To quantify histone protein levels we sought to measure histone protein and DNA content simultaneously using a flow cytometry based assay. This assay has several advantages over other methods of measuring cellular histone protein levels. In normal cells, histone protein levels are dependent on cellular DNA content. This fact has important consequences for the measurement of histone protein levels in different cell lines. Differences in the amount of DNA per haploid complement will result in concomitant differences in histone protein levels during interphase. Additionally, as cells replicate genomic DNA during S-phase, histone protein levels will increase along with the DNA content. Thus, simultaneous quantification of histone protein and DNA content allowed us to normalize histone protein levels to DNA content. The ability to detect the slight increases in both histone protein and DNA content as cells progress through S-phase attests to the quantitative sensitivity of our assay.

Histone H3 and DNA content were measured simultaneously by flow cytometry. The antibody used (see Materials and Methods) is directed against the H3 carboxyl-terminus and recognizes modified forms of H3 and H3 variants. In addition, H3 forms the core heterotetramer of nucleosomes, along with H4. It is important to note that somatic cells, regardless of cell cycle phase, contain no significant histone pools and levels of all four core histones and total linker histone are finely balanced. It has been estimated that even during S phase, the soluble histone pool represents about 0.1% of the total histone in human cells 6667. Therefore, the levels of H3 measured using this antibody are indicative of the nucleosomal levels on the chromosomes. During flow cytometry, the level of anti-H3 antibody bound per cell was measured using an Alexa 488-conjugated secondary antibody. Cells were also stained with propidium iodide (PI) to determine genomic DNA content. Unstained cells, as well as cells treated with primary antibody only, secondary antibody only, PI only, or primary and secondary antibody only were included as controls in every experimental analysis to determine background fluoresence and in instrument calibration.

Mean histone H3 signal intensity was traced from G1 through G2/M and plotted against mean genomic DNA content (Figure 1A). This was accomplished by defining a narrow window on each DNA histogram so that the histone H3 levels for the corresponding cell population defined by the window could be determined at a particular DNA content level. Our results show that upon initiation of DNA replication at the onset of S phase, cells exhibit a near linear increase in both DNA and histone H3 content. As cells approach G2/M histone H3 immunofluoresence was observed to plateau in all T-cell lines examined (both infected and uninfected), while PI fluorescence (DNA content) continued to increase in a near linear fashion. The plateau in H3 immunofluoresence observed may be explained by a decrease in H3 antibody epitope accessibility during chromosome condensation. Nevertheless, the simultaneous measurement of histone levels and DNA content using flow cytometry is clearly sensitive and quantitative enough to accurately measure the slight increases in both DNA and histone associated with genome replication as cells progress through S phase of the cell cycle. To display the histone H3 levels measured for each of the cell lines examined, the average histone H3:DNA ratios corresponding to G1, Mid-S and G2/M cell populations were calculated for each cell line and expressed relative to Jurkat (Figure 1B).

To summarize our results, histone H3 levels are not significantly different between each of the three uninfected T-cell lines, Jurkat, CEM and Molt4. In clear contrast, H3 levels are reduced 20 to 40% in the HTLV-1 infected T-cell lines, SLB-1, HuT102 and C81, relative to Jurkat cells. The p-values associated with the measurement of histone H3:DNA ratios are shown in Table 1. The reduction in histone H3 observed is consistent throughout the cell cycle.

Histone H1 protein levels were determined using the same flow cytometry-based assay. Histone H1 levels are plotted against genomic DNA content (Figure 1C) and histone H1:DNA ratios are determined (Figure 1D). Histone H1 protein levels are not significantly different in the two uninfected T-cell lines examined, Jurkat and Molt4. However, histone H1 protein levels are reduced by 40% in the two infected T-cell lines examined, C81 and HuT102 as compared to the uninfected T-cell lines, Molt4 and Jurkat. The p-values associated with the measurement of histone H1:DNA ratios are shown in Table 2. The reduction of histone H1 is also consistent throughout the cell cycle.

Having unequivocally determined that histone protein levels are reduced in HTLV-1 infected T-cell lines, we proceeded to determine if this was a result of reduced histone message levels.

The replication-dependent histones are expressed from multiple gene copies. While their amino acid sequences are highly conserved they are divergent enough in their DNA sequences as to require a separate primer set for the measurement of transcript from each of the gene copies by real-time PCR. Therefore, Northern blot analysis was employed allowing the measurement of total transcript levels for each replication-dependent histone type. Histone transcript levels in each cell line were normalized against EF1α. Northern blot analyses revealed a reduction of histone transcripts from all of the replication-dependent histone genes (Figure 2A). Quantification of the Northern blot results revealed a reduction in core and linker histone transcript levels ranging from two to five-fold, depending on the histone, in HTLV-1 infected T-cell lines (HUT102, SLB-1 and C81) as compared to Jurkat T-cells (Figure 2B). On average, histone transcripts are reduced by 66%, 61% and 50% in SLB-1, HuT102 and C81 cells respectively. Histone transcript levels in the other uninfected T-cell lines examined, Molt4 and CEM, were either not significantly different from or slightly higher than Jurkat. These results indicate that HTLV-1 infection is reducing histone protein levels through an effect on histone transcript levels.

We used reverse-transcription and real-time PCR to determine the transcript levels of each of the somatic linker histone isoforms. Real-time PCR quantification of linker histone cDNAs generated from an uninfected T-cell line (Jurkat) and from an HTLV-1 infected T-cell line (SLB-1) reveals a significant reduction of somatic linker histone transcripts in the HTLV-1 infected T-cell line when compared to the uninfected T-cell line (Figure 2C). H1^S-1, H1^S-2, H1^S-3, and H1^S-4 transcripts are 6, 14, 9 and 4-fold lower in SLB-1 cells, respectively, relative to Jurkat T-cells. These results indicate that HTLV-1 infection is reducing histone protein levels through an effect on histone transcript levels.

The HTLV-1 provirus encodes several viral proteins, but only the Tax protein is known to be oncogenic. To determine if Tax alone is sufficient to repress histone gene expression, we transiently transfected an uninfected T-cell line (Jurkat) with a Tax expression vector (pSG-Tax) or empty vector (pUC-19). Both sets of cells were cotransfected with a vector encoding a surface molecule (truncated mouse MHC Class I) to allow selection of transfectants by their retention on magnetic beads conjugated to antibody recognizing this unique surface marker. Real-time PCR was used to evaluate all of the somatic linker histone mRNAs and representative family members of the core histone genes in Tax expressing and pUC-19 transfected Jurkat cells. Tax expression was verified by both western blot and real-time PCR (not shown).

Both core and linker histone mRNAs were reduced in Tax expressing Jurkat cells as compared to pUC-19 transfected Jurkat cells (Figure 3A). Core and linker histone mRNA levels are reduced in the range of three- to five-fold, similar to the reduction observed in HTLV-1 infected T-cell lines. Replication-dependent histones are only transcribed in S phase of the cell cycle. To eliminate the possibility that Tax expressing cells reduce replication-dependent histone transcript levels by reducing the proportion of cells in S phase, we determined cell cycle distributions of pSG-Tax and pUC-19 transfected Jurkat cells. pSG-Tax and pUC-19 transfected Jurkat cells used for RNA extraction were subjected to cell cycle analysis. The percentage of cells in S phase of Tax expressing Jurkat cell populations was 31%, while that of pUC-19 transfected Jurkat cells was 36% (Figure 3B). The reduction of histone mRNAs observed in Tax expressing Jurkat cells due to the lower proportion of cells in S phase is no more than 14% (31/36). Therefore, Tax effects on cell cycle distribution do not account for the remaining 52 to 66% reduction in histone transcript levels observed in these Tax transfection studies.

In normal somatic cells, expression of the majority of histone genes is tightly coupled to the rate of DNA replication during S phase of the cell cycle. Thus, they are termed replication-dependent histone genes. Under normal circumstances inhibition of histone synthesis results in DNA replication arrest. Conversely, blocking DNA replication results in a coordinate inhibition of histone expression 68. The coordination of histone expression and DNA synthesis is critical in the maintenance of chromosomal integrity 69. Histone chaperones and chromatin assembly factors normally cooperate in the deposition of histones onto nascent DNA as it emerges from the replication fork 70. The down-regulation of histone gene expression in response to decreased rates of DNA synthesis prevents the accumulation of excess histones, which has been shown to be highly cytotoxic and to severely compromise the transmission of chromosomal information 71. Since there is no significant pool of non-nucleosomal histones, chromatin assembly at the replication fork is dependent on de novo histone expression.

Mammalian cell division requires the synthesis of 10⁸ histone proteins during the brief S phase. In order to meet these demands, cells contain multiple copies of the replication-dependent histone genes. Genomic sequence analysis has revealed 59 replication-dependent core histone genes and 6 replication-dependent linker histone genes, clustered at 4 different chromosomal loci in human cells 72. The expression of all the replication-dependent histone genes is coordinately regulated during S phase. This coordinate regulation is mediated through both transcriptional and post-transcriptional mechanisms. It is not known how each of these different transcriptional and post-transcriptional mechanisms of histone gene regulation are coupled to the rate of DNA synthesis in proliferating cells.

The rate of transcription of all the replication-dependent histone genes increases three- to five-fold as cells enter S phase. Studies have shown that proper cell cycle-dependent transcriptional regulation of each histone subtype involves unique cis-regulatory elements and unique transcription factors for each histone subtype. It is not known how these different DNA-binding transcription factors are coordinated to activate transcription of all the replication-dependent histone genes in response to cell cycle signaling 7374. Recently, non DNA-binding proteins have been identified that may be involved in coordinating histone gene transcription. Hir1p and Hir2p are cell cycle regulators thought to be responsible for transcriptional co-repression of histone gene expression in Saccharomyces cerevisiae 75. HIRA is the mammalian homolog of these yeast proteins and may regulate histone gene expression in mammalian cells. HIRA has been shown to display histone chaperone function independent of DNA synthesis 76. It has also been shown to be phosphorylated by cyclin E/cdk2 during S phase and is thought to be inhibited by this phosphorylation. The ectopic expression of HIRA is known to result in S phase arrest and reduced histone transcript levels, suggesting a negative function for HIRA in replication-dependent histone gene expression outside of S phase 77.

NPAT is another mammalian protein thought to function in the coordinate expression of replication-dependent histone genes 7879. NPAT has also been shown to be an in vivo substrate of cyclin E/cdk2. Phosphorylation of NPAT correlates with the colocalization of NPAT to histone processing bodies. Inhibition of NPAT gene expression impedes cell cycle progression and decreases histone gene expression in mammalian cells 80. Thus, NPAT is thought to participate in the up-regulation of replication-dependent histone gene expression during S phase.

Post-transcriptional mechanisms of histone gene expression are responsible for an additional eight- to ten-fold increase in mRNA levels as cells enter S phase 8182. The replication-dependent histones do not contain introns and are not polyadenylated. Instead, they all contain a very similar 3' stem-loop structure bound by a conserved protein in mammals, stem-loop binding protein (SLBP) 8384. SLBP mediates the only RNA processing event required for production of mature histone mRNAs, a 3'-endonucleolytic cleavage 85. In addition, SLBP binding regulates the half-life and translational efficiency of histone mRNAs 868788.

As the promoters for the genes encoding the four replication-dependent histones have been shown to utilize unique cis-acting elements bound by different sequence-specific trans-acting factors during S-phase induced expression, Tax mediated repression of histone expression is likely to occur through an effect on one or more of the coordinating regulators of histone expression, HIRA, NPAT, or SLBP. The difference in reduction of histone protein levels (20 to 40%) and histone transcript levels (50 to 80%) suggests some compensation may occur through increased mRNA stability, translation efficiency or protein stability. In addition, results from this study suggest that HTLV-1 infection is able to uncouple the interdependence of histone gene expression and DNA synthesis. There is little known about the mechanism coupling histone synthesis and DNA replication. One candidate for this function is the histone chaperone Asf1 (Asf1a and Asf1b in mammalian cells), which has been shown to regulate histone levels during replicational stress 89. Tax may provide a useful probe to investigate this coupling mechanism, as well as the mechanism of coordinate histone gene regulation in general.

We find HTLV-1 infected cells with ~30% reductions in histone levels form apparently normal metaphase chromosomes, at least at the resolution of light microscopy (not shown). This indicates the reduction has more subtle effects. At the same time, we find that lower histone levels are not compensated for by merely increasing the nucleosomal repeat length to evenly distribute the remaining nucleosomes (not shown). Thus, one must assume there is a corresponding increase in non-nucleosomal DNA.

Chromatin structure plays a central role in the regulation of all DNA metabolic processes including transcription, replication, repair and recombination 61. Post-translational modifications on the histone tails and histone variants are thought to be critical in the determination of cellular identity 9091. As every cell in a multi-cellular organism contains the same genomic information, cell differentiation and specialization require that distinct transcriptional programs are established and maintained 9293. Post-translational modifications on the histone tails are thought to impart cellular memory through epigenetic control of the genome 93. Thus, every unique cell type is thought to exhibit a unique set of chromosome "marks", governing the genome-wide transcriptome and ultimately cellular identity. We suggest that a reduction of histone levels in HTLV-1 infected T-cells will result in loss of epigenetic information and may contribute to Tax deregulation of cellular gene expression leading to cancer.

We expect reduced histone levels to have similar effects as those observed with loss of imprinting (LOI) through DNA hypomethylation. These effects include aberrant chromosome rearrangements, deregulation of cellular gene expression and activation of latent viral genomes 949596979899100. Reduction of histone levels could provide a means for HTLV-1 transcriptional reactivation after methylation of the 5' LTR in cultured ATLL cells 101.

Defects in chromatin assembly have been shown to impair all known pathways of double-strand break repair and activate S phase arrest 102103. Furthermore, partial depletion of histone H4 has been shown to increase homologous recombination-mediated genetic instability 104.

Chromatin structure, nucleosome remodeling and the epigenetic information encoded on the histone tails are known to participate in double-strand DNA break repair 105106107108. Defects in nucleosome assembly have been shown to interfere with the repair of double-strand DNA breaks 102. We suggest that a Tax-mediated decrease in histone levels will result in an increase in the incidence of double-strand DNA breaks. The ability of Tax to increase the frequency of double-strand DNA break formation is consistent with its ability to increase the propensity for micronuclei formation 5556. Double-strand DNA breaks are potent inducers of genome instability as they are repaired by the error-prone process of non-homologous end joining, when a sister chromatid is not available for more efficient recombination-mediated repair 109. Decreased histone levels may represent a previously unknown means for Tax induction of DNA damage.

(C81, CEM, HuT102, Jurkat, Molt-4, and SLB-1) T-cell lines were cultured in Iscove's Modified Dulbecco's Media (IMDM; Sigma-Aldrich) supplemented with 10% fetal bovine serum, 50,000 units/liter each streptomycin/penicillin and 2 mM L-glutamine. All T-cell lines were cultured at 37°C in the presence of 5% CO₂.

RNA was extracted from asynchronous cells (10⁷) using Trizol (Invitrogen) in accordance with the manufacturer's protocol. In order to reduce genomic contamination, RNA samples were treated with 20 units of RNase-free DNase I (Roche) at room temperature for 30 minutes before heat inactivating the DNase I at 65°C for 10 minutes. cDNA was synthesized with the iScript cDNA synthesis kit (Bio-Rad) according to manufacturer's protocol using a programmable thermal cycler.

Relative real-time PCR was performed using iQ SYBR Green Supermix (Bio-Rad) on an iCycler (Bio-Rad) and data expressed according to the Michael Pfaffl method, using EF1α as an internal control for normalization 110. A five-point standard curve was constructed by serial diluting target DNA to evaluate PCR efficiency for every primer pair in every experiment. Reactions were performed in triplicate for every experiment. Real-time PCR amplicons were electrophoresed on 1.5% agarose gels to verify the specificity of each primer pair and melt curves were analyzed for every experiment to ensure that amplification products were primer pair specific. The following primer sequences were used for real-time PCR:

H2A Fwd AGCTCAACAAGCTTCTGGGCAA;

H2A Rev TTGTGGTGGCTCTCGGTCTTCTT;

H2B Fwd TGCGCCCAAGAAGGGTTCTAAA;

H2B Rev ACGAAGGAGTTCATGATGCCCA;

H3 Fwd TGCTCATCCGCAAACTGCCATT; H3 Rev AGTGACACGCTTGGCGTGAATA; H4 Fwd ACCGTAAAGTACTGCGCGACAA; H4 Rev TTCTCCAGGAACACCTTCAGCA; H1S-1 Fwd CCTGTAAAGAAGAAGGCGGCCAAA;

H1S-1 Rev CAGAGAAACTCCGCTACGCTCTTT;

H1S-2 Fwd CCCAGTATCTGAGCTTATCACCAAGG;

H1S-2 Rev TTTCTTAAGCGCGGCCAGAGAAAC;

H1S-3 Fwd CCCGGCTAAGAAGAAGGCAACTAA;

H1S-3 Rev GAAAGGCCATTGCGCTCCTTAGAA;

H1S-4 Fwd CCGGTGTCCGAGCTCATTACTAAA;

H1S-4 Rev GCTTTCTTGAGAGCGGCCAAAGAT;

EF1α Fwd GCCTCTCCAGGATGTCTACAAA;

EF1α Rev GTTTGAGAACACCAGTCTCCACTC;

Tax Fwd TTCTACCCGAAGACTGTTTGCCCA;

Tax Rev TGTCCAAATAAGGCCTGGAGTGGT.

Asynchronous cells (5 × 10⁶) from log phase culture were pelleted and suspended in ice-cold phosphate buffered saline (PBS). Cell suspensions were transferred to ice-cold 1% formaldehyde (methanol-free) in PBS and chilled on ice for 15 minutes to fix cells. Cells were washed with ice-cold PBS and transferred to ice-cold 70% ethanol and let stand at -20°C for 2 hours to permeablize cells. Cells were washed with PBS and incubated in 1% BSA-PBS containing 5 μg primary antibody (ab1791 rabbit polyclonal to Histone H3, AbCam or mouse monoclonal antibody directed against histone H1 product # V7013, Biomeda), overnight at 4°C while rotating. Cells were then washed with 1% BSA-PBS and incubated in a 1% BSA-PBS solution containing 5 μg of Alexa Fluor 488 goat anti-rabbit second antibody (product # A11008, Molecular Probes) for Histone H3 analysis or Alexa Fluor 488 goat anti-mouse second antibody (product # A11029, Molecular Probess) for Histone H1 analysis, for 1 hour at room temperature. Cells were washed with 1% BSA-PBS and incubated in PBS containing 25 ug/mL propidium iodide (Sigma) and 40 Kunitz U/mL RNase A (Sigma, certified DNase-free) for 30 minutes at room temperature. Cells were analyzed on a MoFlo (Dako Cytomation) flow cytometer/cell sorter with laser excitation at 488 nm with 110 mW. Data was interpreted using Summit software, version 4.0 (Dako Cytomation).

Total RNA (10 μg) was resolved on 1% agarose/6.6% formaldehyde gels. Samples were transferred to a nylon membrane and probed with histone and EF1α probes simultaneously. PhosphorImager screens were scanned on a STORM (Molecular Probes). Histone signals were quantified with ImageQuant software version 5.1 (Molecular Dynamics) and normalized to EF1α signals. Probes were generated using a random primed DNA labeling kit (Roche). The DNA templates used for the generation of ³²P-radiolabeled probes for histone and EF1α transcripts were prepared by PCR from Jurkat genomic DNA using the following primers:

H1 Fwd AAGGAGCGCAATGGCCTTTCTTTG;

H1 Rev GCCTTCTTGGCCTTTGCAGCTTTA;

H2A Fwd TTCAGTTTCCCGTAGGCCGAGT, H2A Rev GAAGCTTGTTGAGCTCCTCATCGT;

H2B Fwd AAGAAGGATGGCAAGAAGCGCAAG;

H2B Rev ACTTGGAGCTGGTGTACTTGGTGA;

H3 Fwd ATCGCTATCGGCCTGGTACAGT;

H3 Rev AGCTGGATGTCCTTGGGCATGATA;

H4 Fwd ATGTCAGGACGCGGCAAAGGAGGTAA;

H4 Rev GATCACGTTCTCCAGGAACACCTTCA;

EF1α Fwd TCATTGGACACGTAGATTCGGGCA;

EF1α Rev TTCACACCCAGTGTGTAAGCCAGA.

The DNA template used for the generation of ³²P-radiolabeled probes for Tax transcripts was prepared by PCR from pSG-Tax plasmid DNA using the following primers:

Tax Fwd TTCCCAGGGTTTGGACAGAGTCTT;

Tax Rev GCGTGCCATCGGTAAATGTCCAAA.

Jurkat cells were co-transfected with pSG-Tax/pMACs K^k or pUC-19/pMACs K^k at 20 ug total DNA (3.5 molar excess of pMACs K^k) per 10⁷ cells by electroporation in 0.4 cm cuvettes using Genepulse (Bio-Rad). Transfected cells were enriched 24 hours following electroporation using the pMACS K^k magnetic selection system (Miltenyi Biotec) according to the manufacturer's protocol. Enriched cell populations were cultured for an additional 24 hours prior to harvesting cells for RNA extraction and DNA content analysis.

Asynchronous cells (10⁶) from log phase culture were fixed with 70% ethanol and low molecular weight DNA was extracted with a 4 mM sodium citrate/192 mM sodium phosphate wash. Fixed cells were stained with 50 ug/mL propidium iodide in PBS containing 730 Kunitz U/mL RNase A for 20 minutes at room temperature. Data was collected on an EPICS flow cytometer (Beckman Coulter) and cell cycle distributions were calculated using Multicycle (Phoenix Flow Systems).