histone acetylation in zea mays. ii. biological significance of post-translational histone...

THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1991 by The American Society for Biochemistry and Molecular Biology, Inc

Vol. 266, No. 28, Issue of October 5, pp. 18751-18760,1991 Printed in U.S.A.

Histone Acetylation in Zea mays I1 BIOLOGICAL SIGNIFICANCE OF POST-TRANSLATIONAL HISTONE ACETYLATION DURING EMBRYO GERMINATION*

(Received for publication, March 12, 1991)

Elena Ivanova GeorgievaSQ, Gerard0 Lopez-Rodas$Y, Ramon Sendrall , Peter Grobner**, and Peter Loidl$ $$ From the $Department of Microbiology and **Medical Chemistry and Biochemistry, University of Innsbruck, Medical School, A-6020 Innsbruck, Austria and the IIDepartment of Biochemistry and Molecular Biology, Faculty of Sciences, University of Valencia, E-46100 Burjassot, Valencia, Spain

Multiple forms of histone acetyltransferases and histone deacetylases, which have been separated and characterized in the accompanying manuscript (Lopez- Rodas, G., Georgieva, E. I., Sendra, R., and Loidl, P. (1991) J. Biol. Chem. 266, 18745-18750), together with in vivo acetate incorporation, were studied during the germination of Zea mays embryos. Total histone acetyltransferase activity increases during germination with two maxima at 40 and 72 h after start of germination. This fluctuation is mainly due to the cytoplasmic B-enzyme which predominantly acetylates histone H4 up to the diacetylated form. The nuclear histone acetyltransferase A2, specific for H3, is low throughout germination, except at 24 h, when it tran- siently becomes the main activity. Both enzymes are also present in the dry embryo, whereas the second nuclear enzyme A l , specific for H3 and H4, is absent in the initial stage of differentiation. The two histone deacetylases, HD1 and HD2, exhibit entirely different patterns. Whereas HD1 activity is low in the dry embryo and increases during germination, HD2 is the predominant enzyme at the start of differentiation, but almost disappears at later stages. Analysis of the in vivo acetate incorporation reveals that H4 is present in up to tetraacetylated subspecies. The pattern of acetate incorporation into core histones closely resem- bles the fluctuations of histone acetyltransferase B. Based on the analysis of thymidine kinase activity a close correlation was established between histone acetyltransferase B and DNA replication, whereas the A2 enzyme is associated with transcriptional activity. His- tone deacetylase HD1 obviously serves a specific function in the dry embryo and could be a prerequisite for DNA repair processes. The study confirms the idea of multiple functions of histone acetylation and assigns

* This work was supported in part by Grant P7989-MED from the Fonds zur Forderung der wissenschaftlichen Forschung. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduer- tisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

5 Recipient of a short term fellowship from a joint research project of the Austrian and Bulgarian Academies of Sciences. Permanent address: Institute of Genetics, Molecular Genetics Laboratory, Bul- garian Academy of Science, 1113 Sofia, Bulgaria.

ll Recipient of a postdoctoral fellowship from Programa Sectorial de Formacion de Profesorado y Personal Investigador del Ministerio de Educacibn y Ciencia (Spain).

$$TO whom correspondence should be addressed Institut fur Mikrobiologie (Med. Fak.), Universitat Innsbruck, Fritz-Preglstr. 3, A-6020 Innsbruck, Austria.

distinct enzymes, involved in this modification, to certain nuclear processes.

Post-translational acetylation of core histones is an active metabolic process whose precise biological functions remain controversial. Studies of this type of modification in the ciliate protozoan Tetrahymena indicated that the biological function may not be a unique one, but rather variable; in macronuclei of this organism core histone acetylation is involved in the regulation of transcriptional activity, whereas in micronuclei the modification plays a role in the deposition of histones onto chromatin during DNA replication (Chicoine et al., 1986; Lin et al., 1989).

Multiple functions of core histone acetylation are adequate to explain the different and often controversial results in the literature: in a variety of organisms and cell types histone acetylation has been linked to transcription (Doenecke and Gallwitz, 1982; Reeves, 1984; for reviews), DNA replication (e.g. Loidl et al., 1983; Loidl and Grobner, 1987; Weiss and Puschendorf, 1988; Talasz et al., 1990), DNA repair (e.g. Smerdon et al., 1982; Ramanathan and Smerdon, 1989), histone replacement during differentiation processes (e.g. Chris- tensen and Dixon, 1982; Christensen et al., 1984; Grimes and Henderson, 1984; Loidl and Grobner, 1986), and recombination events (Schaffhausen and Benjamin, 1976). The common denominator of these nuclear events is a transient destabili- zation of nucleosome structure and subtle rearrangement reactions occurring within distinct chromatin domains. Based on the numerous results in the literature and on experimental data of the myxomycete Physarum polycephulum our laboratory has proposed a model (Loidl, 1988) which explains the biological role of histone acetylation as a general mechanism to destabilize chromatin structure in a distinct way during different nuclear processes.

Recently, Megee et al. (1990) have pointed out the essential importance of a correct “physiological” acetylation of H4 in yeast; by introducing directed point mutations it was demonstrated that exchange of single lysine residues in the N- terminal histone domains resulted in mating sterility or cell cycle delay.

The effect of histone acetylation on structural properties of nucleosomes or chromatin is not completely clear at present. Although no dramatic changes of the physicochemical properties of nucleosomal structure are observed following hyper- acetylation of chromatin, subtle changes within the nucleosome core as well as alterations of higher order structures occur; recently it was shown that butyrate-induced hypera- cetylation causes an unfolding of the nucleosome core particle

18751

18752 Histone Acetylation in 2. mays

(Oliva et al., 1990) and as a consequence of core histone h~eracetylation also a change in the capacity of histone HI to induce higher order structures (Ridsdale et a[., 1990).

To gain a better understanding of the biological functions of this post-translational modification it is necessary to study the enzymes involved in the dynamic equilibrium of histone acetylation. In the accompanying paper (L6pez-Rodas et at., 1991) we characterized histone acetyltransferase and histone deacetylase activities from Zea mays. We demonstrated multiple forms of both enzymes, thus underlining the essential biological importance of histone acetylation. This paper pre- sents the first simultaneous analysis of histone acetyltransferases, histone deacetylases, and in vivo acetate incorporation during a differentiation pathway. The model system we used is the germination of Z. m y s embryos. We demonstrate that distinct enzyme activities are linked to distinct processes during embryo germination, confirming the idea that histone acetylation serves multiple functions, being involved in DNA replication, transcription, and probably DNA repair in one and the same organism during a differentiation process.

EXPERIMENTAL PROCEDURES

~a~er~a~- [ l -~4C]Acety l -CoA (60 mCi/mmol), [3H]acetate (4.5 Ci/ mmol), and [~~Clthymidine (55 mCi/mmof) were purchased from Amersham International (Amersham, Buckinghamshire, England).

Plant Material-Maize seeds (2. mays M320) were germinated in the dark at 28 "C on cotton layers soaked with water. Total embryos were harvested at 0,24,40, and 64 h after start of germination. At 72 h germination the meristematic part of the root was used. In the case of 40 and 64 h of germination embryos of the same size were selected.

Determination of Histone Acetyltransferase and Histone Deacetylase Activity-The procedure for obtaining cellular extracts and the enzyme assays for histone acetyltransferase (HAT)' and histone deacetylase (HD) are described in detail in the accompanying manuscript (L6pez-Rodas et al., 1991).

In the case of the determination of the total HAT activity (Fig. 7) pooled aliquots of the total chromatographic eluate were dialyzed against buffer B (10 mM NH,CI, 0.25 mM EDTA, 5 mM 2-mercaptoethanol, 10% (v/v) glycerol, 15 mM Tris-HCl, pH 7.9) containing 200 mM NHICl (to prevent chemical acetylation of proteins) and assayed for enzyme activity. For determination of the total HD activity, aliquots of the pooled peak fractions of HD1 and HD2 were dialyzed against buffer B before the enzyme assay.

The analysis of substrate specificity of total HAT under so called in vivo conditions was performed after dialysis of pooled fractions of each of the time points of germination against buffer M (15 mM MOPS, 0.25 mM EDTA, 5 mM 2-mercaptoethanol, 100 mM NH4Cl, 10% glycerol, pH 7.2). Dialyzed pooled fractions were incubated for 30 min at 25 "C in the HAT assay (see accompanying manuscript, L6pez-Rodas et al., 1991) using Z. mays histones as substrate. His- tones were recovered by novobiocin precipitation and analyzed by SDS-polyacrylamide gel electrophoresis (Laemmli, 1970) or acetic ac id /urea /Tr~~n gel electrophoresis (Bonner et at, 1980) with subsequent fluorography (Laskey and Mills, 1975).

In Vivo Incorporation of fHIAcetate"Tota1 embryos at 24,40, and 64 h as well as meristematic parts of the root a t 72 h of germination were collected (1.5 g each) and sliced. Slices (approximately 1 mm) were incubated in incubation buffer (4 mM magnesium acetate, 5 mM MES, pH 6.1) containing cycloheximide (40 pg/ml) for 30 min at 25 "C. Thereafter 20 mCi of 13H]acetate was added and the tissue was labeled for 90 min at 25 "C with shaking. After labeling, the tissue was washed several times with the same buffer without radioactive acetate (until no radioactivity could be detected in the washing supernatant), plunged into liquid nitrogen, and nuclei were isolated as described in the accompanying manuscript (L6pez-Rodas et al., 1991). Extracted histones were analyzed in SDS-polyacrylamide slab gels (Laemmli, 1970) or acetic acid/urea/Triton polyacrylamide gels (Bonner et ah, 1980) with subsequent fluoro~aphy (Laskey and Mills, 1975).

Thymidine Kinase and Glucose-6-P Dehydrogenase-Extracts from

' The abbreviations used are: HAT, histone acetyltransferase; HD, histone deacetylase; MOPS, 4-mo~holinepropanesulfonic acid; MES, 4-morpholineethanesulfonic acid; SDS, sodium dodecyl sulfate.

different stages of embryo germination were analyzed for thymidine kinase (Grobner, 1979) and glucose-6-P dehydrogenase (Loehr and Walter, 1974) activities using established procedures.

RESULTS

In the accompanying manuscript (Lbpez-Rodas et aZ., 1991) we have characterized multiple histone acetyltransferase and histone deacetylase enzyme forms. After chromatography with DEAE-Sepharose we demonstrated the presence of three HAT a cytoplasmic HAT-B, eluting around 200 mM NH,CI, specific for H4 and probably for H2A; a nuclear HAT-A1, eluting around 100 mM NH4CI, specific for H3 and H4; a second nuclear HAT-AB, specific for H3. HAT-A2 and HAT- B elute close to each other and we therefore term this peak A2fB in the chromatographic separation profiles. With respect to HD, two nuclear enzyme activities could be separated: HD1, eluting around 150 mM NH4Cl, and HD2, eluting with approximately 300 mM NH4C1.

A ~ $ y s i s of HAT and HD d ~ r ~ ~ Embryo G e r m ~ ~ t ~ n of Z. mays-Crude cellular extracts of Z. mays seeds were fraction- ated by DEAE-Sepharose chromatography and analyzed for HAT and HD activities and substrate specificity (described in detail in the accompanying manuscript, L6pez-Rodas et a i , 1991) at different times of germination (0, 24, 40, 60, and 72 h; Figs. 1-5). The quantitative evaluation of the results over the whole germination process is summarized and presented in Figs. 6 and 7.

In the dry embryo (0 h of germination) we found a prominent peak of HAT-A2/B (Fig. lA). Analysis of this peak by SDS-polyacrylamide gel electrophoresis with subsequent fluorography revealed that this peak contains HAT-AZ, leading to acetylation of histone H3, and HAT-B, acetylating H4 (Fig. 1D). Integration of specific labeling density values of individual histones among the whole elution volume of the peak provided a quantitative estimate of the contribution of the different HAT enzymes to the acetylation of individual histone species. Fig. 1, C and D, shows that acetylation of H4 by the cytoplasmic enzyme HAT-B was predominant in the dry embryo. The acetylation of H3 by the nuclear enzyme HAT- A2 was relatively low (Fig. 1, B and f)). Interestingly, H3 acetylation also occurred at high ionic strength of the elution gradient (104-110 ml). This particular H3 acetylation was only observed in the dry embryo and cannot be assigned to one of the characterized HAT forms.

The inco~oration of radioactivity into trichloroacetic acid- precipitable material in the HAT assay from 18 to 75 ml is due to the acetylation of nonhistone proteins which elute in vast amounts in this part of the gradient (Fig. lA, upperpart). Dry embryo preparations contain huge amounts of storage proteins, originating from contaminations with endosperm during excision of the embryo from the dry seed. These storage proteins are catabolized during the germination process. Note, that in the dry embryo elution profile we have extremely high protein concentrations, as judged from Asso readings. These values decreased during the differentiation process (Figs. 1- 5A). Acetylation of such proteins is reflected in the HAT activity assay, since proteins are recovered by trichloroacetic acid p~cipitation; in the fluorograms these acetylated proteins do not appear, since we used novobiocin to selectively recover histones from our in vitro system. Both histone deacetylases, HD1 and HD2, were present in the dry embryo.

At 24 h of embryo germination (Fig. 2) a pronounced decrease of the overall protein concentration was observed, as already mentioned above. The most striking change was the more than 10-fold increase of H3 acetylation by HAT-A2. It should be noted that only at 24 h of germination H3 acetylation by HAT-A2 reaches the same high level like H4 acety-

Histone Acetylation in Z. mays 1875.3

A A A 20 - I5 -

10

5

Q

LO

I .5

0 9

1.0 <

n.5

n

20 ..

15-

10 -

5 -

Q -

HAT n 2

0 2 0 40 nfl 1 0 0 0 Volume

I

- H3 ._ ._ H2A :I- 400

n Volume

n 20 40 60 KO 1o(1 120

Vnlume o 20 40 M nn IIKI 1211

Volume I )

"

FIGS. 1-5. Activity and substrate specificity of histone acetyltransferases and activity of histone descetvla.ses during germination of Z. mays embryos. Extracts of size-selected emhrvos at 0 (Fig. 1, column 1 ) . 24 (Pig. 2, column 2) . (Fig. :{, p. 1875,i column I ) , 64 (Fig. 4, p. 18754 column 2). and meristematic parts of the roots at 72 h (Fig. 5 , p. 1875.5 column I ) were applied to a 1)EAE- Sepharose column and eluted with 90 ml of a linear NH,Cl gradient (10-350 mM). A , activities of histone acetyltransferases ( u p p r p o r t ) and activities of histone deacetvlases (lorrwr port of A ) were determined over the whole chromatographic eluate (see "Experimental I'rocedures" of the accompanying manuscript) usingchicken ervthrocvte histones as suhstrate. Protein was estimated hy monitoring the A.,*, in the eluate


I HDI

HDZ

0 20 40 60 80 1 0 0

Volume

300 B - H3

H2A "-

0 20 40 60 80 100 120

Volume ,l

- t!.l 600 -

400 -

200 -

0 20 40 60 80 100 120

Volume

1 .o n

D

0

n t x E e

<

1 HDI

H 3 - H2B- H 2 A- H 4 -

H 3 - HZB- HZA- H 4 -

A

I .o 0 0 n <

0.5

0

0 20 40 60 80 100

Volume I

120

0 20 40 60 80 100 120

Volume

800 - C

-H4

600-

400-

200 -

0 20 40 60 80 100 120

Volume

38 41 46 SI ss sx 61 64 73 79 x3 xx 97 Volume (ml)

(dashed l i n e without symbols). AI , A2 /B , HDI , and HD2 mark the positions of enzymes in the chromatographic elution. After incubation of chromatographic aliquots in the presence of ['4C]acetyl-CoA with chicken erythrocyte histones, the novobiocin precipitate was analyzed by SDS-polyacrylamide gel electrophoresis with subsequent fluorography ( D ) . The specific labeling density ( S L D ) was calculated as the ratio between the incorporation of radioactivity in a histone band in the fluorogram and the amount of protein present in this band in the corresponding Coomassie Blue-stained gel; this evaluation was performed using an image analyzing system. B, SLD for H3 and H2A; C, SLD for H4. Note specific labeling density scale for H4 (C) is different from the scale for H3 and H2A ( B ) .

Histone Acetylation in 2. mays 18755

A

1 HOI

0 20 40 60 80 100 120

Volume

300 7 B

- H3 - H2A

200 -

1 0 0 -

0 20 40 60 80 100 120

Volume

2 0 0 0 - C

lm -

800 200 i : - - A 1200 -

0 7 ' ~ ' Z . . . S . # . 0 20 40 M) 80 100 120

Volume

I )

1.5

0 an

1.0 4"

0.5

0

0 20 40 60 80 B

1

H 3 - HZB- H 2A' H 4 - --I - .

- 1 41 44 48 51 54 61 65 6X 71 75 7X X 2 X I

Volume (ml)

FIGS. 1-5"continued

lation by HAT-B during other stages of germination (Fig. 2, B and D). Concomitantly, a second small maximum of H3 acetylation by HAT-A1 appears, with a faint label also in H4. In comparison to the dry embryo there is no pronounced

0 20 40 60 80

Time of germination (h)

FIG. 6. Total activities of HAT-A1, HAT-AB, and HAT-B expressed as total specific labeling density (SLD) during embryo germination. Specific labeling density and values from Figs. 1-5, B and C, over the whole chromatographic elution volume of the three histone acetyltransferase peaks were integrated and plotted as total specific labeling density against the time course of germination. A, HAT-A1 (label in H3 and H4); B, HAT-A2 (label in H3); C , HAT- B (label in H4 and H2A).

change in H4 acetylation by HAT-B. It is noteworthy, that this time point is the only one when we could detect a faint label in H2B.

The deacetylase activities also changed during the transition from the dry embryo to 24 h of germination. HD1 increased, as judged from the activity over the whole elution volume of its peak, whereas HD2 decreased (Fig. 2A, lower purt) .

At 40 h of embryo germination we observed a dramatic decrease of H3 acetylation by HAT-A2 (Fig. 3, B and D). In contrast, H4 acetylation by HAT-B exhibited a pronounced increase (Fig. 3, C and D). The acetylation of H3 and H4 by HAT-A1 did not change significantly in comparison to 24 h of germination (Fig. 3, B-D). The total amount of HD1 increased, whereas HD2 remained unchanged (Fig. 3, A, lower purt) .

At 64 h of germination H3 and especially H4 acetylation


0 m 40 60 80

Time of germination (b)

FIG. 7. Total HAT, HD1, and HD2 activities during embryo germination. Pooled aliquots of the whole chromatographic eluates a t different time points during germination were dialyzed, adjusted to 200 mM NH,CI (to prevent chemical acetylation) and analyzed for HAT activity. The peak eluates of HD1 and HD2 were pooled, dialyzed, and analyzed for enzyme activity.

by HAT-A1 increased, whereas the acetylation of H3 by HAT- A2 decreased (Fig. 4, B-D). Acetylation of H4 by HAT-B also slightly decreased (Fig. 4C). The total amount of HD1 decreased, whereas HD2 remained on an almost undetectable level (Fig. 4A, lower part).

It should be pointed out that at 40 and 60 h of germination we did not harvest the embryos randomly, but selected embryos of the same size for the experiments, thereby eliminat- ing problems with slightly different stages of the germination process.

Embryo germination at 72 h was characterized by a dramatic increase of H4 acetylation by HAT-B; in comparison to 64 h the level was more than 3-fold (Fig. 5, A, C, and D). HAT-A1 activity slightly increased, as reflected by a small increase in acetylation of H3 and H4 (Fig. 5, B-D). A slight decrease was observed for H3 acetylation by HAT-A2 (Fig. 5B).

Corresponding to the position of HAT-B acetylation of H4 we observed a slight incorporation of label into H2A (Fig. 5, B and D ) . It should be mentioned that we detected a very faint label over H2A during the whole germination process, but a significant increase could only be detected at 64 h and in particular at 72 h. The H2A acetylation only contributes a minor part to the overall acetylation by HAT-B.

With respect to deacetylase activities we found a slight decrease of HD1, but a small increase of HD2 (Fig. 5A, lower part).

Fluctuations of HAT-catalyzed Chicken Erythrocyte Acety- lation during Embryo Germination-To elucidate the fluctuations of acetylation of different histone species we integrated the areas of individual histones in Figs. 1-5, B and C, by the aid of an image analyzer system (Kontron IBAS 2000). Fig. 6, A-C, show the fluctuations during the whole germination period with respect to acetylation by HAT-A1, HAT-A2, and HAT-B. It is interesting that HAT-A1 was absent in the dry embryo, whereas HAT-A2 and HAT-B could be detected. The most prominent activity throughout germination was the cytoplasmic HAT-B, which acetylates H4 and to a negligible extent H2A. Values for HAT-B (Fig. 6C) also contain this H2A acetylation; however, this does not significantly contrib- ute to the observed pattern. Even in the dry embryo the acetylation of H4 by HAT-B was more pronounced than the total acetylation catalyzed by HAT-A1 and -A2 in any other stage of germination (Fig. 6, A-C). The only exception was the dramatic increase of H3 acetylation by HAT-A2 at 24 h. HAT-B catalyzed H4 acetylation remained unchanged during

the transition from the dry embryo until 24 h after the start of germination. The progression until 40 h was accompanied by a 2-fold increase of activity. At 64 h, HAT-B activity towards H4 decreased. The transition to 72 h again was accompanied by a more than %fold increase. It seems that HAT-B catalyzed H4 acetylation undergoes transient eleva- tions from a more or less constant basic value (Fig. 6C).

The acetylation of H3 by HAT-A2 exhibited a relatively low level during germination, except at 24 h, where a more than 10-fold increase in comparison to the dry embryo could be observed. Only at 24 h HAT-A2 became the predominant acetylating enzyme. During the transition from 24 to 40 h HAT-A2 declined to approximately a quarter of its peak activity and decreased slightly at later stages of germination (Fig. 6B).

As mentioned above, HAT-A1 was completely absent in the dry embryo. The acetylation of H3 and H4 by HAT-A1 increased during the germination; only at 40 h of germination HAT-A1 slightly decreased (Fig. 6A). Note, the scales for the specific labeling density values are different in each of Fig. 6, A, B, and C; in particular, the scale of Fig. 6A is lower by a factor of 10 and 20, respectively, in comparison to Fig. 6, B or C.

Total HAT and HD Activities during Germination-In order to get a quantitative estimate of activities we measured the HAT activity in the total DEAE-Sepharose elution in the presence of 200 mM NH4C1 (to prevent chemical incorporation of label from [14C]acetyl-CoA). The total HAT activity (Fig. 7) basically resembled the pattern of H4 acetylation by the B enzyme, which was expected with respect to the results of Fig. 6, since HAT-B was the predominant enzyme activity. Only at 24 h we observed an elevated value in comparison to HAT- B, since at that time the contribution of the A2 enzyme is predominant.

In order to quantitatively estimate HD activities throughout germination, the eluted peak volumes of HD1 and HD2 were dialyzed for analysis of enzyme activity. Dialysis is necessary, since increasing salt concentrations of the elution gradient inhibit HD1 and HD2 activities to a different extent, as demonstrated in the accompanying manuscript (L6pez- Rodas et al., 1991). The data of Fig. 7 are therefore not directly comparable with the results shown in Figs. 1-5A (lower part), since the ionic strength of the elution volumes of HD1 and HD2 is different. In dry embryos HD2 was the predominant deacetylating activity. It amounted to approximately twice the activity of HD1. During the initial 24 h of germination HD2 dramatically decreased reaching a very low level that remained almost unchanged during the later stages of germination (Fig. 7). This indicated a specific function in the dry embryo and during the very first steps of germination. HD1 was the predominant deacetylase during germination and stayed on an almost constant level, except at 40 h, when a transitory increase was observed (Fig. 7).

In Vivo Acetylation-To compare our in vitro measure- ments with more physiological conditions, we analyzed pooled fractions of the total chromatographic eluate with homologous 2. mays histones (pH 7.2, 100 mM NH4C1, 25 "c). After incubation, samples were electrophoresed and analyzed by fluorography (Fig. 8A). The results represent the simultaneous action of all HAT enzymes together with histone deacetylases under conditions which more resemble in vivo conditions. This gives us an estimate of the steady state level of acetylation at different times of germination. The overall acetylation at each time point increased with the progression from the dry embryo until 40 h. After a transient decrease at 64 h, acetylation again increased at 72 h of germination. The


A

H 2 A - H 3 H 2 R -

H 4 - " I - 0 24 40 64 72 time of germination ( h )

H4 ac2D ac1 b -

FIG. 8. In vitro acetylation of maize histones during embryo germination. A , pooled chromatographic eluates from different time points of germination were dialyzed. adjusted to 100 mM NH'CI, pH 7.2, and incuhated with homologous Z. mqyn histones and ["Clacetyl- CoA at 25 "C for 30 min. After electrophoresis in SDS-polyacrylamide gels histones were analyzed hy fluorography; R, the peak eluates of HAT-H from 72 h of germination were incubated under the same conditions as in A. Samples were analyzed hy acetic acid/urea/Triton gel electrophoresis with subsequent fluorography; H4ac1, mono- acetylated H4; H4ac2, diacetylated H4.

most prominent fluctuation, reflecting this pattern, was exhibited by H4 acetylation with two maxima a t 40 and 72 h in accordance with the fluctuation of the HAT-B enzyme (Fig. 6C). The pattern of H3 acetylation had a maximum a t 24 h, which is in line with the activity pattern of HAT-A2 (Fig. 6 R ) ; it has to be considered that the dramatic increase seen in Fig. 6 R is diminished by the action of deacetylases. H2A is faintly labeled throughout germination (Fig. 8.4). Analysis of maize H4 acetylated under in vivo conditions by pooled HAT- R from 72 h of germination in acetic acid/urea/Trit,on gels with subsequent fluorography revealed that H4 is modified to mono- and diacetylated form (Fig. 8R).

To extend this study of acetylation under in vivo conditions we also analyzed the in viuo incorporation of ["Hlacetate into core histones. This was of course not possible with the dry embryo, so Fig. 9A shows the in vivo incorporation of radioactive acetate from 24 until 72 h after the start of germination. The labeling was performed in the presence of cycloheximide to inhibit co-translational acetylation of histones. The overall pattern of acetate incorporation closely resembled the total fluctuation revealed in Fig. 8, as well as the total HAT activity of Fig. 7. In contrast to the results of our in vitro system, H3 is the predominant acetylated species, whereas H4 acetylation

SDS A U T - * "

- H 3 - H 2 A

c H 2 B H2A - H3lH2R-

H 4 - 3 'lime of 24 40 64 72

germination (h)

FIG. 9. In vivo acetate incorporation into core histones during different time points of embryo germination. Selected em. hryos at 24. 40, and G4 h. as well as meristematic. parts o f the roots from 72 h of germination, were incuhated In r , i l w with [ 'H]acetnte in the presence of cycloheximide for 90 min. Isolatetl histones were analyzed hy electrophoresis with snhsequent fluorography either in SDS-polyacrylamide gels or acetic acid/urea/'I'riton gels fAr ''/'I, as shown for 72 h. Note, that the right fluorogram shows mono- up t n tetraacetylated H4.

is rather low. However, labeling was performed for 90 min in the presence of cycloheximide wit.h an additional 30 min of cycloheximide preincubation. Therefore the inhibition of protein synthesis has to be considered in the interpretation of this result. The pattern of H4 acetylation clearly follows the activity profile of HAT-R. Acetate incorporation into H3 has two maxima a t 40 and 72 h of germination. To discriminate between acetylation of Z. mays H2R and H3 (both comigrate on SDS-polyacrylamide gels) we separated histones on acetic acid/urea/Triton gels with suhsequent fluorography. Acetic acid/urea/Triton gels allow the clear separation of H2R and H3. Fig. 9R clearly shows that almost all incorporation occurred into H3, whereas only minor incorporation was oh- served in H2R. Acetic acid/urea/Triton gels also allow the separation of the acetylated subspecies of H4. The fluorogram of Fig. 9R shows that mono-, di-, tri-, and tetraacetylated subspecies are present. The overall distrihution of acetylated H4 subspecies does not change significantly during the germination process (result not shown). As already seen in Fig. 8, there is a faint in vivo labeling of H2A during the whole germination period (Fig. 9A ).

Course of Macromolecular Synthmis durinc F h b p o Gcr- mination-To locate periods of DNA replication during the germination process we tried to stain isolated nuclei from different stages of germination with propidium iodide in order to analyze them by FACScan. Unfortunately this approach was not successful, since we found it impossible to reproduc- ibly stain all nuclei of a preparation. Since routine DNA- labeling procedures with radioactive thymidine are not fully reliable during a differentiation process, where dramatic changes in endogenous precursor pools must he expected. and because thymidine incorporation directly depends on the enzyme thymidine kinase, we measured this salvage pathway enzyme during germination and took the enzyme activity as a measure for DNA replication. Thymidine kinase activity is strongly increased during DNA replication (Griihner and Sachsenmaier, 1976; Griihner and Loidl, 1983). Fig. 10 shows the course of this enzyme activity during germination. The enzyme is low in the dry embryo, increasing slightly within the initial 24 h, hut increased approximately 3-fold at 40 h of germination. During progression until 64 h the activity remained unchanged, but it increased dramatically from 64 until 72 h. According to this pattern, there are two cycles of DNA synthesis around 40 and 72 h of germination.


To exclude the possibility that other enzymes, which are not involved in DNA replication, follow a similar pattern, we measured a housekeeping enzyme of the carbohydrate metabolism as a control. Fig. 10 also shows the pattern of activity of glucose-6-phosphate dehydrogenase during germination. The enzyme activity was high in the dry embryo, which is in line with the mobilization of energy from internal carbohydrate stores. It decreased strongly within 24 h and remained on a constant level until 64 h, before it increased again at 72 h of germination, This pattern is similar to the total amount of protein during germination, as judged from Azso readings of Figs. 1-5A. It is obvious that the activity patterns of thymidine kinase and glucose-6-phosphate dehydrogenase are entirely different.

We also measured endogenous RNA polymerase activity in isolated nuclei during 2. mays embryo germination. Due to active RNases it was impossible to quantitate this enzyme activity correctly during germination. Therefore we related our results to published data which show that transcription has a pronounced maximum at 24 h of germination in this specific variety (M320) of 2. mays (Georgieva and Karap- chanska, 1989).

DISCUSSION

In the past decade it has become clear that the reversible acetylation of core histones serves multiple functions in nuclear metabolism. A huge body of evidence has accumulated which implicates histone acetylation in DNA replication (e.g. Loidl et al., 1983; Talasz et al., 1990), transcription (e.g. Allegra et al., 1987; Hebbes et al., 1988), DNA repair (e.g. Ramanathan and Smerdon, 1989), DNA recombination (Schaffhausen and Benjamin, 1976), and differentiation-associated chromatin rearrangements (e.g. Christensen and Dixon, 1982; Loidl and Grobner, 1986). Although these processes are diverse, a common feature is a transient rearrangement of the involved chromatin structure. Based on this assumption models have been proposed for this type of modification which attribute a general function to histone acetylation for chromatin structural transitions (Perry and Chalkley, 1982; Loidl, 1988; Thorne e t al., 1990). In line with this proposal is the observation in the ciliate protozoan Tetrahymena that histone acetylation is linked to histone

0 20 40 60 80

Time of germination (h)

FIG. 10. Activities of thymidine kinase (TK) and glucose-6- phosphate dehydrogenase (GGP-DH) during embryo germination. Extracts of embryos at 0, 24, 40, and 64 h as well as meristematic tissue at 72 h of germination were analyzed for thymidine kinase and glucose-6-phosphate dehydrogenase activities.

deposition and to transcription in one and the same organism (Lin et al., 1989).

It has been demonstrated that the 4 N-terminal lysine residues of histone H4 which are subject to reversible acetylation are not modified randomly, but in an ordered site- specific way (Chicoine et al., 1986; Couppez et al., 1987). This indicates that the effect of acetylation is not merely due to an overall reduction of the positive charge in the N-terminal histone domain, but to a more subtle physicochemical mechanism (Loidl, 1988). The sequential site specificity of acetylation is strict for H3, whereas the occupancy of lysine residues in H4 exhibits greater variability (Thorne et al., 1990). In yeast it has been demonstrated that the N-terminal core histone domains containing the amino acid residues for acetylation are dispensable for vegetative growth, but are essential for the maintainance of distinct chromatin configurations during development (Kayne et al., 1988). To elucidate the differential function of these N-terminal lysine residues, Me- gee et al. (1990) recently performed an elegant genetic study, where they substituted lysine residues by directed point mutations; the results show that substitution at all four positions was lethal, whereas the substitution of single lysine residues resulted in different effects, like mating sterility, delay of cell cycle traverse or prolongation of DNA replication. Thus, these data support the impact of histone acetylation for various cellular functions.

In contrast to the advanced understanding of the sequential acetylation of core histones, almost nothing is known about the association of different histone acetyltransferases and deacetylases with different types of histone acetylation. We assume that the diversity of histone acetylation is also mani- fested on the enzymatic level. It was the aim of this paper to correlate the different enzymes, exhibiting a distinct histone specificity, with nuclear processes occurring during a differentiation process.

Total histone acetyltransferase activity exhibited a pronounced fluctuation during maize embryo germination. This pattern was mainly due to the cytoplasmic HAT-B, which acetylates predominantly H4 and to a low extent H2A. The HAT-B pattern clearly correlates with the activity pattern of thymidine kinase, which we used to detect periods of DNA replication during the differentiation pathway. The peak of DNA replication at 40 and 72 h is in line with a number of investigations in the literature, where it was shown that nuclei in germinating maize embryos have a maximum of replicative DNA synthesis around 40 h after start of inbibition and a second peak in meristematic tissue around 70 h (Deltour and Jacqmard, 1974; Baiza et al., 1989; for review, see Deltour, 1985). We conclude that the function of HAT-B is mainly associated with DNA replication, most likely responsible for the cytoplasmic acetylation prior to the deposition and assem- bly of newly synthesized H4 into chromatin.

The nuclear HAT-A2 which acetylates H3 exhibits a dramatic peak at 24 h after start of germination. This maximum is concomitant with the maximum of RNA synthesis, as measured by uridine incorporation (Georgieva and Karap- chanska, 1990), as well as with an initial increase of rRNA synthesis and processing (Deltour, 1985). A correlation of transcription-linked acetylation of H3 by HAT-A2 could cor- respond to the specific H3 acetylation in transcriptionally active chromatin as proposed by Allegra et al. (1987). HAT- A2 could also represent the histone acetyltransferase activity localized in transcriptionally active chicken erythrocyte chromatin which is enriched in hyperacetylated H3, described by Chan et al. (1988). It may be speculated that at this time of embryo germination the most dramatic structural re-


arrangements necessary for transcriptionally competent genes take place. At 24 h of germination we observed a faint acetylation of H2B. The functional significance of this is unclear, but it could be associated with additional structural changes necessary for the activation of specific genes.

The interpretation of the pattern of the nuclear HAT-A1, which is responsible for H3 and H4 acetylation, is not clear. I t is noteworthy that this enzyme, in contrast to HAT-A2 and HAT-B, is completely absent in the quiescent embryo. The continuous increase during the germination process indicates a yet unclear role associated with growth and active metabolism. It could represent an enzyme activity necessary for the structural remodeling of chromatin during the transition from quiescence to active growth. It should be pointed out that HAT-A1 contributes only a minor part of total HAT activity. The acetylation by HAT-A1 is 10 times less than the modification catalyzed by HAT-A2 and HAT-B. Nevertheless, it may have an important, yet unclear function within distinct chromatin domains.

It should be pointed out that such correlations are only revealed if the different enzyme forms are separated and analyzed in terms of substrate specificity; analysis of total unfractionated activity would conceal such specific correlations, as can be seen for H3 acetylation by HAT-A2.

The basic HAT activity in the quiescent dry embryo is in line with the high levels of H4 acetylation seen in dormant microsclerotia of P. polycephulum (Loidl and Grobner, 1986). During the initial stage of plant embryo germination these activities could be associated with DNA repair which takes place until 12 h after start of germination in 2. mays (Zlatan- ova et al., 1987). According to the results of Smerdon and co- workers, DNA repair can be expected to depend on a certain level of histone acetylation (Smerdon et al., 1982; Smerdon, 1983; Ramanathan and Smerdon, 1989). Since the mainte- nance of a distinct acetylated state in chromatin is determined by the equilibrium between acetyltransferases and deacetylases, one should also expect the action of deacetylase during DNA repair. Interestingly, HD2 is the predominant deacetylating activity in the dry embryo, but it almost completely disappears during progression of germination. This indicates a specific function of HD2 in the dry embryo during the initial phase of germination. The most important nuclear process at this stage is certainly the repair of DNA damage, which accumulated during natural seed dehydration and subsequent inbibition (Cheah and Osborne, 1978).

HD1 is present during the whole germination period with a more or less basic constant activity, except for 40 h where we observe a transient increase. This transient increase could be due to the deacetylation of newly assembled chromatin at this time point, since histone deacetylation is required for the maturation of newly synthesized chromatin in order to form higher order structures (Annunziato and Seale, 1983).

In the dry embryo we observed an additional labeling of H3 in chromatographic fractions eluting with high ionic strength. This activity is certainly different from HAT-AS. Since it is specific for the dry embryo, it could represent an additional differentiation-specific HAT.

Apart from these possible correlations there are apparent discrepancies with respect to the substrate specificity of the different histone acetyltransferases when we compare our i n vitro system with i n vivo acetate incorporation. The most pronounced difference is that i n vivo H3 is the main substrate for acetylation in contrast to H4 in the i n vitro analysis. This is explained by the specific labeling protocol we used. In order to inhibit the co-translational acetylation of core histones during the long incubation period (90 min) with radioactive

acetate, we had to perform the labeling in the presence of cycloheximide after an additional preincubation (30 min without radioactivity). It is clear that 2 h of cycloheximide treat- ment inhibit the de novo synthesis of histones. Since only newly synthesized H4 is acetylated by the cytoplasmic HAT- B, we cannot expect high levels of H4 acetylation. Neverthe- less, the basic pattern of HAT-B is still reflected on a low level by the H4 acetylation in the fluorogram (Fig. 9A). Therefore it is obvious that H3, which is acetylated in chromatin, is the predominantly acetylated histone species under these experimental conditions. The fact that we observe a relatively low i n vivo H3 acetylation at 24 h of germination in comparison to 40 h is not at all surprising, since inhibition of protein synthesis by cycloheximide affects HAT-A2 at its most at 24 h, where this enzyme is predominantly synthesized. Apart from this, the apparent differences in substrate specificity between i n vitro and i n vivo may have additional rea- sons. It can be expected that acetylation within intact chromatin is different from the modification of isolated histones, although the basic substrate specificity of histone acetyltransferases is not entirely different, when we compare chicken erythrocyte histones or homologous maize histones in our enzyme assay (see accompanying paper, L6pez-Rodas et al., 1991). Additional factors are present in chromatin which cannot be substituted in an i n vitro system. Another important point is that i n vivo histones are occupied by other post- translational modifications as well, which may influence the extent of mutual modification; it has been demonstrated that acetylated histones may at the same time be modified by ADP-ribosylation (Malik and Smulson, 1984; Wong and Smulson, 1984) or methylation (Hendzel and Davie, 1991). It should also be considered that the histones undergoing acetylation i n vivo have a distinct predisposed acetylation pattern; considering the nonrandom utilization of acetylation sites, this pre-existing acetylation pattern could strongly influence the further modification. Moreover, differences in the turn- over of histones may quantitatively change the substrate specificity in comparison to an i n vitro system.

As can be seen from our results, i n vivo analysis of histone acetylation may, under certain experimental conditions, give a misleading picture, especially when long labeling periods in the presence of a protein synthesis inhibitor are necessary.

We could demonstrate the existence of H4 in up to tetraacetylated form. On the other hand, the homologous maize H4 was only modified up to diacetylated subspecies in the HAT-B assay. This is in line with data in the literature that newly synthesized H4 is assembled into chromatin preferen- tially in the diacetylated form. We assume that the shift from diacetylated H4 to tri- and tetraacetylated forms is catalyzed by the nuclear HAT-A1, probably at different positions as compared to HAT-B. This could be a specific function for this enzyme limited to certain areas of the chromatin.

For the first time we could show that linkage of a specific histone acetylation to a distinct biological process is regulated on an enzymatic level. The concept that multiple enzyme forms are associated with specific processes is attractive to explain the various functions attributed to this modification. Together with the site specificity it extends the possibilities to introduce a variety of different acetylation states. This correlation represents a new level for the regulation of histone modifications.

Acknowledgments-We wish to thank Dr. G. Stoffler (Innsbruck) and Dr. L. Franco (Valencia) for their interest and encouragement. The critical reading of the manuscript by Dr. I. Pashev (Sofia), Dr. B. Redl (Innsbruck), and Dr. V. Tordera (Valencia) is gratefully acknowledged.


REFERENCES Allegra, P., Sterner, R., Clayton, D. F., and Allfrey, V. G. (1987) J.

Annunziato, A. T., and Seale, R. L. (1983) J. Biol. Chem. 258,12675-

Baiza, A. M., Vasquez-Ramos, J. M., and Sanchez de JimCnez, E.

Bonner, W. M., West, M. H. P., and Stedman, J. D. (1980) Eur. J.

Chan, S., Attisano, L., and Lewis, P. N. (1988) J. Biol. Chem. 263,

Cheah, K., and Osborne, D. J. (1978) Nature 272,593-599 Chicoine, L. G., Schulman, I. G., Richman, R., Cook, R. G., and Allis,

Christensen, M. E., and Dixon, G. H. (1982) Deu. Biol. 93, 404-415 Christensen, M. E., Rattner, J. B., and Dixon, G. H. (1984) Nuckic

Couppez, M., Martin-Ponthieu, A., and Sautibre, P. (1987) J. Biol.

Deltour, R. (1985) J. Cell Sci. 7 5 , 43-83 Deltour, R., and Jacqmard, A. (1974) Ann. Bot. (Lond.) 38, 329-334 Doenecke, D., and Gallwitz, D. (1982) Mol. Cell. Biochem. 4 4 , 113-

Georgieva, E., and Karapchanska, N. (1989) C. R. Acad. Bulg. Sci.

Grimes, S. R., and Henderson, N. (1984) Exp. Cell Res. 152 , 91-97 Grobner, P. (1979) J. Biochem. (Tokyo) 86, 1595-1605 Grobner, P., and Sachsenmaier, W. (1976) FEBS Lett. 71 , 181-184 Grobner, P., and Loidl, P. (1983) Exp. Cell Res. 144 , 385-391 Hebbes, T. R., Thorne, A. W., and Crane-Robinson, C. (1988) EMBO

Hendzel, M. J., and Davie, J. R. (1991) Biochem. J. 273, 753-758 Kayne, P. S., Kim, U. J., Han, M., Mullen, J . R., Yoshizaki, F., and

Laemmli, U. K. (1970) Nature 2 2 7 , 680-685

Mol. Bid. 196 , 379-388

12684

(1989) J. Plant Physiol. 135,416-421

Biochem. 109 , 17-23

15643-15651

C. D. (1986) J. Bid. Chem. 261, 1071-1076

Acids Res. 12,4575-4592

Chem. 262,2854-2860

128

42,95-97

J. 7, 1395-1402

Grunstein, M. (1988) Cell 5 5 , 27-39

Laskey, R. A., and Mills, A. D. (1975) Eur. J. Biochem. 5 6 , 335-341 Lin, R.. Leone, J. W.. Cook, R. G.. and Allis. C. D. (1989) J. Cell Biol.

ios,' 1577-i588 . .

Loehr. G. W.. and Walter. H. D. (1974) in Methoden der Enzvma- tischen Analyse (Bergrneyer, H. U., ed) Vol. I, pp. 673-681, &lag Chemie, Weinheim, Federal Republic of Germany

Loidl, P. (1988) FEBS Lett. 2 2 7 , 91-95 Loidl, P., and Grobner, P. (1986) Nucleic Acids Res. 14, 3745-3762 Loidl, P., and Grobner, P. (1987) Nucleic Acids Res. 15 , 8351-8366 Loidl, P., Loidl, A., Puschendorf, B., and Grobner, P. (1983) Nature

L6pez-Rodas, G., Georgieva, E. I., Sendra, R., and Loidl, P. (1991) J.

Malik, N., and Smulson, M. (1984) Biochemistry 23, 3721-3725 Megee, P. C., Morgan, B. A., Mittman, B. A,, and Smith, M. M.

Oliva, R., Bazett-Jones, D. P., Locklear, L., and Dixon, G. H. (1990)

Perry, M., and Chalkley, R. (1982) J. Bid. Chem. 2 5 7 , 7336-7347 Ramanathan, B., and Smerdon, M. J. (1989) J. Biol. Chem. 2 6 4 ,

Reeves, R. (1984) Biochim. Biophys. Acta 782,343-393 Ridsdale, J. A., Hendzel, M. J., Delcuve, G. P., and Davie, J. R. (1990)

Schaffhausen, B. S., and Benjamin, T. L. (1976) Proc. Natl. Acad.

Smerdon, M. J. (1983) Biochemistry 2 2 , 3516-3525 Smerdon. M. J.. Lan. S. Y.. Calza. R. E.. and Reeves. R. (1982) J.

305,446-448

Biol. Chem. 266,18745-18750

(1990) Science 2 4 7 , 841-845

Nucleic Acids Res. 18 , 2739-2747

11026-11034

J. Biol. Chem. 265, 5150-5156

Sci. U. S. A. 73, 1092-1096

Bid. Chem. 2 5 7 , 13441-13447 '

Talasz, H., Weiss, G., and Puschendorf, B. (1990) FEBS Lett. 2 6 4 , 141-144

Thorne, A. W., Kmiciek, D., Mitchelson, K., Sautiere, P., and Crane-

Weiss, G., and Puschendorf, B. (1988) FEBS Lett. 238 , 205-210 Wong, M., and Smulson, M. (1984) Biockmistry 23 , 3726-3730 Zlatanova, J. S., Ivanov, P. V., Stoilov, L. M., Chimshirova, K. V.,

Robinson, C. (1990) Eur. J. Biochem. 193 , 701-713

and Stanchev, B. S. (1987) Plant Mol. Biol. 10 , 139-144

histone acetylation in zea mays. ii. biological significance of post-translational histone...

Documents