Genetic Evidence for the Importance of Protein Acetylation and Protein Deacetylation in the Halophilic Archaeon Haloferax volcanii

Abstract

Protein acetylation and deacetylation reactions are involved in many regulatory processes in eukaryotes. Recently, it was found that similar processes occur in bacteria and archaea. Sequence analysis of the genome of the haloarchaeon Haloferax volcanii led to the identification of three putative protein acetyltransferases belonging to the Gcn5 family, Pat1, Pat2, and Elp3, and two deacetylases, Sir2 and HdaI. Intriguingly, the gene that encodes HdaI shares an operon with an archaeal histone homolog. We performed gene knockouts to determine whether the genes encoding these putative acetyltransferases and deacetylases are essential. A sir2 deletion mutant was able to grow normally, whereas an hdaI deletion mutant was nonviable. The latter is consistent with the finding that trichostatin A, a specific inhibitor of HdaI, inhibits cell growth in a concentration-dependent manner. We also showed that each of the acetyltransferases by itself is dispensable for growth but that deletion of both pat2 and elp3 could not be achieved. The corresponding genes are therefore “synthetic lethals,” and the protein acetyltransferases probably have a common and essential substrate.

Protein acetylation is one of several classes of posttranslational regulatory processes that occur in living cells. In contrast to those of other protein modifications, such as phosphorylation, which has been studied for over 50 years, the role of protein acetylation in cellular events is less well understood (23, 35). Two types of protein acetylation take place in the cell. N acetylation is the acetylation of the amino termini of newly synthesized proteins that mostly follows the removal of the first methionine residue. This type of acetylation is common in eukaryotes but also occurs to some extent in archaea (1, 14). The second type of acetylation is a reversible modification that occurs on lysine residues of mature proteins, resulting in charge neutralization of these residues. The latter type of protein acetylation is catalyzed by a family of histone acetyltransferases (HATs) that function to transfer an acetyl group from acetyl-coenzyme A (acetyl-CoA) to the ɛ-amino group of certain lysine side chains. The HAT family is composed of five enzyme subfamilies, of which the Gcn5-related N-acetyltransferase (GNAT) subfamily is the best characterized (reviewed in references 11, 17, 26, 29, and 46). Acetylation reactions catalyzed by HATs can be reversed by a family of histone deacetylases (HDACs). HDACs are divided into three classes: class I HDACs are related to yeast Rpd3, class II HDACs are related to yeast HdaI, and class III HDACs are related to the yeast transcriptional repressor Sir2. Class I and class II HDACs share some homology in their catalytic domains and hydrolyze the acetamide bond in similar manners, whereas class III HDACs share no homology with class I and II HDACs and employ a different enzymatic mechanism (reviewed in references 7, 10, 20, 39, and 47).

The reversible posttranslational protein acetylation is not restricted to eukaryotes and has been demonstrated to take place in bacteria and archaea. A metabolic role for the Sir2 homolog CobB was reported in the bacterium Salmonella enterica. It was shown that CobB activates the acetyl-CoA synthetase (ACS) via deacetylation (42), while Pat, an acetyltransferase that exhibits homology in its C-terminal 95-amino-acid-residue region to the eukaryotic Gcn5 acetyltransferases, deactivates it (43). Recent studies of the archaeon Sulfolobus solfataricus revealed that the archaeal homolog of Sir2 forms a stable complex with Alba, one of the most abundant archaeal chromatin proteins. In vitro, Sir2 is responsible for deacetylation of Alba, causing an increase in its affinity to DNA and thereby repressing transcription (5, 57). Subsequent studies showed that Alba acetylation is carried out by a homolog of S. enterica Pat (31). Sir2 homologs are not the only HDACs in archaea and are not even the most common. All archaea (with the exception of Nanoarchaeum equitans) possess homologs of HdaI. However, no genetic analysis has been performed to establish the significance of HDACs and HATs in archaea.

Haloferax volcanii is an obligate halophilic and aerobic archaeon of the euryarchaeota lineage. It has become a model organism for molecular genetic studies of archaea due to the wide range of available genetic tools (2). Also, the complete genome nucleotide sequence of the organism was recently determined and annotated by The Institute for Genomic Research. In this communication, we present bioinformatic evidence for the existence in H. volcanii of genes encoding the reversible protein acetylation/deacetylation reactions and provide genetic evidence for their essentiality.

MATERIALS AND METHODS

Strains and culture conditions.

The properties of the various H. volcanii strains used in this work are given in Table 1.

TABLE 1.

Strains used in this study

Strain	Description	Derivation or reference
H133	DS70 ΔpyrE2 ΔtrpA ΔleuB ΔhdrB	3
WR580	H133 Δsir2	This study
WR643	H133 Δpat1::trpA	This study
WR644	H133 Δpat2::hdrB	This study
WR645	H133 Δpat2::hdrB Δpat1::trpA	This study
WR660	H133 Δelp3::leuB	This study
WR669	H133 sir2::pMM915	This study
WR671	H133 hdaI::pMM1028	This study
WR709	H133 Δpat1::trpA Δelp3::leuB	This study
WR718	H133 ΔhdaI 71-920 (pMM1142)	This study

H. volcanii was routinely grown in rich (HY) medium containing (per liter): 150 g of NaCl, 36.9 g of MgSO₄ · 7H₂O, 5 ml of a 1 M KCl solution, 1.8 ml of a 75-mg/liter MnCl₂ solution, 5 g yeast extract (Difco), and Tris-HCl (pH 7.2) at a final concentration of 50 mM. After autoclaving and cooling of the medium, 5 ml of 10% (wt/vol) CaCl₂ was added. Agar plates contained 18 g of Bacto Agar (Difco) per liter. Casamino Acids (CA) medium contained the same components as the HY medium except that the yeast extract was replaced by 5 g/liter of CA (Difco). Minimal medium contained (per liter) 150 g of NaCl, 36.9 g of MgSO₄ · 7H₂O, 5 ml of a 1 M KCl solution, 50 ml of 1 M NH₄Cl, 45 ml of 10% (vol/vol) glycerol, 5 ml of 10% (wt/vol) sodium succinate, 2 ml of 0.5 M K₂HPO₄, and Tris-HCl (pH 7.2) at a final concentration of 50 mM. After autoclaving and cooling of the medium, the following materials were added: 5 ml of 10% (wt/vol) CaCl₂, 1 ml trace element solution, 0.8 ml of 1-mg/ml thiamine, and 0.1 ml of 1-mg/ml biotin.

For counterselection for uracil auxotrophs, 5-fluoroorotic acid (U.S. Biological) was added to the medium at a final concentration of 100 μg/ml. When needed, novobiocin (Sigma-Aldrich) was added to the medium at a final concentration of 2 μg/ml. When required, thymidine was added to a final concentration of 40 μg/ml, and leucine, tryptophan, and uracil were added to a final concentration of 50 μg/ml. Trichostatin A (TSA) (Sigma-Aldrich) was added to the growth media at the concentrations indicated in the text.

Gene knockouts and gene replacements.

Gene knockouts and gene replacements were performed according to the “pop-in pop-out” methodology as described previously (3, 6). In this methodology, the upstream and downstream flanking regions of the genes to be deleted are PCR amplified and cloned together into the “suicide plasmid” pGB70 or pTA131, which carries the pyrE selectable genetic marker but cannot replicate autonomously in H. volcanii. The plasmids are transformed into an H. volcanii ΔpyrE mutant, and transformants in which the plasmids have been integrated into the chromosome are selected on plates that lack uracil. Upon counterselection on plates containing uracil and 5-fluoroorotic acid, the only cells that survive are those in which the integrated plasmids have been excised by spontaneous intrachromosomal homologous recombination, either restoring the wild-type gene or resulting in its deletion. Gene replacements were performed according to the method of Allers et al. (3).

HY medium was used as a thymidine-minus medium for hdrB cassette selection. CA medium was used as a uracil- and tryptophan-minus medium for trpA cassette selection. Minimal medium was used as a leucine-minus medium for leuB cassette selection.

The pop-out strains were screened using pairs of external “short up” and “short down” primers located approximately 100 bp upstream and 100 bp downstream of the entire flanking construct. All the deletions were also verified by the inability to PCR amplify the coding region of the deleted genes in mutants running in parallel control reactions with the “wild-type” strains.

A list of all integrative plasmids, shuttle vectors, and other vectors that were used for this study is given in Table 2. A list of all primers used in this study is given in Table S1 in the supplemental material.

TABLE 2.

Plasmids used in this study

Plasmid	Relevant properties	Source or reference
pGB70	pUC19 containing the H. volcanii pyrE2 gene	6
pTA105	pUC19 containing the leuB cassette (H. volcanii leuB gene expressed from the ferredoxin promoter)	Thorsten Allers
pTA106	pUC19 containing the trpA cassette (H. volcanii trpA gene expressed from the ferredoxin promoter)	Thorsten Allers
pTA131	pBluescript II containing the H. volcanii pyrE2 gene	3
pTA187	pUC19 containing the hdrB cassette (H. volcanii hdrB gene expressed from the ferredoxin promoter)	Thorsten Allers
pMM915	H. volcanii sir2-flanking regions cloned into pTA131	This study
pMM956	trpA cassette cut from pTA106 and cloned into pMM951 between the pat1-flanking regions	This study
pMM1028	Plasmid designed to allow the introduction of an in-frame deletion of nucleotides 71-920 of H. volcanii hdaI	This study
pMM1033	hdrB cassette cut from pTA187 and cloned into pMM961 between the pat2-flanking regions	This study
pMM1108	leuB cassette cut from pTA105 and cloned into pMM952 between the elp3-flanking regions	This study
pMM1142	H. volcanii hdaI gene under the control of the tryptophanase promoter; cloned between the NdeI and EcoRI sites	28

Transformation procedures.

Transformation of H. volcanii was carried out using the polyethylene glycol method as described previously (6).

Determination of growth rates and TSA inhibition.

To determine growth rates, cells were grown in HY medium at 42°C to the stationary phase, diluted (1:50), grown to an optical density at 600 nm (OD₆₀₀) of 0.6 to 0.7, and then diluted again to an OD₆₀₀ of 0.05 in fresh medium. When the effect of TSA on the growth rate was determined, the fresh medium was supplemented with the indicated amount of TSA. OD₆₀₀ measurements were taken at 2- to 4-h intervals following an overnight lag phase.

Genomic-data analyses.

H. volcanii genome sequence data were obtained from The UCSC Genome Browser (http://archaea.ucsc.edu/cgi-bin/hgGateway?db=haloVolc1.) Multiple-alignment analysis was performed using Multalin software (12; http://bioinfo.genopole-toulouse.prd.fr/multalin/multalin.html). Multiple-alignment figures were created using BoxShade 3.21 software (http://www.ch.embnet.org/software/BOX_form.html).

RESULTS

H. volcanii contains genes predicted to encode putative protein acetylases and deacetylases.

To identify putative components of the H. volcanii posttranslational protein acetylation/deacetylation machinery, its genome database (see Materials and Methods) was screened using sequences of known HATs and HDACs as BLAST queries. Several putative open reading frames that showed significant homology to known HATs and HDACs were identified. The H. volcanii genome was found to contain genes belonging to two HDAC families: one gene is a homolog of the Sir2 family (HVO_2194) (40% identity and 57% similarly to S. solfataricus Sir2), and another is a homolog of the HdaI family (HVO_0522) (37% identity and 55% similarity to Saccharomyces cerevisiae HdaI). Figures 1 and 2 present alignments of the yeast and archaeal Sir2 homologs and the yeast and archaeal HdaI homologs, respectively, and show that the proteins are closely related and conserve important functional domains.

FIG. 1.

Alignment of the Sir2 homologs of different organisms versus the Sir2 homolog of H. volcanii. Several amino acid residues that were proven to be essential for deacetylation activity are marked by gray arrows. Highly conserved amino acids that interact with NAD⁺ are marked with black arrows (4). The CXXC(15-20X)CXXC zinc finger motif is marked by black lines. Black and gray shading represent amino acid identity and similarity, respectively. A. fulgidus, Archaeoglobus fulgidus.

FIG. 2.

HdaI alignment. A partial alignment of the halophilic H. volcanii HdaI and H. marismortui HdaI versus the amino-terminal part of the S. cerevisiae HdaI and the Homo sapiens HDAC1. Known conserved domains are marked by black lines. Amino acid residues that were proven to be essential for deacetylation activity are marked by black arrows. Black and gray shading represent amino acid identity and similarity, respectively.

The hdaI gene encoding HdaI occurs in an operon containing a histone gene (Fig. 3). The putative histone is predicted to have a tandem H3-H4 core domain similar to that present in the Methanopyrus kandleri HMk histone, which specifically binds DNA (13, 33, 41). Interestingly, hdaI overlaps with the CCA-adding tRNA nucleotidyltransferase gene, which is on the opposite DNA strand. The same genome arrangement is found in other halophilic archaea whose genomes have been sequenced but not in any other archaea.

FIG. 3.

Halophilic hdaI operon. The halophilic hdaI genes share an operon with the archaeal “double-histone” homolog. The hdaI start codon overlaps with the histone stop codon. The 3′ region of the HdaI open reading frame overlaps the 3′ region of the essential gene that encodes CCA tRNA nucleotidyltransferase, which is transcribed on the opposite DNA strand.

Three homologs of known HATs were identified in the H. volcanii genome. The first two, Pat1 (HVO_1756) and Pat2 (HVO_1821), belong to the Gcn5 family. They are related to S. solfataricus Pat, and the C-terminal 95-amino-acid region shares similarity with the S. enterica enzyme (26% identity and 45% similarity to S. solfataricus Pat and 34% identity and 45% similarity to the C terminus of S. enterica Pat, respectively) (Fig. 4). The third HAT (HVO_2888) is a homolog of yeast Elp3, which also belongs to the Gcn5 family (Fig. 5) (37% identity and 58% similarity to S. cerevisiae Elp3). Elp3 is a subunit of the “elongator” complex possessing acetyltransferase activity (51, 53).

FIG. 4.

Pat alignment. The S. solfataricus Pat and the carboxy-terminal 95 amino acid residues of S. enterica Pat were aligned against H. volcanii Pat1 and Pat2. Motifs A and B of the Gcn5 family are marked by black lines (32, 46), with the partial sequence of S. cerevisiae Gcn5. The highly conserved R/QXXGXG/A motif, which is important for acetyl-CoA recognition and binding (38, 54), is marked by arrows. Black and gray shading represent amino acid identity and similarity, respectively.

FIG. 5.

Alignment of archaeal and S. cerevisiae Elp3 sequences. Motifs A and B of the Gcn5 family are marked with black lines. The highly conserved R/QXXGXG/A motif is marked by black arrows. Two tyrosine residues that were shown to be highly important for Elp3 acetylation activity are marked by gray arrows. HV, HL, HM, MB, MM, MK, and SC stand for H. volcanii, Halorubrum lacusprofundi, H. marismortui, Methanosarcina barkeri, Methanosarcina mazei, M. kandleri, and S. cerevisiae, respectively. Black and gray shading represent amino acid identity and similarity, respectively.

The H. volcanii sir2 homolog is dispensable, but hdaI is essential.

To determine the essentiality of the two HDACs, we employed the “pop-in pop-out” strategy for constructing gene knockouts previously developed for H. volcanii (3, 6) (see Materials and Methods). In this procedure, if the deletion of the target gene has no effect on the growth properties of the cells, it is expected that in about half of the cells excision of the chromosomally integrated plasmid leaves behind the wild-type allele of the target gene and in about half of the cells the excision creates the desired deletion. The sir2 genomic deletion plasmid pMM915 was transformed into H. volcanii strain H133 and integrated into its chromosome to create the sir2 “pop-in” strain WR669. Following “pop-out” counterselection, it was found that in about half of the cells in which the pyrE-containing plasmid was excised, deletion of sir2 had occurred (to give strain WR580), as determined by PCR analysis (see Fig. S1 in the supplemental material). The H. volcanii sir2 homolog is therefore not essential. The growth rate of the mutant strain in rich HY medium was similar to that of the parental strain, H133, grown under the same conditions. Unlike the S. enterica cobB knockout mutant (45, 48), the H. volcanii sir2 deletion mutant grew normally on low concentrations of acetate as the sole energy source. The H. volcanii Δsir2 mutant also showed no apparent growth impairment compared to the wild-type strain when cultured at salt concentrations ranging from 4 M to 1.5 M NaCl and at temperatures ranging from 37°C to 45°C (data not shown).

The same “pop-in pop-out” strategy was employed in attempts to inactivate hdaI. Since the hdaI gene partially overlaps the essential CCA-adding transferase gene that is involved in tRNA maturation (Fig. 3), attempts were made to delete nucleotides 71 to 920 of the nonoverlapping region of the hdaI gene. Plasmid pMM1028 was used to create the “pop-in” strain WR671. However, no hdaI deletions were obtained following the “pop-out” counterselection, suggesting that hdaI is essential.

To confirm that hdaI is essential, a plasmid carrying the complete hdaI gene was cloned into the pRV1-ptna-bgaH plasmid (28) between the NdeI and EcoRI sites, placing the hdaI gene under the control of the tryptophanase promoter (pMM1142). Plasmid pMM1142 transformed into the “pop-in” strain, WR671. In this genetic background, it was possible to knock out the chromosomal gene and create strain WR718 (see Fig. S2 in the supplemental material). The novobiocin resistance gene present on the pMM1142 plasmid was used to confirm the presence of the plasmid in the cell, together with the genomic knockout background. Other evidence to support the essentiality of hdaI was obtained from the effect of the HdaI-specific inhibitor TSA (16, 50, 55) on the growth rate of H. volcanii. Figure 6 shows that TSA inhibits the growth of H. volcanii in a concentration-dependent manner.

FIG. 6.

Growth curve of the H. volcanii wild-type (wt) strain in the presence of TSA. H. volcanii was grown on rich HY medium without TSA or at different concentrations of TSA, as indicated. The culture turbidity was measured as the OD₆₀₀.

A pat1 and pat2 double-knockout strain is viable.

H. volcanii contains genes coding for two homologs of the S. solfataricus Pat. Plasmid pMM956 was designed to allow replacement of pat1 by the trpA cassette, and pMM1033 was designed to allow replacement of pat2 by the hdrB cassette (see Fig. S3 in the supplemental material). The growth rates in rich HY media of the two single-deletion mutants (WR643 for pat1 replacement with the trpA cassette and WR644 for pat2 replacement with the hdrB cassette) and the double-deletion mutant (WR645) were comparable to that of the parental strain, H133 (data not shown).

The elp3 knockout is viable, but elp3 and pat2 are “synthetic lethals.”

The elp3 gene was replaced by the leuB selectable marker using the pMM1108 plasmid. It was possible to create an elp3 knockout in the wild-type background (WR660), as well as in the background of the Δpat1 strain (WR709). However, attempts to create an elp3 null strain in the Δpat2 single-mutant or the Δpat1 Δpat2 double-mutant strain failed. These results imply that elp3 and pat2 are “synthetic lethals,” namely, their products may share the same target(s).

Acetyltransferase knockout strains show no resistance to TSA.

The growth rates of all acetyltransferase knockout strains (Δpat1, Δpat2, Δpat1 Δpat2, Δelp3, and Δpat1 Δelp3) in rich HY medium were found to be comparable to those of the parental H133 strain (data not shown). We also examined the growth properties of the mutants in rich media containing TSA. However, none of the single-mutant strains were resistant to TSA. Similarly, attempts to knock out hdaI in strains carrying the various acetyltransferase gene knockouts also failed.

DISCUSSION

Posttranslational protein modification plays a key role in many cellular processes. The natures of these modifications cannot, in most cases, be deduced from genomic information and have to be determined by elaborate procedures. While archaeal genomic data are accumulating rapidly with innovations in genome-sequencing techniques, progress regarding archaeal protein modification is much slower. The present study describes a genetic approach to explore the potential of protein acetylation and deacetylation in H. volcanii by identifying the putative protein acetylase and protein deacetylase genes, followed by attempts to delete these genes. The results of this study are summarized in Table 3.

TABLE 3.

Summary of knockouts of H. volcanii acetylation-deacetylation genes

Gene	Status	Source
Single knockout
sir2	Viable	No selectable marker was used
hdaI	Essential	Knockout was obtained only in the presence of an extrachromosomal copy
pat1	Viable	Knockout was obtained using a selectable marker
pat2	Viable	Knockout was obtained using a selectable marker
elp3	Viable	Knockout was obtained using a selectable marker
Double knockout
pat1 pat2	Viable	Knockout was obtained using a selectable marker for both genes
pat1 elp3	Viable	Knockout was obtained using a selectable marker for both genes
pat2 elp3	Synthetic lethal	Knockout could not be obtained

The plethora of acetyltransferases and deacetylases in yeasts is not observed in archaea in general and in H. volcanii in particular. Unlike many organisms (8, 9), H. volcanii (and other archaea [27, 30]) contains only one putative Sir2 homolog. The second H. volcanii histone deacetylase is an HdaI homolog. Interestingly, Sir2 is widely distributed among bacteria and eukarya (9, 15), whereas in archaea, HdaI is more common. In fact, HdaI is present in all archaeal orders and is missing only from the Nanoarchaeota genome. Similarly, elp3 is widely distributed and highly conserved among all archaeal genomes (37, 49).

S. cerevisiae sir2 knockout mutants, as well as those in other eukaryotes, display a wide range of phenotypes, whereas the single H. volcanii sir2 knockout mutant had no recognizable phenotype. In contrast, the yeast hdaI knockout mutant had a mild phenotype and the hdaI rpd3 double mutant was viable (39). Since addition of TSA to the culture medium causes a severe defect in the growth of H. volcanii and deletion of the hdaI gene is lethal, it seems likely that HdaI is the main protein deacetylase in H. volcanii.

Single null mutations of S. cerevisiae gcn5 or elp3, homologs of H. volcanii pat2 and elp3, are viable, and the growth properties of the single-mutant strains are comparable to those of the wild-type strain under most conditions (53). However, S. cerevisiae elp3 and gcn5 double mutants, though viable, have more severe growth defects. These defects can be partly relieved by an accompanying hdaI deletion (52).

Similarly, single-deletion mutations in each of the three putative H. volcanii acetyltransferase genes had no discernible effect on growth or on TSA resistance. Nevertheless, it was not possible to delete both pat2 and elp3. These results imply that Pat2 and Elp3 share a substrate whose acetylation is essential and that this substrate cannot be acetylated by Pat1. The fact that hdaI is also essential indicates that indiscriminate acetylation of lysine residues is harmful and necessitates selective removal by protein deacetylases. So far, we have been unable to create a strain in which pat2, elp3, and hdaI are all deleted in analogy to the situation described above for S. cerevisiae.

Given that the acetylation/deacetylation machinery is an essential cellular process, we can address the issue of the nature of the possible protein targets at which it acts. The archaeal chromatin protein Alba was shown to undergo acetylation/deacetylation (31). Alba is found in many archaea but is missing in all halophilic archaea (40, 49). ACS was recently shown in S. enterica to be acetylated on residue K609 by Pat and deacetylated by CobB (a Sir2 homolog). Residue L641 was found to be important for enzymatic acetylation (44). H. volcanii has four ACS homologs. Two ACS genes (HVO_1585 and HVO_0894) are on the chromosome, while the other two genes (HVO_A0158 and HVO_A0156) are on an extrachromosomal megaplasmid (pHV4). The four enzymes show considerable similarity to the bacterial ACS, and their acetylation sites are also well preserved. Many other archaea have ACS homologs and have preserved the acetylation site (some examples are given in Fig. 7). Nevertheless, the H. volcanii sir2 deletion mutant displays no phenotypic growth impairments and, unlike the cobB mutant, can grow on minimal media using low concentrations of acetate as the sole carbon source. The only haloarchaeal protein known so far to be acetylated in vivo is the 2Fe-2S ferredoxin. The amino acid sequences of the Halobacterium salinarum (18) and Haloarcula marismortui (19) ferredoxins were determined and shown to contain a unique acetylated lysine close to their carboxyl termini. The conservation of the acetylated lysines in the two distantly related halophilic archaea might indicate their functional significance.

FIG. 7.

Amino acid alignment of different archaeal ACS enzymes. Lys609 of S. enterica ACS, which is known to undergo acetylation by Pat and deacetylation by Sir2, is marked with an arrow. Leu641 of S. enterica ACS, which is important for the acetylation/deacetylation process, is also marked. Black and gray shading represent amino acid identity and similarity, respectively. H. lacusprofundi, Halorubrum lacusprofundi; N. pharaonis, Natronomonas pharaonis; A. fulgidus, Archaeoglobus fulgidus.

Among the eukaryotes, the best-studied protein acetylation process is that of the amino-terminal histone tails. Their acetylation, and other posttranslational modifications, plays an important role in the regulation of gene expression (some models for regulation are reviewed in references 22, 24, 25, 34, and 36). H. volcanii histone lacks the tail extension and consists only of sequences homologous to the core sequence of the eukaryotic H3-H4 histones. Evidently, the potential target of the acetylation-deacetylation machinery in H. volcanii cannot be the histone tail. While acetylation of the histone tail is well documented, evidence regarding histone core modifications is gradually accumulating. Recently, a core acetylation modification was observed in Lys79 of histone H4 in yeasts (21, 56). This lysine residue is conserved in the core H3-H4 histone of most archaea and also in H. volcanii histone. This conservation, along with the physical location of the halophilic histone gene in an operon with the essential HdaI deacetylase gene, suggests that the histone is a possible target for acetylation.