.=) 1991 Oxford University Press

Nucleotide sequence histone cDNA

Nucleic Acids Research, Vol. 19, No. 7 1491

and

expression

of a

maize Hi

P.Razafimahatratra, N.Chaubet, G.Philipps and C.Gigot* Institut de Biologie Moleculaire des Plantes du CNRS, Universite Louis Pasteur, 12 rue du General Zimmer, 67084 Strasbourg Cedex, France Received January 8, 1991; Revised and Accepted March 1, 1991

ABSTRACT The first complete amino acid sequence of a Hi histone of a monocotyledonous plant was deduced from a cDNA isolated from a maize library. The encoded Hi protein is 245 amino acid-long and shows the classical tripartite organization of this class of histones. The central globular region of 76 residues shows 60% sequence homology with HI proteins from dicots but only 20% with the animal Hi proteins. However, several of the amino acids considered as being important in the structure of the nucleosome are conserved between this protein and its animal counterparts. The N-terminal region contains an equal number of acidic and basic residues which appears as a general feature of plant Hi proteins. The 124 residue long and highly basic C-terminal region contains a 7-fold repeated element KA/PKXA/PAKA/PK. Southern-blot hybridization showed that the HI protein is encoded by a small multigene family. Highly homologous Hi gene families were also detected in the genomes of several more or less closely related plant species. The general expression pattern of these genes was not significantly different from that of the genes encoding the corehistones neither during germination nor in the different tissues of adult maize. INTRODUCTION For several reasons the histone HI occupies a particular position major histones ubiquitously found in eukaryotes. Interacting with the intemucleosomal DNA the histone HI has been shown to play a specific role in the higher structure of the chromatin (1). This structural role in addition to several observations of tissue specificity related to physiological changes have strongly suggested that HI also plays a role in regulation of gene expression. A very recent review (2) shows that not only is HI considered as a general repressor of gene expression but also as a control element of the transcriptional activity of particular genes. Although such a role seems to be more and more documented the basic mechanism by which HI could control gene expression is far from being understood. Moreover, it is still not clear whether or not the 'active chromatin' contains HI or not.

among the five

*

To whom correspondence should be addressed

EMBL accession

no.

X57077

That HI plays an important structural role at the level of the chromatin is also suggested by the fact that it is one of the major targets for kinases intervening at specific steps of the cell-cycle (3). However, the possible role of the cyclic phosphorylation of HI during the cell-cycle is still a matter of speculations. At the level of its primary structure HI is the less conserved histone among the five major classes typically found in higher eukaryotes. It has been shown that in many organisms several HI subtypes differing in size, sequence and in their ability to condense chromatin could coexist (4). Nevertheless, despite such an extraordinary variability, all the Hi subtypes, including Hi0 and H5, share the same organization into three domains: a conserved central hydrophobic region, flanked by highly variable N- and C- terminal domains. The conserved globular region was shown to seal up the DNA at its entry and exit of the nucleosome (1). The N-terminal region was suggested to play a role in positioning the central domain with respect to the nucleosome (5) while the highly basic C-terminal domain is involved in the formation of higher order structure of the chromatin (1,6). HI histones as well as the corresponding genes have been extensively studied in a wide variety of animal species, leading to the general representation of the structure of the HI protein discussed above. Multiple subtypes of HI were also detected in several plant species (7) and immunological studies have shown that they also contain highly conserved elements as compared to their animal counterparts (8). Up to now the complete amino acid sequences of only 2 plant Hi histones have been deduced from the nucleotide sequences of a pea Hi cDNA (9) and an Arabidopsis genomic clone (Gantt, personal communication). Very partial sequences of the globular regions of two wheat H1 subtypes (10) and a maize H1 (I1) had been published earlier. All these sequences show several conserved elements within the central globular regions but highly divergent N- and C-terminal domains. Here we present the nucleotide sequence of a nearly full-length HI cDNA encoding a maize HI histone. The cDNA was used as a probe to study the expression of the corresponding gene under different physiological conditions. Southern-blot analysis showed that the corresponding gene belongs to a small multigene family which also exists in the genomes of several plants related to maize.

1492 Nucleic Acids Research, Vol. 19, No. 7

MATERIALS AND METHODS Preparation of a polyclonal antiserum Total histones were extracted from purified nuclei prepared from 8-day old maize plantlets and the histones Hi were purified as previously described (12). The purified HI histones were injected to rabbits and the specificity and titer of the antiserum was controlled by Western-blotting. The serum was shown to react exclusively with the three major variants of histone HI identified in maize.

Hi cDNA: purification and analysis A Xgtll cDNA library constructed from 11 day-old maize plantlets poly(A)+ RNAs (Clontech) was screened using the anti HI polyclonal antiserum. Positive clones were revealed using [1251] iodinated protein A. Several positive clones were revealed and one of them, cH1C21 was further plaque purified and the insert transferred into the plasmid pUC9. A series of overlapping deletions was generated using exonuclease 111 digestion (doublestranded nested deletions kit Amersham) and adequate deletions were sequenced using the dideoxy chain termination method (13).

Northern-blot analysis Total RNAs were extracted from frozen manually excised embryos or different tissues using the phenol/chloroformlisoamyl alcohol (25:24:1) technique previously described (16). The purified RNAs were analysed by formaldehyde agarose gel electrophoresis (17) and subsequently transferred to Hybond N nylon membranes (Amersham). The hybridization conditions were as indicated for Southern blots.

RESULTS Characterization of a maize Hi cDNA and corresponding protein A recombinant phage showing strong immunological reaction with a polyclonal antiserum raised against purified maize HI histones was plaque purified through three rounds of screenings. The complete nucleotide sequence of the 1.1 kbp insert was established (accession number X57077). The translation of cH1C21 shows that it contains an open reading frame of 735 nucleotides encoding a protein of 245 amino acids. The 5' flanking region is only 19 nucleotide-long and is probably uncomplete as it has been shown that several maize H3 and H4 mRNAs start approximately 70 nt upstream from the initiation AUG (16). The 3' untranslated region is 318 nt long and is followed by a 36 nt long poly(A) tract. The presence of a poly(A) tail and a long 3'-untranslated region (3'-UTR) shows that the

*

v

v

v

IV

50

v.**

ATDVTETPAPVD 70

A

APK

90

TSDDH

110

A;8A

t_

170

150

130

230

210

190

*K AGXK

.CSKDPKKPIUWKKAPVKP

.

.

KPA

.

.

B

A Camrd

tea"

A

C

Southern-blot analysis The DNA was extracted from frozen tissues as previously described (14). After digestion with the appropriate restriction endonuclease and analysis on a 0.8% agarose gel, the DNA fragments were transferred onto a Hybond N (Amersham) nylon membrane (15) under the conditions indicated by the manufacturer. When stringent conditions were applied, hybridization with the random-primed 32P-labelled probe was during 24 to 48 h at 42°C in hybridization medium containing 50% formamide. Carrier DNA was sonicated E. coli DNA. The membrane was subsequently washed twice for 15 min. in 2xSSC; 0.1% SDS, and for 5-10 min. in0.lxSSC; 0.1% SDS at 420C.

30

10 v

+-

1

0.00 000

0

.T

-

-10

+

.

J

00

*

00

*

0 0 0

_4)4+

+

+

. .----20 ----30 . .---40

Bc Conserved

real& A 00 @000 +++@

*- -

50

- -

0 @00 +++ 00000 +@+000 000 +:j

60 +........ 70

-.-.-

77

Figre 1. Remarkable elements in the amino acid sequence of the maize HI protein deduced from the nucleotide sequence of cHIC21. (A) Tripartite organization of the maize HI protein. N-terminal region (residues 1 to 45): black and open arrow-heads indicate the acidic and basic amino acids respectively. The central globular region (residues 46 to 121) is underlined and the amino acids conserved between plant and animal HI proteins are boxed. The repeated elements of the C-terminal region (residues 122 to 245) are underlined with broken lines. (B) Sequence comparison of the globular regions of HI proteins from different species. a, b: partial sequences of 2 wheat HI variants (10). c: partial sequence of a maize HI histone (I1). d: sequence of the maize HI protein deduced from the nucleotide sequence of cHIC21 (this article). e, f: sequences of HI variants of pea (9) and Arabidopsis (Gantt, personal communication), respectively. g: animal consensus sequence deduced from the comparison of 30 sequences of HI proteins (23). Amino acids are numbered from 1 to 77 and gaps have been introduced to optimize alignment. The open bars named A, B and C delimit the helical domains determiined on chicken H5 histone (25). Circles indicate the conserved amino acids among plant HI proteins and black circles those also conserved between plants and animals. Under the amino acid alignments are indicated the positions of the basic (+) and acidic (-) residues conserved among plants (line P) and their position in the animal consensus (line A). The circled signs represent those which were suggested in animal HIs to interact with the internucleosomal DNA.

gene(s) encoding this H I subtype is transcribed into polyadenylated mRNAs as previously shown for H3 and H4 genes in different plant species including maize (18,19). This structure constitutes one of the major differences with the animal replication-dependent histone genes which are transcribed into non-polyadenylated mRNAs with short 3'-UTRs (20,2 1). Furthermore, the sequence AATGAAGA located 33 nt upstream of the 3'-end is very similar to the consensus motif A/GATGAAATG found at equivalent positions in several plant H3 and H4 genes and which is supposed to correspond to a polyadenylation signal. Within the coding region, only 12.5% of the codons have A or T at the third position. Such a strickingly biased codon usage

Nucleic Acids Research, Vol. 19, No. 7 1493 M T C Mi S B 0 W A s

M

T

C

Mi

S

B

0

W

A

0

~~~~kb Hi

12 24 36 48 .......

72 96 108 120 Hrs .......

H3 .14.3

2~~~~~~

>

Figure 3. Expression of HlC21 during germination (0 to 120 hours). 25 ug of total RNA extrd from maize embryos were analyzed by agarose-formaldehyde gel electrophoresis and transferred onto Hybond N nylon membrane. The northernblot was successively hybridized to radioactively labelled cH 1C21 and the maize H3 coding region as probes.

>

6 1.6

A

P

B

Figure 2. Southern blot hybridization of DNA from different plant species more or less related to maize with radioactively labelled cHlC21. 3 jg DNA from each plant was digested with EcoRI (A) or BamHI (B), analysed on a 0.8% agarose gel, transferred to a nylon membrane, and the resulting genomic-blot hybridized to cHlC21 under stringent conditions (42°C, 50% formamide). M: maize (Zea mays, inbred F7), T: teosinte (Zea diploperennis), C: Coix lacryma jobi, Mi: millet (Pennisetum typhoides), S: sorgho (Sorghum bicolor-Texas 615), B: barley (Hordeum vulgare), 0: oat (Avena sativa), W: wheat (Triticum aestivum), A (Arabidopsis thaliana).

was also found in all the genes encoding the histones H3 and H4 cloned from three different gramineae but not in the corresponding genes of dicotyledonous plants (22). Whether it plays any role on the levels of transcription or translation is still a matter of speculation. The primary structure of the protein deduced from the nucleotide sequence is of particular interest as it is only the third plant histone Hi to be completely sequenced and the first from a mono-cotyledonous species. The predicted protein contains 245 residues and has a molecular weight of 25,500 daltons (Fig. IA). This size corresponds to that of one of the major variants of the histones HI in maize (12). The amino acid composition is very similar to other Hi histones with high numbers of lysine (60), alanine (59) and proline (35) residues and a very high lysine to arginine ratio (60/6). Like the other HI proteins studied up to now, the maize HI has a typical tripartite organization with a N-terminal region of 45 amino acids, a central globular region of 76 amino acids and a C-terminal domain of 124 residues. As in other HI proteins, the N-terminal region can be subdivided into an acidic domain (residue 1-30) containing 6 acidic residues followed by a basic part (residue 31-45) containing 6 basic amino acids (Fig. IA). The number of acidic residues is higher than in any other animal HI N-terminal region (23) and equals the number of basic residues. The same kind of organization exists in the two HI proteins sequenced from dicots: the pea (9) and Arabidopsis (Gantt, personal communication) H1 histones contain in their N-terminal regions 12 and 9 acidic and II and 13 basic residues respectively. However, in addition to the two classical acidic and basic domains existing in the maize and animal HIs, the N-terminal regions of these two proteins show a central part where acidic and basic amino acids are intermingled. The C-terminal region is 124 amino acid-long (residues 122-245) and contains a high number of basic residues as in all the other HI proteins from animals and plants. Of particular

interest is the presence of a 7 times repeated motif of 9 amino acids with a general consensus sequence: KA/PKXA/PAKA/PK (Fig. IA). The existence of such repeated peptides mainly containing alanine, lysine and proline was reported in almost all the Hls and was suggested to originate from internal sequence duplications (24). The central globular region has been shown to be the most conserved region among the Hi histones (23,24). In the protein described here it extends from amino acid 46 to 122 and mainly contains hydrophobic residues (Fig. iB). Partial amino acid sequences of the corresponding region have been previously established for one maize (I1) and two wheat (10) H1 proteins. The maize Hi described here shows very extensive homology with these partial sequences but only a 60% homology with the pea and Arabidopsis HI globular regions (Fig. iB) [(9); Gantt, personal communication]. Interestingly, the comparison of these sequences with the consensus sequence resulting from the compilation of the globular regions of about 30 animal HI proteins (23) revealed the existence of 15 conserved residues. The tertiary structure of the central globular region of the chicken H5 histone has been established and extended to different animal HI proteins (25,26). When superimposing this model to the central region of the maize Hi, the conserved residues between animal and plant HI mainly fall into the A, B and C helices and between helices A and B (Fig. IB). From the folded structure of the central region of H5, 20 invariant and conservative surface residues had been found among 24 different animal HI and H5 histones (26). Among these residues, 9 are also conserved in the 3 HI from plants, all being localized from residue 9 to 48 of the maize globular region except serine 68. Four of them are basic amino acids and could possibly interact with the DNA. These conserved surface residues are probably responsible for establishing the basic interactions between the HI globular region and the internucleosomal DNA.

Genomic organization of the Hi genes in the genomes of maize and related plant species DNA from maize (inbred line F7), teosinte, Corx, millet, sorgho, barley, oat, wheat and from one dicot, Arabidopsis were blothybridized under stringent conditions to cHIC21. The DNA from all species except Arabidopsis showed hybridization with more or less intensity (Fig. 2). In maize, 3 bands hybridized whether the DNA was digested with EcoRI or BamH. The difference of hybridization intensities between the 3 bands could be due to the presence of different copy numbers or to the fact that weaker bands correspond to genes encoding other closely related HI

1494 Nucleic Acids Research, Vol. 19, No. 7

subtypes. When hybridization was carried out under less stringent conditions (37°C-40% fonnamide) 2 additional faint bands were revealed (data not shown) suggesting the existence of other subfamilies of HI genes likely encoding variants of HI proteins. Multiple bands hybridized with variable intensities in teosinte (one of the wild ancestors of cultivated maize) when the DNA was digested with EcoRI. This multiplicity of bands is most probably due to the existence in the HI genes of internal EcoRI sites as the BamHl blot only reveals 3 fragments as in maize. Coix, which belongs to the same tribe as maize, and sorgho which belongs to another tribe within the same subfamily, the Andropogonoideae, also show DNA fragments hybridizing with high intensity. It is more surprising that strong homology was found in the genome of millet (Pennisetum) which belongs to another subfamily, the Panicoideae and which did not show to be strongly related to maize at the level of the H3 and H4 multigene families (14). Faint hybridization was also found with the DNA of barley, oat and wheat which belong to a third even more distant subfamily, the Poordeae. If significant, this faint hybridization possibly relates to the conserved amino acid sequences within the globular regions of HI histones from different gramineae. There is for example a 80% homology between the HI histone described here and the partial sequences of two wheat HI variants (see Fig. iB). As the codon usage has been shown to be very similar in the histone genes of other gramineae (16) such a homology at the amino acid level is probably sufficient to produce a hybridization signal at the DNA level. No hybridization was found with the DNA of Arabidopsis nor was any signal detected when the pea Hi cDNA used as a probe was hybridized to a maize DNA blot (data not shown). This result is not surprising as between the HI proteins of maize and these 2 dicots the homology with the most conserved central globular regions is only 60% (Fig. IB) and the codon usage is not the same (22). Expression of the Hi genes In the animal kingdom the expression of the majority of histone genes has been shown to be coupled to the synthesis of DNA, although a restricted number of genes escapes this general rule (27,28) especially some variants of HI (29,30). Such a coupling was never clearly demonstrated in plants although the amounts of mRNAs corresponding to the core-histones H3 and H4 were found to be much higher in meristematic than in non-proliferating tissues (our unpublished results). Whether the expression of the HI genes corresponding to cHIC21 and closely related subtypes parallels that of the core-histone genes or shows any tissue or developmental specificity was estimated by studying the steady state levels of HI mRNAs in maize germinating embryos and in different tissues of maize adult plants. Expression during germination. Total RNA was extracted from manually excised maize embryos every 12 hours from day 0 to 5 of germination, and the resulting 'Northern-blot' was hybridized to the HI cDNA used as a probe. The autoradiogram shows that there is no stored HI mRNA in the dry embryo (Fig. 3). The steady-state level of the HI mRNA increases from day 1 to 4 and decreases after the 4th day. A densitometric evaluation showed that the amounts of HI mRNAs roughly change in parallel with the DNA synthesis as estimated by 3H-TTP incorporation into the DNA (data not shown). There is no major difference between the evolution of the amounts of HI mRNAs and those of the core-histone H3 mRNA for example (see on

IN BS

HI

R R~1

rn

rl

Wl-*i-;

cm

.

H3 Figure 4. Expression of HlC21 in different tissues of adult maize. 25 isg of total RNA extracted from different tissues of adult maize were analyzed on a formaldehyde-agarose gel, and transferred onto a Hybond N nylon membrane. The northern-blot was successively hybridized to radioactively labelled cHlC21 and the maize histone H3 coding region. BS: leaf-sheath, L: adult leaves, R: roots (except the tips), RT: root tips, N: nodes, IN: internodes (1: lower, m: middle, h: higher), o: male fully expanded inflorescence, 9: female embryonic inflorescence.

Fig. 3) except that HI mRNAs are hardly detectable after 12 hours of germination whereas H3 mRNAs are clearly present at that stage.

Expression in different tissues. Tissues from different organs of adult maize plants have been collected and the total RNAs were extracted. The tissues were chosen with respect to the presence of meristematic tissues (root-tips, embryonic inflorescence, nodes) or the absence of cell divisions (adult leaves, leaf-sheath, internodes) and the resulting 'Northern-blot' was hybridized to cHlC21 used as probe (Fig. 4). As expected, H1 transcripts are most abundant in the embryonic female inflorescence, root-tips and nodes, all these tissues being known to contain meristematic activity. However, more surprising is the presence of significant amounts of transcripts in tissues where the cells are elongating but no longer dividing like the internodes and the elongating part of the roots. This result might be due to the presence of secondary or residual meristems, to endopolyploidisation which was shown to occur in several plant tissues (31) or simply to the existence of an uncoupled constitutive histone gene expression in some plant tissues. Finally, no or very low amounts of HI mRNAs were detected in fully expanded organs such as leaf-sheath, leaves, and lower internodes. These results compared to the expression pattern of the H3 core histone genes (Fig. 4) indicate that the gene(s) encoding the Hi subtype(s) described here does not show any strict tissuespecificity although some quantitative differences were found between the HI and H3 mRNAs steady-state levels in some particular tissues such as root-tips, lower and middle nodes (see Fig. 4). More accurate quantitative estimations possibly using more subtype-specific probes will be necessary to determine whether these differences are of any functional significance.

DISCUSSION Our knowledge of the structure of histones as well as that of the organization and expression of the corresponding genes in higher plants is still limited when compared to the animal kingdom (20,21, 32). Genes encoding histones H3 or H4 have been studied in a series of plant species including monocots such as wheat (33,34), maize (35,36) and rice (37), and dicots such as alfalfa (38) and Arabidopsis (22). Their structures as well as their general organization particularly in the genome of maize is now well documented (14). However, informations concerning the two other core-histones namely H2A and H2B are very sparse as only one complete sequence of a wheat H2A (39) and partial sequences

Nucleic Acids Research, Vol. 19, No. 7 1495 of H2B from pea (40) and wheat (41) have been published up to now. As far as the histones HI are concerned only two complete sequences had been deduced from a pea cDNA (9) and an Arabidopsis genomic clone (Gantt, personal communication). In monocots only very partial sequences had been established of wheat (10) and maize (I1) HI proteins. Thus, the sequence presented here is the first complete primary structure of a HI histone from a monocot. The encoded protein is of larger size (245 amino acids) than the majority of its counterparts in the animal kingdom and exhibits the typical organization into three domains found in this class of histones (42,43). The 45 amino acid-long N-terminal region contains more acidic residues than any other animal HI protein and an equal number of basic residues. The presence of a high number of acidic amino acids is even more pronounced in the two HIs sequenced from dicots. Generally the acidic and basic domains are clearly delimited, the acidic domain being at the N-terminal part and clusters of basic residues immediately in front of the globular region.This organization also exists in the 3 HI proteins from plants but in the case of the dicots the acidic and basic domains are separated by an additional region where positively and negatively charged amino acids are intermingled. More sequences of HI proteins from plants are needed to determine if such a structure of the N-terminal region is specific to dicots. Otherwise, the N-terminal region is generally considered as being involved in the proper positioning of the globular region of the protein at the DNA entry and exit site of the nucleosome (5), and thus, it certainly would be of interest to estimate if the presence of such a long and highly negatively and positively charged Nterminal region might influence the structure of the internucleosomal domain in plants. The central globular region of 76 residues contains 42 residues which are conserved in the 3 plant HI proteins, and 15 of them are also found in the animal HI consensus sequence. Among these 15 residues 9 were shown to be located at the surface of the molecule and are highly conserved in all the HI/H5 animal histones (26). Otherwise, the animal HI globular regions have 8 conserved lysine residues shown to be located on the surface and which are supposed to play a fundamental structural role by sealing off the two turns of the DNA around the nucleosome (25). Among these 8 basic residues, 4 are conserved in plants (positions 19, 21, 31 and 48 on figure iB) and the 4 others could possibly be replaced by lysine or arginine residues located near up- or downstream. The size of the C-terminal region (124 amino acids) is slightly higher than in the majority of animal HI proteins and contains a similar number of basic residues. Like in several other HI proteins, it contains a number of repeated motifs consisting of Lys, Pro and Ala whose origin and role remain to be determined. In the genome of maize, the HI protein described here is encoded by a small multigenic family showing a much simpler hybridization pattern than the genes encoding H3 and H4 (14). Highly homologous gene families likely encoding the same HI variant exist not only in species belonging to different tribes within the same taxonomic subfamily, but also in the genome of millet which belongs to another subfamily. This result shows that this multigenic family has quite an ancient origin and has been conserved through the evolution and divergency of tribes and species. The existence of additional faint hybridization bands indicate that using the maize cHIC21 clone as a probe can reveal the existence of HI histone gene families likely encoding other HI variants in the genome of maize or closely related species

as well as in morphologically more distant species such as wheat, oat and barley. In this case the sequence homology likely restricts to the globular domain which thus could be used as a probe to clone other HI genes from a wide variety of gramineae. Estimation of the expression of the gene(s) corresponding to cHIC21 failed to show any strict tissue specificity, although some quantitative differences could be detected as compared to the expression of the genes encoding the core histones. It certainly would be interesting, using the globular region of cHIC21 as a probe to clone the genes encoding the other HI variants existing in maize in order to see whether tissue specific or developmentally regulated genes encoding HI variants similar to HI1 and H5 (29,30) found in animals also exist in plants.

ACKNOWLEDGEMENTS We thank Dr. S.Gantt for having communicated unpublished informations concerning the gene encoding the Arabidopsis H1

protein.

REFERENCES 1. Thoma, F. (1988) In Kahl, G. (ed.), Architecture of Eukaryotic Genes. VCH Verlagsgesellschaft, Weinheim, pp. 163- 186 2. Zlatanova, J. (1990) Trends Biochem. Sci. 15, 273-275 3. Lewin, B. (1990) Cell 61, 743-752 4. Cole, R.D. (1984) Anal. Biochem. 136, 24-30 5. Allan, J., Mitchell, T., Harborne, N., Bohm, L. and Crane-Robinson, C. (1986) J. Mol. Biol. 187, 591-601 6. Thoma, F., Losa, R. and Koller, T. (1983) J. Mol. Biol. 167, 619-640 7. Spiker, S. (1988) In Kahl, G. (ed.), Architecture of Eukaryotic Genes. VCH Verlagsgesellschaft, Weinheim, pp. 143-162 8. Mazzolini, L., Vaeck, M. and Van Montagu, M. (1989) Eur. J. Biochem. 178, 779-787 9. Gantt, J.S. and Key, J.L. (1987) Eur. J. Biochem. 166, 119-125 10. Brandt, W.F. and Von Holt, C. (1986) FEBS Lett. 194, 282-286 11. Hurley, C.K. and Stout, J.T. (1980) Biochemistry 19, 410-416 12. Langenbuch, J., Philipps, G. and Gigot, C. (1983) Plant Mol. Biol. 2, 207-220 13. Sanger, F., Nicklen, S. and Coulson, A.R. (1977) Proc. Natl. Acad. Sci. U.S.A. 74, 5463-5467 14. Chaubet, N., Philipps, G. and Gigot, C. (1989) Mol. Gen. Genet. 219, 404-412 15. Southern, E.M. (1975) J. Mol. Biol. 98, 503-5 17 16. Gigot, C., Chaubet, N., Chaboute, M.-E., Ehling, M. and Philipps, G. (1987) Plant Physiol. Biochem. 25, 235 -247 17. Lehrach, H., Diamond, D., Wozney, J.M. and Boedtker, H. (1977) Biochemistry 16, 4743-4751 18. Chaubet, N., Chaboute, M.-E., C1ement, B., Ehling, M., Philipps, G. and Gigot, C. (1988) Nucleic Acids Res. 16, 1295 - lf4 19. Chaboute, M.-E., Chaubet, N., C1ement, B., Gigot, C. and Philipps, G. (1988) Gene 71, 217-223 20. Hentschel, C.C. and Birnstiel, M.L. (1981) Cell 25, 301-313 21. Maxson, R., Cohn, R. and Kedes, L. (1983) Annu. Rev. Genet. 17, 239-277 22. Chaboute, M.-E., Chaubet, N., Philipps, G., Ehling, M. and Gigot, C. (1987) Plant Mol. Biol. 8, 179-191 23. Wells, D. and Mc Bride, C. (1989) Nucleic Acids Res. 17, r31 1-r338 24. Doenecke, D. (1988) ) In Kahl, G. (ed.), Architecture of Eukaryotic Genes. VCH Verlagsgesellschaft, Weinheim, pp. 123-142 25. Clore, G.M., Gronenborn, A.M., Nilges, M., Sukumaran, D.K. and Zarbock, J. (1987) EMBO J. 6, 1833-1842 26. Crane-Robinson, C. and Ptitsyn, O.B. (1989) Protein Eng. 2, 577-582 27. Schumperli, D. (1986) Cell 45, 471-472 28. Busslinger, M., Schiimperli, D. and Birnstiel, M.L. (1985) Cold Spring Harbor Symp. Quant. Biol. 50, 665-670 29. Lennox, R.W. and Cohen, L.H. (1984) In Stein, G.S., Stein, J.L. and Marzluff, W.F. (eds.), Histone Genes. Structure, Organization and Regulation. Wiley-Interscience Publication, New York, pp. 373-395 30. Harvey, R.P. and Wells, J.R.E. (1984) In Stein, G.S., Stein, J.L. and Marzluff, W.F. (eds.), Histone Genes. Structure, Organization and Regulation. Wiley-Interscience Publication, New York, pp. 245-262

1496 Nucleic Acids Research, Vol. 19, No. 7 31. D'Amato, F. (1964) Caryologia 17, 41-52 32. Old, R.W. and Woodland, H.R. (1984) Cell 38, 624-626 33. Tabata, T., Sasaki, K. and Iwabuchi, M. (1983) Nucleic Acids Res. 11,

5865-5875 34. Tabata, T., Fukasawa, M. and Iwabuchi, M. (1984) Mol. Gen. Genet. 196, 397-400 35. Philipps, G., Chaubet, N., Chaboute, M.-E., Ehling, M. and Gigot, C. (1986) Gene 42, 225-229 36. Chaubet, N., Philipps, G., Chaboute, M.-E., Ehling, M. and Gigot, C. (1986) Plant Mol. Biol. 6, 253-263 37. Peng, Z.-G. and Wu, R. (1986) Gene 45, 247-252 38. Wu, S.C., Bogre, KL., Vincze, E., Kiss, G.B. and Dudits, D. (1988) Plant Mol. Biol. 11, 641-649 39. Rodrigues, J.D.A., Brandt, W.F. and Von Holt, C. (1985) Eur. J. Biochem. 150, 499-506 40. Hayashi, H., Iwai, K., Johnson, J.D. and Bonner, J. (1977) J. Biochem. 82, 503-510 41. Von Holt, C., Strickland, W.N., Brandt, W.F. and Strickland, M.S. (1979) FEBS Leu. 100, 201-218 42. Cary, P.D., Hines, M.L., Bradbury, E.M., Srmith, B.J. and Johns, E.W. (1981) Eur. J. Biochem. 120, 371-377 43. Allan, J., Smith, B.J., Dunn, B. and Bustin, M. (1982) J. Biol. Chem. 257, 10533-10535