decay, we examined the tissue expression profile of KSRP. KSRP expression was examined using a rabbit polyclonal antibody (C2742). Strong expression was found in the spleen and thymus, as it was for AUF1 (Fig.  5). Moderate expression was found in the ovary, testes, and uterus, with weak expression in all other tissues and organs except the intestine, kidney, and skeletal muscle where the protein was not detected. Thus, the predominant expression profile of KSRP protein clearly overlaps that of AUF1, but it is the most widely expressed AUBP analyzed.
Immunohistochemical Localization of AUF1 Protein-To study AUF1 protein expression at the cellular level and its distribution among different organs, we performed immunohistochemical staining to determine whether the strong expression of AUF1 in the spleen and thymus was limited to the lymphocytes in the cortex or medullary regions or to other structural regions such as the capsule or septum. Tissues from 6-week-old mice were embedded in paraffin and sectioned for immunohistochemical staining using anti-AUF1 polyclonal antibody and a biotinylated secondary antibody (Fig. 6, A and B). Sections were counterstained with hematoxylin, visualized, and photographed by light microscopy. Strong immunoreactive staining of AUF1 was localized to the nuclei of most of the splenocytes and thymocytes within the thymus, cortex, and medulla (Fig. 6, A and B, insets), indicating a high level of AUF1 expression in lymphocytes. Interestingly, not all lymphocytes express AUF1 at an equally strong level, and some are negative for staining (Fig. 6, A and B), raising the possibility that AUF1 may play a role in the development of lymphocytes at certain (but not all) stages. There was little if any staining of the capsule or other structures evident. Expression of AUF1 in thymic and splenic cells is therefore restricted to lymphoid cells, suggesting that AUF1 is primarily expressed in B-and T-lymphocytes. In the absence of an addition of AUF1 antibody, there was no staining evident in any tissues, as shown representatively in the spleen and testicle (Fig. 6, G and H, respectively).
Studies were also conducted to localize AUF1 protein expression by staining other mouse organs including the brain, testicle, ovary, and lung. Intense AUF1 staining was localized in the nuclei of neuronal cells in different areas of the murine cerebrum such as the cortex, hippocampus, and thalamus (Fig. 6C). Given the fact that other hnRNP proteins such as hnRNP A1, C, K/J, and U are also expressed strongly in brain neurons (52), AUF1 may function together with other hnRNP proteins to regulate mRNA turnover in the central nervous system. AUF1 expression in the reproductive system was also examined. Spermatogenesis involves maturation of germ cells toward the center of seminiferous tubules, with proliferation of spermatogonia, meiotic division of spermatocytes, differentiation of spermatids, and release of spermatozoa into the lumen of seminiferous tubules. Staining of AUF1 showed intense localization to the nuclei of spermatocytes in the seminiferous epithelium but not to the mature spermatids and interstitial cells (Fig. 6D). Thus, AUF1 expression is primarily restricted in adult testis to meiotic germ cells. Similarly, AUF1 was found to be strongly expressed in the ovarian follicles at various stages (Fig. 6F) as well as the mucosa epithelium of the oviduct (data not shown). This expression profile suggests the possibility that AUF1 may function in reproductive cell development, possibly in regulation of short-lived mRNA stability in this context. In the lung (Fig. 6E), the ciliated cuboidal epithelium lining of the bronchiole showed strong immunostaining both in the nucleus and cytoplasm, whereas most of the alveolar cells displayed only nuclear immunoreactivity to the AUF1 antibody. In most tissues, no cytoplasmic AUF1 staining was detected, except in the mucus-secreting bronchiole epithelium. This is consistent with a report indicating that other hnRNP proteins are also localized primarily to nuclei (52). Given the strong nuclear staining of AUF1 in various tissues, it would not be surprising to find nuclear functions for certain AUF1 isoform(s) in addition to cytoplasmic ARE mRNA decay.
Summary-HuR, AUF1, TTP, and possibly KSRP proteins bind to AU-rich sequences in the 3Ј-untranslated region of FIG. 6. Immunohistochemical localization of AUF1 protein in various mouse organs. A-F, Tissues from sacrificed 6-week-old male C57BL/6 mice were fixed in buffered formalin, embedded in paraffin, and sectioned into 5-m thin sections for immunohistochemical staining with polyclonal AUF1 antibody PcAb 995 (1:200 dilution) and visualized by 3,3Ј-diaminobenzidine. A, spleen; B, thymus; C, brain; D, testicle; E, ovary; F, lung; G and H, spleen and testicle, respectively, stained in the absence of AUF1 antibody. Phosphate-buffered saline was used to replace AUF1 antibody 995 followed by secondary antibody incubation, 3,3Ј-diaminobenzidine development, and hematoxylin counterstaining. cytokine and proto-oncogene mRNAs and regulate their rapid degradation. The pattern of distribution of these proteins was important to characterize, because their expression levels may be associated with cytokine expression and regulation and with the development of various carcinomas. In this regard, overexpression of HuR is associated with brain and lung tumors in humans (35,36), and AUF1 is associated with sarcomas in transgenic mice (38). In the present study, we have characterized the tissue and organ distribution of all four proteins in the young adult mouse. We show a strong tissue expression profile of the four AUF1 isoforms in lymphocytes in the spleen and thymus. These data implicate AUF1 in the regulation of lymphocyte cytokine mRNA stability, which is likely also regulated by the high levels of HuR found in these cells. It is also very likely that different AUF1 protein isoforms may function in different tissues. For example, the major AUF1 isoforms detected in the brain were p45 and p42/p40, whereas no p45 isoform was detected in lung. Further study using an isoformspecific AUF1 antibody could resolve this issue. Future studies are needed to determine whether the balance of HuR to AUF1 proteins is important in the maintenance of normal lymphocyte cytokine production and whether an imbalance is associated with the development of inflammatory disease or B-and T-cell lymphomas. The distribution of TTP, as reported previously (10,11) is largely restricted to the liver, probably in macrophages, where it is a primary regulator of TNF mRNA stability.