Chromatin structure and conserved sequence elements in genes encoding ribosomal proteins in Tetrahymena thermophila

Eur. J. Biochem. 210,621 -627 (1992) FEBS 1992

Chromatin structure and conserved sequence elements in genes encoding ribosomal proteins in Tetvahymena thevmophila Peter N0RGAARD, Hanne DREISIG and Karsten KRISTIANSEN Department of Molecular Biology, Odense University, Denmark

(Received August 11, 1992) - EJB 92 1163

The chromatin structure of the macronuclear genes encoding ribosomal proteins S25 and L1 in the ciliated protozoan Tetrahymena thermophila was analyzed. Using the indirect end-labelling technique, DNase-I-hypersensitive regions were located in the promoter regions as well as in the 3’ regions of the genes. The DNase-I-hypersensitive regions were present in chromatin of exponentially growing cells, where the rate of ribosomal-protein gene transcription is high, and in chromatin from starved cells, where transcription of ribosomal-protein genes is severely depressed. Micrococcal- nuclease-digestion experiments revealed that the promoter regions of the S25 gene and the L1 gene are devoid of nucleosomes in exponentially growing cells. In starved cells, no nucleosomal organisation of the promoter region of the L1 gene could be detected, whereas nucleosomal structures were discernible in the promoter region of the S25 gene. A conspicuous polypurine sequence motif, AARGGGAAA, is present within or adjacent to the DNase-I-hypersensitive regions in the promoter of the S25 and the L1 gene, and interestingly, the same motif is found also in the promoter regions of the genes encoding ribosomal proteins L21 and L37.

Although a considerable number of protein-encoding genes from ciliates has been cloned and characterized (for a compilation, see Brunk and Sadler, 1990; Csank et al., 1990), detailed functional analyses of the promoter regions of these genes are still lacking. So far, it has not been possible to develop ciliate-derived effective RNA polymerase I1 in vitro transcription systems, and few examples of successful trans- formation with class-I1 genes have been reported (Godiska et al., 1987; Yao and Yao, 1991). Thus, identification of putative promoter elements of class-I1 genes in ciliates has been restricted to computer-assisted analyses of upstream regions for known typical eukaryotic promoter elements. The most com- prehensive study of this kind was published by Brunk and Sadler (1990). Taking advantage of the fact that the coding regions of the divergently transcribed histone genes H4II and H31I are separated only by about 345 bp in Tetrahymena species, they amplified, cloned and analyzed this region from 22 species of Tetrahymena. This analysis revealed a single putative promoter element containing a canonical CCAAT box. Furthermore, Brunk and Sadler (1990) demonstrated that putative CCAAT boxes were also present in the promoter region of a number of other ciliate class-I1 genes.

We have cloned, sequenced and characterized four genes encoding ribosomal proteins S25, L1, L21 and L37, respectivc- ly, in Tetrahymena thermophilu (Nielsen et al., 1986; Rosen-

Correspondence lo K. Kristiansen, Department of Molecular Bio- logy, University of Odense, Campusvej 55, DK-5230 Odense M, Denmark

Enzymes. a-amylase, porcine pancreas a-amylase (EC 3.2.1 . l ) ; DNase I, bovine pancreas deoxyribonuclease I (EC 3.1.21.1); micrococcal nuclease, Staphylococcus aureus endonuclease (EC 3.1.31 .I); proteinase K, Tritirachium album proteinase K (bC 3.4.21.14); RNase A, bovine pancreas RNase A (EC 3.1.27.5); RNase TI , Aspergillus oryzae RNase TI (EC 3.1.27.3).

dahl et al., 1991 ; Hansen et al., 1991 ; Kristiansen et al., unpublished results). As in other eukaryotes, the expression of these genes is coordinately regulated (Dreisig et al., 1984a, b; Andreasen et al., 1984). Thus, we would expect these genes to contain common regulatory cis-acting sequence elements in their promoter regions. Indeed, as related in this paper, computer-assisted analyses of the upstream regions of these four ribosomal-protein genes revealed a set of common sequence elements constituting putative regulatory regions.

An alternative approach to identify putative regulatory cis-acting sequence elements involves mapping of nuclease- hypersensitive sites in chromatin, since such cis-acting sequence elements generally reside within or adjacent to regions exhibiting hypersensitivity to nucleases (Gross and Garrard, 1988). Accordingly, we analyzed the chromatin structure of the genes encoding ribosomal proteins S25 and L1, and corre- lated the results of these analyses with those obtained in the computer-assisted comparison of the promoter regions of the four ribosomal-protein genes cloned and sequenced so far.

MATERIALS AND METHODS

Enzymes and isotopes

Restriction endonucleases were from New England Bio- labs (Beverly, MA), RNase-free DNase 1 and proteinase K were from Boehringer Mannheim (Mannheim), RNase A, RNase T1 and micrococcal nuclease were from Sigma Chemi- cal Co. (St. Louis, MO). [a-32P]dGTP and [cY-~*P]~TTP (both 1 11 TBq/mmol, 3000 Ci/mmol) and [rnethyl-’H]thymidine (1.85 - 3.33 TBq/mmol, 50 - 90 Ci/mmol) were purchased from Du Pont Danmark A/S (Brrandby).

622

Strains and growth conditions

T. thermophila B1868 (mating type VII) was grown in complcx protcose peptone/yeast extract medium, starved in 10 mM TrisjHCl, pH 7.5, and refed with double-concentrated proteose peptonelliver extract medium (Andreasen et al., 1984). For quantitation of DNase-I digestion, cells were radiolabelled by adding 0.15 pCi/ml [~~ethyl-~H]thymidine at the time of inoculation.

Preparation of nuclei and treatment of DNA with DNase I and micrococcal nuclease.

Macronuclei were prepared from cultures at 20-25 x lo4 cellsiml according to the method of Gocke et al. (1978). For digestion with DNase I, nuclei were suspended in the nuclear buffer of Wu (1 984) to 1.5 - 1.9 x l o7 nucleiiml corresponding to 236-295 pg/ml DNA. DNase I, at the concentrations stated in the figures, was added to aliquots of nuclear suspensions after equilibration at 25°C for 2 min. Digestion was allowed to proceed for 6 min, stopped by the adjustment of EDTA to 20 mM. NaC1 to 500 mM, proteinase K to 200 pg/ ml and SDS to 0.5%, followed by incubation at 50°C for 2 h. Macronuclear DNA was purified by repeated extraction with phenol/CHC13/isoamyl alcohol and precipitated with isopro- panol. For digestion with micrococcal nuclease, nuclei were resuspended in the micrococcal-nuclease buffer of Bonven and Westergaard (1 982) to the same density as that described above. Nuclear suspensions were allowed to equilibrate at 25 C for 2 min. Micrococcal-nuclease concentrations were as stated in the figures. After 4 min, digestion was stopped and DNA purified as described above. After resuspension the DNA was treated with 100 pgjml RNase A, 100 U/ml RNase T1 and 20 pg/ml a-amylase for 1 h at 37"C, then subjected to another round of purification as described above. Finally, DNA was resuspended in 10 mM Tris/HCI, pH 8.0, and 1 mM EDTA.

Quantitation of the extent of DNase-1 digestion

Duplicate aliquots were removed from each sample after proteinase-K treatment. Half of each aliquot was spotted di- rectly onto a GFjC filter and dried down. The other half was precipitated with 10% trichloroacetic acid/0.5% sodium pyrophosphate, filtered through a GFjC filter, washed with 30 ml 5% trichloroacetic acid followed by 2 ml 96% ethanol, then dried under vacuum at 80°C. Radioactivity in filters was measured in a mark-I11 liquid-scintillation spectrometer using a xylene-based scintillation cocktail. 3 - 5"h total DNA was made trichloroacetic acid soluble with the highest DNase-I concentrations used in the experiments.

Agarose-gel electrophoresis, Southern blotting and hybridization conditions

For DNase-I experiments, macronuclear DNA was restricted with the appropriate restriction endonuclease according to the manufacturers recommendations, clectrophoresed on 1.2% horizontal agsrose gels and blotted onto BA85 nitrocellulose membranes (Schleicher & Schuell, Keene, NH) with 3 M NaCl and 0.3 M sodium citrate, pH 7.0. For micrococcal- nuclease experiments, the isolated DNA was electrophoresed on 1.4% horizontal agarose gels and blotted onto Zeta-Pro- beTM nylon membranes (Bio-Rad, Richmond, CA) with 0.4 M NaOH using a transfer time of 4 h. Following transfer of

DNA, nitrocellulose membranes were baked for 2 h and nylon membranes for 30 min under vacuum at 80°C. Gel-purified DNA fragments were used as probes and radiolabelled by the random primer procedure of Feinberg and Vogelstein (1983) according to the modifications proposed by Hodgson and Fisk (1 987). For AIT-rich upstream probes, labelling was with [ u - ~ ~ P I ~ T T P , otherwise [ x - ~ ~ P I ~ C T P was used. Stringent hybridization and washing conditions were employed (Sam- brook et al., 1989). Membranes were exposed at -70°C to Fuji RX X-ray films with Du Pont Cronex Lightning Plus intensifying screens for 6 - 60 h.

RESULTS

DNase-I-hypersensitive regions in genes encoding ribosomal proteins S25 and L1

To map DNase-I-hypersensitive sites in genes encoding ribosomal proteins S25 and L1 and their flanking regions, the indirect end-labelling procedure of Wu (1980) was employed. Nuclei were isolated from cxponcntially growing cclls and from cells starved for 24 h, and digested lightly with DNase 1. The DNA was isolated, digested to completion with the appropriate restriction endonucleases, electrophoresed on agarose gels, blotted onto nitrocellulose, and hybridized with appropriate probes. Fig. 1 depicts the architecture of the genes encoding ribosomal proteins S25 and L1, and, furthermore, indicates the probes used for mapping of DNase-l-hypersensitive sites in the 5' and the 3' ends of the genes, as well as the probes used for analysis of the nucleosomal organization of the genes.

Fig. 2 shows that DNase-I-hypersensitive regions are Io- cated in the promoter regions of the S25 and L1 genes. In particularly, in the S25 gene, hypersensitivity to DNase I is prominent. In both genes, two closely spaced regions exhibit hypersensitivity to DNase I. In the S25 gene, thcse sites are located 100 - 200 bp and 250 - 400 bp upstream of the translation-start codon. In the L1 gene, the hypersensitive sites are located 60- 130 bp and 160 - 240 bp upstream of the translation-start codon. The constitutive nature of these DNase- I-hypersensitive sitcs is apparent. In cells starved for 24 h, transcription of ribosomal-protein genes is severely curtailed (Andreasen and Kristiansen, unpublished rcsults). Neverthe- less, the pattern and degree of DNase-I hypersensitivity is unaffected. The difference in transcriptional activity, however, is clearly reflected in thc heightened general DNase-I sensi- tivity of the genes in chromatin from exponentially growing cclls wherc the rate of ribosomal-protein transcription is high. As a control, purified macronuclear DNA was digested with DNase I and appropriate restriction enzymcs and analyzed as dcscribed for isolated nuclei. No DNase-I-hypersensitive sitcs could be detected (not shown).

Fig. 3 illustrates the results of mapping of DNase-I-hypersensitive sites in the 3' end of the genes by indirect end- labelling. As in the promoter region, DNase-I-hypersensitive sites are most prominent in the S25 gene, but a DNase-I- hypersensitive region is clearly present also in the L1 gene. All DNase-I-hypersensitive sites are present in chromatin from exponentially growing cells as well as in chromatin from cells starved for 24 h.

Fig. 4 summarizes the locations of DNasc-I-hypersensitive regions in the S25 and L1 genes. In both gcnes, constitutive DNase-I-hypersensitive regions flank the bodies of the genes. In the promoter region of the S25 gene, the hypersensitive sites encompass two closely spaced regions of 100 bp and

623

E E E

s25 gene I / 1 INTRON

I /

- MN probe - MN probe - DNase I probe - DNase 1 probe 200 bp Y

Fig. 1. The architecture of the genes encoding ribosomal proteins S25 and L1 in T. themophila, and the location of probes used for analyses of DNase- I-hypersensitive sites and nucleosomal organization by micrococcal-nuclease-digestion experiments. In the map which shows thc organization of the genes, boxes indicate exons and coding regions are hatched. The regions covered by the probes are indicated by filled boxes below each map of the genes. A, AhalII; B, BglII; Bs, BspHl; E, EcoRI; E5, EroRV; H, HinDIII; HI, HpaI; H2, HpaII; N, NsiT; R, RsuI; MN, micrococcal nuclease.

Fig. 2. Mapping of DNase-I-hypersensitive sites in the promoter region of the genes encoding ribosomal proteins S25 and L1. Nuclei isolated from exponentially growing cells and cells starved for 24 h were incubated with increasing amounts of DNase I: lanc 1, undigcsted; lane 2, 0.1 U/ml; lane 3, 0.25 U/ml; lane 4, 0.50 Uiml; lane 5 , 0.75 U/ml; lanc 6, 1.00 U/ml; lane 7, 2.00 U/m1. The DNA was extracted and digested with EcoRI for analysis of the S25 gene, and RsaI for analysis of the L1 gene. DNA rragments were separated by electrophoresis in I .2% agarose gels, blotted onto BA85 nitrocelldose membranes and hybridized to appropriate probes. The S25 probe was a 415-bp HpnII - EcoRI fragment, and the L1 probe was an L1 cDNA fragment of 451 bp, starting at T in the translation-start codon ATG of the L1 gene and ending at the RsaI sitc at the 3’ end of exon 1 (Fig. 1) .

Fig.3. Mapping of DNase-I-hypcrsensitive sites in the 3’ end of the genes encoding ribosomal proteins S25 and L1. Nuclei isolated from exponentially growing cells and cells starved for 24 h were incubated with increasing amounts of DNase I. Lane 1, undigested; lane 2, 0.1 Ujml; lane 3 , 0.25 U/ml; lane 4, 0.50 U/ml; lane 5 , 0.75 Ulrnl; lanc 6, 1.00 U/ml; lane 7, 2.00 U/ml. The DNA was extracted and digested with EcoRI for analysis of the S25 gene, and BgKI and Hind111 for analysis of the L1 gene. DNA fragments were separated by electrophoresis in 1.2% agarosc gcls, blotted onto BA85 nitrocellulose membranes and hybridized to appropriate probes. The S25 probe was a 400-bp EcoRI-EgZII fragment, and the L1 probe was a 756 bp BgnI - BspI fragrncnt (Fig. 1).

624

S25 gene

-130, i60

INTRON L1 gene

200 bp

Fig. 4. Location of DNase-I-hypersensitive regions in genes encoding ribosomal proteins S25 and L1. Filled boxes indicate DNase-I-hypersensitive regions; open boxes indicate non-coding rcgions; hatched boxes are coding regions of exons. Locations are given relative to the translation- start codon ATG where the first nucleotide upstream of A is - 1, and A is + 1.

150 bp, just upstream of the two mapped transcription-start points. A similar organization of hypersensitive sites is observed in the promoter region of the L1 gene, although in this gene the hypersensitive regions appear to be more narrowly defined, spanning 70 bp and 80 bp.

In the S25 gene, a hypersensitive region of 150 bp is located in the 3’ end of the single intron found in this gene (corresponding to fragments of 350-500 bp in Fig. 3). The probe used for analysis of hypersensitive sites in the 3’ end of the L1 gene did not permit mapping of DNase-I-hypersensitive sites in the corresponding intron region in this gene. A DNase-I- hypersensitive region coincides with the region of the S25 gene encoding the 3‘ untranslated trailer (corresponding to fragments of 950 - 11 50 bp in Fig. 3). Finally, downstream of the polyadenylation site in both genes, a DNase-l-hypersensitive region is present. In the S25 gene, this region encompasses 200 bp (corresponding to fragments of 1500- 1700 bp in Fig. 3), in the L1 gene, 250 bp (corresponding to fragments of 1650 - 1900 bp in Fig. 3) . These hypersensitive sites may be related to transcription termination and/or polyadenylation, in analogy with hypersensitive regions present in the 3’ end in class-I1 genes in yeast (Ptrez-Ortin et al., 1989).

Nucleosomal organization of the genes encoding ribosomal proteins S25 and L1

Since little is known about the nucleosomal organization of class-I1 genes in Tetrahymena, and DNase-I-hypersensitive sites in general are characterized by the absence of nucleosomes (Gross and Garrard, 1988; Pkrez-Ortin et al., 1989), the nucleosomal organization of the S25 and L1 genes were analyzed. Nuclei wcre isolated from exponentially growing cells and from cells starved for 24 h, then digested with micrococcal nuclease. DNA was extracted, subjected to electrophoresis in 1.4% agarose gels, blotted onto Zeta-Pro- beTM and hybridized to probes specific for the promoter regions and exons of the S25 and L1 genes, respectively.

Fig. 5 shows ethidium-bromide-stained gels of micrococcal-nuclease-digested chromatin from exponentially growing and starved cells, both exhibiting the typical digestion pattern of nucleosomal organized chromatin. The slightly fuzzy pattern is characteristic of macronuclear chromatin in Tetrahymena (Gorovsky, 1986). To ensure comparable hybridization and exposure, a control lane of restricted macronuclear DNA was included on each gel. Furthermore, to ensure that lack of hybridi~ation with a given probe was not due

Growing cells Starved cells 1 2 3 4 6 5 7 8 9 1 2 3 4 5 6 7 8 9

955 * 585 +

341 P 258

141 *

Fig. 5. Micrococcal-nuclease digestion of chromatin from exponentially growing cells and cells starved for 24 h. Nuclei isolated from cells growing exponentially and from cells starved for 24 h were incubated with micrococcal nuclease. Lane 1 , purified macronuclear DNA rc- stricted with RsaI; lane 2, undigested; lane 3, 0.01 U/ml; lane 4, 0.025 U/ml; lane 5 , 0.050 U/ml; lane 6, 0.100 U/ml; lane 7, 0.250 U/ ml; lane 8, 0.500 U/ml; lane 9, 1.00 U/rnl. DNA was extracted, run on horizontal 1.4% agarose gels and stained with ethidium bromide.

to loss of DNA from the filters each filter was analyzed in succession with the different probes in the following order: exon-specific probe; upstream-specific probe; exon-specific probe. The probes were selected so that for a given gene, the upstream probe and the exon probe were of comparable size. The L1 exon-specific probe was a 71 8-bp EcoRV - HpaI fragment, and the L1 upstream-specific probe was a 780-bp RsaI - EcoRV fragment. The S25 exon-specific probe was the 556- bp cDNA insert of prB2 (Nielsen et al., 1Y86), and the S25 upstream-specific probe was a 396-bp Aha111 - NsiI fragment (Fig. 1).

Fig. 6 shows the results obtained when blots of DNA from micrococcal-nuclease-digested chromatin of exponentially growing and starved cells were probed with the L1 exon- specific and upstream-specific probes. In exponentially growing cells, the pattern of the hybridization signals of the L1 exon-specific probe is clearly nucleosomal. In the starved cells, the pattern is also nucleosomal, but less distinct. In contrast, using the L1 upstream-specific probe, no nucleosomal patterns were observed irrespective of whether blots of DNA

585 - 341 > 258

585 D

341 - 258 - 141 *

Fig. 6. Nucleosomal organization of the gene encoding ribosomal protein L1 in exponentially growing cells and cells starved for 24 h. Nuclei from cells growing exponentially or fromcells starved for 24 h were isolated and treated with micrococcal nuclease. Lane 1, purified macronuclear DNA restricted with RsaI; lane 2, undigested; lane 3 , O . O l U/ml; lane 4,0.025 U/ml; lane 5,0.050 U/ml; lane 6,0.200 U/ml; lane 7,0.250 U/ ml; lane 8,0.500 U/ml; lane 9, 1.00 U/ml. After extraction, the DNA was clcctrophorescd, blotted onto (-ProbeTM nylon membranes and hybridized in succession to the L1 exon-1-specific 718-bp EcoRV- HpuI fragment and the L1 upstream-specilic 780-bp RsaI - EcoRV fragment.

from exponentially growing or starved cells were hybridized. Using S25-specific probes, the same type of results were obtained, except that a weak nucleosomal pattern was observed in the S25 upstream region in chromatin from starved cells. In exponentially growing cells, the S25 upstream region appeared devoid of nucleosomes, like the L1 upstream region (Fig. 7). An attempt to analyze the nucleosomal organization using the indirect end-labelling technique resulted in no dis- crete patterns of micrococcal nuclease cleavage or protection (not shown), indicating that nucleosomes are not phased in the transcribed regions of the genes.

A common sequence element in the promoter region of genes encoding ribosomal proteins in T. thermophila

The sequences of the DNase-I-hypersensitive sites and their flanking regions of the S25 and L1 genes were compared with sequences of the promoter regions of the genes encoding ribosomal proteins L21 (Rosendahl et al., 1991) and L37 (Hansen et al., 1991) using the Beckman MicroGenieTM program. A polypurine sequence motif, AARGGGAAA, was found to be located within or adjacent to the DNase-I-hypersensitive regions of the S25 and L1 genes, and the same motif

Fig. 7. Nucleosomal organization of the gene encoding ribosomal protein S25 in exponentially growing cells and cells starved for 24 h. Nuclei from cells growing exponentially or from cells starved for 24 h were isolated and treated with micrococcal nuclease. Lane 1, purified macronuclear DNA restricted with Rsal; lane 2, undigested; lane 3, 0.01 U/ml; lane 4, 0.025 U/ml; lane 5 , 0.050 U/ml; lane 6, 0.300 U/ ml; lanc 7, 0.250 U/ml; lane 8, 0.500 U/ml; lane 9, 1.00 U/ml. After extraction, the DNA was electrophoresed, blotted onto Zeta-ProbeTM nylon membranes and hybridized in succession to the S25 exon- specific 556-bp cDNA insert of prB2 (Nielscn et al., 1986) and the S25 upstream-specific 396-bp AhnIII -NsiI fragment.

was also present in the promoter region of the genes encoding ribosomal proteins L21 and L37. In the S25 and L1 genes, this motif occurs twice, whereas the L21 and L37 genes each harbor one copy. Table 1 summarizes the locations of these sequence elements in the four ribosomal-protein genes. Con- sidering the scarcity of G (and C) in the promoter region of Tetruhynzenu class-I1 genes (Nielsen et al., 1986; Rosendahl et al., 1991; Hansen et al., 1991), it is highly improbable that this G-rich motif should occur purely by chance in comparable locations in four of four sequenced ribosomal-protein genes. Therefore, we consider this purine-rich sequence as a good candidate for a functionally important promoter element in ribosomal-protein genes in T. thermophila.

DISCUSSION

Our results show that the genes encoding ribosomal proteins L1 and S25 in T. thermophilu are flanked by constitutive DNase-I-hypersensitive regions. Furthermore, we present evi- dence that in chromatin from exponentially growing cells, where the rate of ribosomal-protein gene transcription is high, the promoter regions of the L1 and S25 genes are not or-

626

Table 1. The location of the G-rich polypurine sequence element in the promoter regions of ribosomal-protein genes in T. thermophila. The promoter regions of thc T. thermophila ribosomai-protein genes were analyzed using the Beckman MicroGenieTM program Compare. The locations of the GGGAA polypurine motif, the DNase-I-hypersensitive regions and the transcription-start points are given relative to the translation-start codon ATG, where the first nucleotide upstream of A is - 1 . References: S25 (Nielsen et al., 1986); L1 (Kristiansen el al., unpublished results); L21 (Rosendahl et al., 1991); L37 (Hansen et al., 1991). n.d., not determined.

Gene GGGAA polypurine Location DNase-I-hypersensitive Transcription-start points sequence motif region

S25 AAGGGGAAA -427 to -419 -400 to -250 -95, -70 AACIGGGAAA -158 to -150 -200 to -100

AA AGGGAAT -120 to -112 -330t0 - 60 L1 A AAGGGAAA -145 to -137 -240 to -160 -63, -61, -57, -53, -47

L2 1 AGAGGGAAA - 86 to - 78 n.d. -73, -71, -65, -61, -56 L3 7 AAAGGGAAA -271 to -263 n.d. -63, -61, -49, -46, -39

ganized into nucleosomes, whereas the transcribed regions clearly are. Nucleosome removal or displacement, onc of the essential steps in the generation of DNase-I-hypersensitive sitcs and transcription-competent chromatin, is often me- diated by the binding of trans-acting protein factors. Alterna- tively, the intrinsic structure of the DNA may prevent its assembly into nucleosomes (reviewed in Gross and Garrard, 1988; PCrez-Ortin et al., 1989). The presence of nucleosomes in the promoter region has been shown to repress initiation of transcription (Lorch et al., 1987). Conversely, nucleosome removal has been shown to correlate with active transcription initiation (Han et al., 1988) and to activate downstream pro- moters (Han and Grunstein, 1988). In the L1 gene, the promoter region appears to be devoid of nucleosomes, irrespective of the transcriptional state of the gene. In the S25 gene, no nucleosomes could be detected in chromatin from cells with a high level of transcription of the gene, whereas Faint hybridization signals reflecting possible nucleosomal organization of the promoter region were observed using chromatin from starved cells.

Constitutive nuclease-hypersensitive regions have been shown to flank class-I (Bonven and Westergaard, 1982; Palen and Cech, 1984; Budarf and Blackburn, 1986), class-I1 (Pcdcrson et al., 1986) and class-I11 (Pederson et al., 1984) genes in T. therrnophilcl, and it appears that these sites are generated during macronuclear development following conju- gation (Pederson et al., 1986). Thus, analyzing the chromatin structure of the histone-H4-I gene in T. therrnophila, Pederson et al. (1986) demonstrated that the constitutive nuclease-hypersensitive sitcs which flank the the macronuclear gene were absent in the micronuclear copy.

We have not analyzed the organization of the S25 and L1 genes in the transcriptionally inactive micronucleus, so, formally, the presence of DNase-I-hypersensitive sites in the promoter regions of the macronuclear S25 and L1 gcnes could be due to intrinsic structural properties of the DNA. However, the organization and substructure of the hypersensitive regions in the promoter region of the S25 and L1 genes are remarkably similar to thosc of the histone-H4-1 gene, and the occurrence of a protected region within closely spaced hypersensitive sites is consistent with the binding of a trans- acting factor.

Computer-assisted analysis of the promoter regions of the S25 and L1 genes showed that a common polypurinc motif was present within or adjacent to the DNase-I-hypersensitive regions, and that the same sequence motif was also present in the promoter region of genes encoding ribosomal proteins L21 and L37 (Table 1). Interestingly, our analysis showed that

the motif AARGGGAAA is also found within the DNase-I- hypersensitive region of the histone-H4-I gene, lending further support to the notion that this sequence element is of functional importance. Thus, this sequence element might function in analogy with the yeast UAS,,, element, which is functionally important for controlling the exprcssion of many ribosomal-protein genes in yeast, but is also involved in controlling the expression of a number of genes unrelated to ribosomes or protein synthesis (Mager, 1988; Warner, 1989).

We thank Carina Nymann for expert technical assistance. This work was supported by grants from the Danish Natural Science Rcscarch Council, the Carlsberg Foundation and the Ib IIenriksen Foundation.

REFERENCES Andreasen, P. H., Dreisig, H. & Kristiansen, K . (1984) Eur. J . Bio-

Bonvcn, B. J. & Wcstcrgaard, 0. (1982) Nucleic Acids Rr,s. 2-?, 7593-

Rrunk, C. F. & Sadler, L. A. (1990) Nucleic Acids Res. 18, 323 - 329. Budarf, M. L. & Blackburn, E. H. (1986) J . Biol. Chem. 261, 363-

Csank, C., Taylor, F. M. & Martindale, D. W. (1990) Nucleic Acids

Dreisig, H.. Andrcasen, P. H. & Kristiansen, K. (1984a) Eur. J .

Dreisig, H., Andreasen, P. H. & Kristiansen, K. (1984b) Eur. J .

Feinberg, A. P. & Vogclstein, B. (1983) Anal. Biochem. 132, 6-13. Gocke, E., Leer. J. C., Nielsen, 0. F. & Westergaard, 0. (1978) Nucleic

Acids Res. 5 , 3993-4006. Godiska, R., Aufderheide, K. J., Gillcy, D., Hendrie, P., Fitzwater,

T., Preer. L. B., Polisky, R. & Preer, Jr., J. R. (1987) Yroc. Nut1 Acad. Sri. 84, 7590- 7594.

Gorovsky, M. A. (1986) in The molecular biology of'cilintedprotozoa (Gall, J . G., ed.) pp. 227-261, Academic Press, Inc., New York.

Gross, D. S. & Garrard, W. T. (1988) Annu. Rev. Biochem. 57, 157- 197.

Han, M. A'L Grunstein, M. (1988) CelZ5.5, 1137-1145. Han, M., Kim, U.-J., Kayne, P. & Grunstein, M. (1988) EMBO J . 7,

Hansen, T. S . , Andreascn, P. H., Dreisig, H., HnJrup, P., Nielsen. H.. Engberg, J. & Kristiansen, K. (1991) Gene (Amsf.) 105, 143-150.

Hodgson, C. P. 81 Fisk, R. Z. (1987) Nucleic Acids Res. 15, 6295. Lorch, Y . , LaPoinle, J. W. & Koriiberg, R. D. (1987) Cell 49, 203 -

Mager, W. H. (1988) Biochim. Biophys. Actn 949. 1 - 15. Nielsen, I I . , Andreasen, P. H., Dreisig, H., Kristiansen, K. & Engberg,

rhem. 140,485-492.

7608.

369.

RPS. 18, 51 33 - 5141.

Biochern. 140, 469-475.

Biorhem. 140,476-484.

2221 -2228.

210.

J. (1986) EMBO J. 5, 2711 -2717.

627

Palen, T. E. & Cech, T. R. (1984) Cell 36,937 -942. Pcderson, D. S., Shupe, K . & Gorovsky, M. A. (1984) Nucleic Acids

Pederson, D. S., Shupe, K., Bannon, G. A. & Gorovsky, M. (1986)

Ptrez-Ortin, J. E., Matallana, E. & Franco, L. (1989) Yeast 5, 219-

Rosendahl. G., Andreasen, P. H. & Kristiansen, K. (1991) Gene

Res. 12, 8489 - 8507.

Mol. Cell Biol. 6, 3014-3017.

238.

[Amst.) 98, 161 - 161.

Sambrook, J.. Fritsch, E. F. &Maniatis, T. (1989) Molecular cloning; a laborator-v manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

M'arner, J. R. (1989) Microbiol. Rev. 53, 256-271. Wu, C. (1980) Nature 286, 854-860. Wu, C. (1984) h'ature 311, 81 -84. Yao, M.-C. &Yao,C.-H. (1991) Proc. NatlAcud. Sci. IiSA88,9493-

9497.

Documents

Chromatin structure and conserved sequence elements in genes encoding ribosomal proteins in Tetrahymena thermophila