The TATA-binding Protein and Its Associated Factors Are Differentially Expressed in Adult Mouse Tissues*

We have investigated the expression levels of the TATA-binding protein (TBP) and several TBP-associated factors (TAFIIs) in differentiated adult mouse tissues. Immunoblots performed using monoclonal antibodies show that there are considerable variations in the levels of TBP and many TAFIIproteins present in various tissues. Consequently, the relative levels of TAFIIs and TBP vary significantly from one tissue to another. TBP and several TAFIIs are overexpressed in both testis and small intestine, while in marked contrast, many of these proteins, including TBP itself, were substantially down-regulated in nervous tissues and in the heart. These tissues do, however, show a high expression level of the TBP-like factor, which thus may represent an alternative factor for the specialized transcription program in some differentiated tissues. While there are significant variations in the levels of TAFII28 protein, reverse transcription-coupled polymerase chain reaction shows similar expression of the TAFII28 mRNA in different tissues. The variations in TAFII28 protein levels therefore result from post-transcriptional regulatory events.

TFIID 1 is a multiprotein complex, which together with TFIIA, TFIIB, TFIIE, TFIIF, and TFIIH assists RNA polymerase II to correctly initiate transcription (1). TFIID is composed of the TATA-binding protein (TBP), which specifically binds the TATA element, and a series of evolutionary conserved TBPassociated factors (TAF II s). TAF II s have been shown to be involved in promoter recognition (2,3) and to act as specific transcriptional coactivators in vitro and in transfected mammalian cells (Refs. 4 -8 and references therein; for review, see Ref. 9). Genetic experiments in yeast have shown a variable requirement for TAF II s, some of which are required for the expression of only a subset of promoters involved for example in cell cycle control, while others are more generally required (10 -14). Recently, a subset of TAF II s have been found in other complexes devoid of TBP, such as the PCAF⅐SAGA complex in humans and in yeast (15)(16)(17), and the TBP-free TAF II -containing complex (TFTC) (18). Despite the fact that TFTC does not contain TBP it can replace TFIID in both basal and activated transcription in vitro, suggesting that TBP may not always be an essential transcription factor in vivo.
While much has been learned about the function of TFIID in biochemical assays and in yeast, little is known concerning the expression of its constituent subunits in animal tissues. Previous studies on TBP (19) have demonstrated an overexpression of TBP mRNA and to a lesser extent of the TBP protein in testis. The mRNAs of several TAF II s have been shown to be equally expressed in several rat tissues (20). However, the TAF II 105 mRNA is widely expressed yet the corresponding protein shows cell specificity, being much more abundant in mature lymphoid B cells (8).
The above observations prompted us to investigate the expression of TBP and TAF II proteins rather than their mRNAs in a variety of adult murine tissues. Immunoblots performed with a series of monoclonal antibodies show that the relative expression levels of these TAF II s, and TBP can vary extensively from tissue to tissue, suggesting that the transcription program in different tissues may have differential requirements for TFIID components. Furthermore, the levels of TBP and many TAF II s is significantly reduced in extracts from the nervous system (brain, cerebellum, eye, spinal cord), kidney, and in the heart. Interestingly, several of these tissues show high expression levels of the previously described TBP-like factor (TLF) (18), raising the possibility that TLF may functionally substitute for TBP in certain tissues. In the case of TAF II 28, whose mRNA is equivalently expressed in many tissues, the variations in TAF II 28 protein must result from post-transcriptional events.

MATERIALS AND METHODS
Preparation of Murine Tissue Extracts-Four individual 6-week-old Black 6 mice were sacrificed and the tissues extracted and immediately frozen in liquid nitrogen. Protein extracts were made as described (21) by shearing the tissues in 2 ϫ boiling Laemmli buffer containing 10 mM ␤-mercaptoethanol. The extracts were analyzed by SDS-PAGE and staining with Coomassie Blue stain to normalize each preparation.
Preparation of Cell Line Extracts-Cell extracts were prepared as described previously (5) by three cycles of freeze-thaw in 100 l of buffer A (50 mM Tris-HCl (pH 7.9), 20% glycerol, 1 mM dithiothreitol, and 0.01% Nonidet P-40) containing 0.5 M KCl and 2.5 g/ml leupeptin, pepstatin, aprotinin, antipain, and chymostatin. The proteins were quantified by Bradford test and the equivalent amounts were used for immunoblots.
RNA Preparation and RT-PCR-RNA from tissue samples was prepared as described previously (26). RT-PCR was performed on 1 g of total RNA using the following primers 5Ј-GGACAAGAAGGAGAA-GAA-3Ј and 5Ј-CTTCTTGTGCTTTGAGTTGGGGAT-3Ј specific to different exons of mTAF II 28 generating a 360-base pair fragment. Samples were denatured for 3 min at 94°C and annealed for 10 min at 50°C. A mix of avian myeloblastosis virus reverse transcriptase and Taq polymerase was added and incubated for another 20 min at the same temperature. 30 cycles of PCR were then performed. After 15, 23, and 30 cycles an aliquot of each sample was removed and electrophoresed, transferred to a hybond membrane, and hybridized with a 32 P-* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
labeled TAF II 28-specific oligonucleotide probe. As a control a 200-base pair fragment of the hypoxanthine guanine phosphoribosyltransferase (HPRT) gene was amplified in the same reactions and detected by hybridization using an HPRT-specific oligonucleotide probe. Amplification with no avian myeloblastosis virus reverse transcriptase was also performed as a negative control.

RESULTS AND DISCUSSION
Variations in TAF II Protein Content of Adult Murine Tissues-To investigate the levels of TFIID components in adult murine tissues immunoblots were performed using mAbs against a selection of TAF II s, which are either TFIID-specific (TAF II 28, TAF II 18) or are present in other TAF II -containing complexes (TAF II 135, TAF II 100, TAF II 55, TAF II 30, TAF II 20). Six-week-old mice were sacrificed, dissected, and their organs were immediately frozen in liquid nitrogen. Equivalent amounts of the proteins extracted from each tissue (see "Materials and Methods" and Fig. 1) were used to make several replica immunoblots along with extracts from human HeLa and murine F9 cells as controls. All the antibodies used detected both the human TAF II s and their murine counterparts. Analogous results to those shown below were observed in blots from independently prepared extracts (data not shown).
TAF II 135 and TAF II 100 could be detected in all tissues, with the exception of the spinal cord where TAF II 100 was seen only very weakly, while TAF II 135 was undetectable (Fig. 2, see lane 17). Both TAF II 135 and TAF II 100 were strongly overexpressed in the testis where, and for the sake of clarity, a 5-fold shorter exposure is shown (lane 15, Fig. 2). An exposure time comparable with that shown in the other lanes resulted in a saturated black signal (data not shown). Varying levels of TAF II 55 could also be detected in all tissues with overexpression in the testis being less dramatic than for TAF II 135 and TAF II 100 (note that for TAF II 55 the same exposure time is shown in all tissues).
Although these TAF II s are widely expressed, their relative expression levels vary from tissue to tissue. For example, equivalent signals for TAF II 135 and TAF II 100 are seen in the liver, lung, and adrenal gland (Fig. 2, lanes 9 -11, respectively), while the signal for TAF II 100 is stronger than that for TAF II 135 in the pituitary and the small intestine (lanes 3 and 4, respectively). The opposite relationship is observed in the eye, tongue, and spleen (lanes 2, 7, and 8, respectively). Similarly, the ratio of TAF II 55 and TAF II 100 signals changes when one compares the pituitary or the liver, where the signal for TAF II 100 is the stronger, with the heart and lung, where the opposite is seen (lanes 3, 9, 6, and 10, respectively). Therefore, not only do the expression levels of a given TAF II vary from tissue to tissue, but the relative abundance of TAF II s also varies.
The presence of several other TAF II s in these tissues was also assayed. TAF II 30 can be readily detected in most tissues with the exception of the eye and the pituitary, where only low levels of expression are seen (lanes 2 and 3, respectively). In contrast, TAF II 30 is below the limit of detection in the heart and spinal cord (lanes 6 and 17, respectively). Like the other TAF II s, it is strongly overexpressed in the testis (lane 15).
The histone fold-containing TAF II 28 can be clearly detected only in the small intestine and the testis (lanes 4 and 15, respectively), while it is barely detectable in most other tissues and undetectable in the heart, brain, kidney, and spinal cord (lanes 6, 13, 16, and 17, respectively). A similar expression pattern was seen for its heterodimeric partner TAF II 18 (27). The histone fold-containing TAF II 20 was also up-regulated in the testis and small intestine and down-regulated in heart, brain, kidney, and spinal cord.
These results again highlight some significant variations in the ratios of TAF II s present in different tissues. For example, equivalent amounts of TAF II 30 are seen in small intestine and spleen (lanes 4 and 8, respectively), while TAF II 28, TAF II 20, and TAF II 18 are down-regulated in spleen, whereas TAF II 55 is up-regulated. Similarly, equivalent amounts of TAF II 30 are seen in the brain and cerebellum (lanes 13 and 12, respectively), while all the other TAF II s are down-regulated in brain compared with cerebellum. Furthermore, while TAF II 30, TAF II 28, TAF II 20, and TAF II 18 are down-regulated in the eye; TAF II 135, TAF II 100, and TAF II 55 are expressed at levels comparable with those of several other tissues.
These results also reveal a general pattern of TAF II expression. Many TAF II s are overexpressed in the testis. This was most dramatic for TAF II 135, TAF II 100, and TAF II 30, while TAF II 55 was only mildly overexpressed. Overexpression of other RNA polymerase II transcription factors, TBP (also confirmed by this study, see below), TFIIB, and the largest subunit of RNA polymerase II, have been described previously in testis (19). It is possible that the TAF II s, like these other factors, are overexpressed in the round haploid spermatids.
In addition to testis, most of the TAF II s tested were strongly expressed in the small intestine. This is particularly true for TAF II 28, TAF II 20, and TAF II 18 which were as well expressed  Fig. 1. The positions of migration of each TAF II are indicated. As long exposures of the panels for TAF II 30 and TAF II 28 are shown, several other proteins are detected nonspecifically by these antibodies. The dash indicates the bone fide position of migration of TAF II 30 and TAF II 28 to distinguish them from closely migrating species, while "o" indicates the presence of an artifact seen with both the anti-TAF II 30 and anti-TAF II 28 antibodies. The signals for TAF II 135 and TAF II 100 in lanes 14 and 15 represent a 5-fold shorter exposure than those in the other lanes. TAF II 135, TAF II 100, and TAF II 55 were detected on the same blots as were TAF II 28, TAF II 20, and TAF II 18, while TAF II 30, which closely comigrates with TAF II 28, was taken from another blot. as in the testis. In contrast, many TAF II s were down-regulated to the point of being undetectable in tissues such as brain, heart, kidney, and spinal cord. This is also the case in the kidney with the exception of TAF II 30, which is as abundant as in intestine. Comparison of the signals observed in the brain, kidney, and lung with those obtained with serial dilutions of the small intestine extract showed that the levels of TAF II 135 and TAF II 100 were 5-fold lower in the brain and kidney than in intestine, while those in the lung were around 3-fold lower (data not shown). Note that the levels of these TAF II s are even lower in the spinal cord and heart than in the brain or kidney. Similar titrations showed that TAF II 55 levels were 10-fold lower in the brain and kidney than in the testis, while the levels in the lung were 2-3-fold lower (data not shown). This suggests that the distinct transcriptional programs of each tissue show differing requirements for a given TAF II .

Partially Complementary Expression of TBP and TLF in Mouse Tissues-
The same extracts were also tested for the expression of TBP. TBP is strongly expressed in the testis (Fig.  3A, lane 15) and in the small intestine and the pituitary (lanes 8 and 9, respectively). It is interesting that one of the highest levels of TBP is found in the pituitary, since many TAF II s are under expressed in this extract. Intermediate expression levels were detected in the adrenal, lung, liver, spleen, and tongue (lanes 2-6, respectively). Strikingly, only very low levels of TBP could be detected in the brain and cerebellum (lanes 12 and 13, respectively), and TBP was virtually undetectable in the heart, eye, kidney, and spinal cord (lanes 7, 10, 16, and 17, respectively; note that since comparable exposures of two different blots are presented, the adrenal gland was included in both to allow comparison of the left and right panels). In these experiments, TBP was detectable in brain, cerebellum, heart, eye, kidney, and spinal cord only when very long saturating exposures of the blots were made (data not shown), while the nonsaturating exposures shown in Fig. 3A highlight the differences in expression levels. Titration experiments using serial dilutions of the testis and small intestine extracts showed that TBP levels were 3-5-fold lower in the small intestine than in the testis, 5-6-fold lower in the lung, and more than 10-fold lower in the brain and kidney (data not shown). These results reveal a considerable variation in TBP expression levels among the different tissues.
The above result is rather unexpected considering the important role which TBP is thought to play in transcription. This prompted us to look at the expression of TLF, a factor highly related to the TBP core domain (18) 2 and which consequently may be able to functionally substitute for TBP.
The highest levels of TLF were detected in the adrenal, small intestine, brain, and spinal cord (Fig. 3A, lanes 2, 8, 13, and 17, respectively). TLF was also present in the liver, tongue, heart, pituitary, eye, cerebellum, and kidney (lanes 4, 6,7,9,10,12, and 16, respectively), but was undetectable in the lung and spleen (lanes 3 and 5, respectively). TLF was expressed in the testis, but in contrast to the other factors examined, it was under, rather than overexpressed, in this tissue (compare the contrasting levels of TBP and TLF in testis, lane 15, with brain or spinal cord in lanes 13 and 17, respectively, and the expression of TBP and TLF in the pituitary and eye, lanes 9 and 10, respectively). The fact that TLF expression can be readily detected in the eye, heart, spinal cord, and kidney (note also that TAF II 30 is readily detectable in the kidney extract) extracts shows that there is no intrinsic defect in these extracts which would explain the observed low levels of TAF II s and TBP. The presence of TLF in these extracts rather underlines the real differences which exist in the expression levels of TBP and TAF II s.
TLF was also present in the extracts from several cultured cell lines, being readily detected in total cell extracts from pluripotent murine F9 embryonal carcinoma cells (Fig. 3A, lane  1, and Fig. 3B, lane 2) and embryonic stem cells (Fig. 3B, lane 5) or from differentiated 3T3 fibroblasts and simian COS cells (lanes 3 and 4, respectively), but much more weakly in HeLa cells (lane 1).
Previous studies on TBP protein expression have been limited to only a few tissues and have employed polyclonal antisera. Here we have used a very sensitive monoclonal antibody against TBP that reveals unexpected and very significant variations in TBP expression. As described previously (19), TBP is overexpressed in the testis. This, however, is not unique since high expression was also observed in the small intestine and the pituitary. In contrast, TBP like many TAF II s, was strongly down-regulated in the nervous tissues, eye, kidney, and in the heart.
In many of the tissues with low TBP expression, especially those of the nervous system, prominent levels of TLF were observed. Nevertheless, TLF expression was not limited only to nervous tissues or to tissues with low TBP levels, since it was also abundantly expressed in the adrenal and the small intestine extracts. Immunohistochemistry will help determine whether TBP and TLF are overexpressed in the same cell populations in these organs. Similarly, it will be interesting to determine which cells within the nervous system express TLF. The available antibody does not yet permit such studies.
In yeast and in mammalian cells, TAF II s are essential for cell cycle progression and they regulate the expression of cell cycle genes (14, 28 -31). Thus, while there is a stringent requirement for high levels of TBP and TAF II s in proliferating cells, nothing is known concerning the requirement for these proteins in terminally differentiated tissues. Our finding that the levels of TBP and several TAF II are very dramatically reduced in several differentiated tissues suggests that there is a differential requirement for TFIID in rapidly proliferating cells versus differentiated tissues.
The low expression of TBP and many TAF II s does not reflect an absence of polymerase II transcription in these tissues. Previous measurements of polymerase II transcription rates in different organs have shown only two to three-fold reductions in the kidney and the brain compared with the liver (32). Moreover, transcription rates were lower in the lung than in the brain and kidney. There is therefore no correlation between global transcription rates and measured TBP levels.
Our results would rather suggest that in certain differentiated tissues, TLF may play an important role in very specialized transcription program. While this manuscript was in preparation, Ohbayashi et al. (33) described the expression of TLF (TBP-like protein) in a limited set of rat tissues. As observed here, TLF levels were especially high in the brain and heart. However, these authors (33) showed that recombinant TLF does not support transcription in vitro and does not bind to adenovirus E4 and major late promoter TATA boxes. Further experiments will be required to determine whether TLF is a transcription factor under more physiological conditions using extracts from TLF-expressing tissues and natural promoters. In addition, as suggested for TBP-related factor (34), TLF may itself be associated with a specialized set of associated factors programming it for the transcription of specific genes.
TAF II 28 Protein Levels Do Not Correlate with That of Its mRNA-Among the TAF II s examined here, TAF II 28 is typical of a TAF II , whose expression varies considerably from tissue to tissue. We tested whether these variations could result from differences in mRNA levels.
RT-PCR was performed with total RNA from each tissue and exon-specific TAF II 28 primers (see "Materials and Methods"). Aliquots of the reaction were analyzed after 15, 23 (shown in Fig. 4), and 30 cycles of PCR. Although the TAF II 28 protein was most abundant in testis and small intestine, yet undetectable in the heart, kidney, and spinal cord, only minor differences in the corresponding levels of mRNA were observed. These variations closely mirrored those of the HPRT control and thus are probably due to intrinsic variations in the RNA samples rather than real significant differences in TAF II 28 mRNA levels. The TAF II 28 mRNA was, however, expressed at significantly lower levels in the adrenal gland.
The above results (and Northern blotting) 3 indicate that there is no direct correlation between the levels of TAF II 28 protein and those of its mRNA. Most of the variations in TAF II 28 protein expression therefore result from post-transcriptional events. This is in keeping with previous studies that showed rather homogenous expression of several TAF II s mRNAs in normal rat tissues (20). The increases in several TAF II proteins seen in testis do not, as is the case for TBP (19), require a strong up-regulation of their mRNAs. Moreover, while TBP mRNA levels are increased in testis, correlated with increased protein expression, TBP mRNA levels are not downregulated in brain and kidney (19), where there is a considerable reduction in the corresponding protein. These results in conjunction with the previous results concerning TAF II 105 (8) would indicate that the levels of many TAF II proteins are mainly regulated post-transcriptionally.
The nature of this post-transcriptional regulation is at present unknown. It is possible that the efficiency of translation of the mRNAs varies in different tissues. Alternatively, it is interesting to note that when a given TAF II is depleted in yeast, the integrity of the TFIID complex is compromised and the levels of other TAF II s are also strongly reduced (12,35,36). This suggests that many TAF II proteins accumulate only when they are stably associated in the TFIID complex, otherwise they are be rapidly degraded. Therefore, the levels of one TAF II may indirectly control those of others, if it becomes limiting for TFIID complex assembly. In the nervous tissues, it may even be the low levels of TBP itself that are limiting for TFIID assembly. Further knowledge of how the different TAF II -containing complexes are assembled inside cells and what the limiting factors in this process are will help in understanding the mechanisms which regulate TAF II expression.