Báo cáo khoa học: Characterization, localization and possible anti-inflammatory function of rat histone H4 mRNA variants potx

Characterization, localization and possible anti-inflammatory function of rat histone H4 mRNA variants Rene ´ Poirier, Irma Lemaire and Simon Lemaire Department of Cellular and Molecular Medicine, University of Ottawa, Canada Histones are known to play a key role in the pack- aging of DNA within eukaryotic cells. The majority of histone proteins or ‘core histones’ are produced during the synthesis (S) phase of the cell cycle [1]. Core his- tone mRNAs do not end with a polyadenylated tail but, instead, contain within their 3¢UTR a conserved stem–loop sequence that is involved in their matur- ation and function [2]. There are also replication- independent histone variants that transcribe poly- adenylated mRNAs and whose translation products accumulate preferentially in nondividing, terminally differentiated tissues [3]. In contrast with core histone mRNAs, histone mRNA variants can be expressed throughout all phases of the cell cycle in inducible and tissue-specific ways [3]. Histones derived from replica- tion-independent mRNAs were originally suggested to Keywords C-terminal H4 peptides; extracellular function; histogranin; histone H4 mRNA variants; H4-v.1 Correspondence S. Lemaire, Department of Cellular and Molecular Medicine, Faculty of Medicine, University of Ottawa, 451 Smyth Road, Ottawa, Ontario, Canada K1H-8M5 Fax: +1 613 562 5646 Tel: +1 613 562 5800 ext. 8350 E-mail: slemaire@uottawa.ca (Received 7 July 2006, accepted 1 August 2006) doi:10.1111/j.1742-4658.2006.05444.x Two histone H4 mRNA variants, H4-v.1 and histogranin mRNAs, were detected in the rat genome and measured in various tissues and isolated alveolar macrophages. Medium to high levels of both mRNAs were present in the liver, adrenal glands, thymus, bone marrow and alveolar macrophag- es. H4-v.1 cDNA contained an open reading frame that coded for unmodi- fied whole histone H4, whereas histogranin cDNA lacked the first ATG codon and contained an open reading frame that coded for modified (Thr89) H4-(84–102). The two genes displayed a sequence homologous (> 80%) to the open reading frame of core H4 somatic (H4s) and H4 ger- minal (H4g) and their variant nature was supported by the absence of histone consensus palindromic and purine-rich sequences in the proximal 3¢UTR, and the presence of a polyadenylation signal in the distal 3¢UTR and of specific upstream transcription factor-binding sites. H4-v.1 and his- togranin transcripts, but not H4s transcript, were selectively induced by lipopolysaccharide and ⁄ or interferon gamma in alveolar macrophages. In vitro transcription ⁄ translation experiments with H4-v.1 and histogranin cDNA pCMV constructs produced peptides with the molecular mass (2 kDa) of the alternative histone H4 translation product which, like syn- thetic H4-(86–100) and [Thr89]H4-(86–100) or rat histogranin, inhibited lipopolysaccharide-induced prostaglandin E 2 release from rat alveolar macrophages. The synthetic peptides also inhibited the secretion of the CXC chemokine interleukin-8 (GRO ⁄ CINC-1) in response to lipopolysac- charide. The presence of H4-v.1 and histogranin mRNAs in tissues wherein immune reactions take place and the inhibitory effects of their translation products on prostaglandin E 2 and interkeukin-8 secretion by activated alveolar macrophages suggest an anti-inflammatory function. Abbreviations AM, alveolar macrophage; AP, amplification primer; BAL, bronchoalveolar lavage; EST, expressed sequence tag; GSP, gene-specific primer; HN, histogranin; IFN, interferon; IL, interleukin; LPS, lipopolysaccharide; NF-jB, nuclear factor kappa B; OGP, osteogenic growth peptide; PGE 2 , prostaglandin E 2 ; SP1, stimulating protein 1. 4360 FEBS Journal 273 (2006) 4360–4373 ª 2006 The Authors Journal compilation ª 2006 FEBS constitute a replacement pool of histones for nucleo- somal maintenance throughout the cell cycle [4]. Recently, their expression was shown to regulate var- ious processes that comprise heterochromatin ectopic spread [5–7], DNA transcription (H2A.Z) [8–10], cen- tromere formation (CENP-A) [11–13], X chromosome inactivation (macroH2A) [14,15] and DNA repair (H2AX) [16–19]. The observation that the translation of histone mRNA variants follows the rules of typical poly(A) track-containing mRNAs [20] suggests that histone variants may also exert extranuclear functions. In this regard, histones were reported to modulate pituitary hormone secretions [21–24], pathogenic anti- body production [25–27], microbial [28,29] and tumor- al [30] cell growth, osteogenesis [31,32], pain [33–36] and macrophage proinflammatory functions [37]. Histogranin (HN), a slightly modified C-terminal histone H4 peptide homologous to histone H4-(86– 100), was first isolated in our laboratory from bovine adrenal medulla [38]. The immunoreactive peptide was detected in various rat tissues, including the pituitary, adrenal glands, lungs, spleen, brain and plasma [39]. Synthetic HN was initially shown to block N-methyl- d-aspartate-induced convulsions in mice [38]. More recently, HN and related peptides were also shown to display in vivo nonopioid analgesic effects and in vitro anti-inflammatory activity [33–40]. Our initial search to determine the structure of the HN gene was unsuc- cessful but led to the discovery of the H4 mRNA vari- ant H4-v.1 [41]. H4-v.1 was first isolated and sequenced from a bovine adrenal medullary cDNA phage library [41]. Bovine H4-v.1 was then shown to be a polyadenylated mRNA coding for unmodified whole histone H4. A similar mRNA variant was also detected in the rat using a cDNA probe that recog- nized part of the bovine H4-v.1 coding region, although its sequence was not determined [42]. A close correlation was then observed between the level of H4- v.1 in various rat tissues and alveolar macrophages (AMs) and the amounts of the histone H4 C-terminal peptides, osteogenic growth peptide (OGP) [31] and H4-(86–100), but not whole histone H4 protein or core H4 mRNA [42]. This study suggested that the extra- cellularly acting unmodified C-terminal histone H4 peptides OGP and H4-(86–100) may be translation products of the alternative AUG start codon in H4- v.1, but not core H4 mRNA. On the other hand, the modified nature of the C-terminal histone H4 peptide HN indicated that its synthesis might depend upon the expression of another H4 mRNA variant akin to some other types of histone variant, such as the H3 mRNA variants that produce modified histone H3 proteins [43,44]. As no report has indicated the structure of rat H4- v.1 and HN mRNAs, we herein used rat genome data- bases to search for the H4 variant candidates as well as specific molecular approaches and in vitro assays to assess their structure, expression and function. We also verified whether the cell cycle regulatory region and site II element, known to regulate the expression of H4 genes [45–47] or other specific elements, were present upstream of the H4 mRNA variants. The results con- firm the existence and illustrate the structures of H4- v.1 and HN genes, two polyadenylated histone H4 mRNA variants with characteristics of the replication- independent histone genes coding for unmodified whole histone H4 and a modified C-terminal histone H4 peptide, respectively. The particular localization of the two genes in rat tissues, the identification of upstream gene-specific regulatory elements and the in vitro transcription ⁄ translation experiments with gene-specific cDNA constructs indicate that the two genes are independently expressed and produce C-ter- minal H4 peptides with in vitro anti-inflammatory activity. Results Gene BLAST search A blast search of the rat TIGR database provided an expressed sequence tag (EST) sequence (TC: 322388) that resembled that of the bovine H4-v.1 mRNA variant [41]. Like bovine H4-v.1, this EST sequence contained an ORF that coded for unmodi- fied histone H4 and a 3¢UTR that ended with an AATAAA polyadenylation signal. Conversely, the EST sequence was not complete at its 5¢ end, lack- ing the 5¢UTR and the first ATG initiation codon present in bovine H4-v.1. On the other hand, a tblastn search of NCBI for histone H4 in the rat genomic sequence provided another sequence (NW_047492.1|Rn17_2014:1861737–1862016) that did not code for whole histone H4, but a modified his- tone H4 C-terminal peptide, [Thr89]H4-(84–102). Since histone H4 is one of the most evolutionarily conserved proteins [48], it was assumed that if a gene with this modified H4 coding region was expressed, it could be the gene encoding the modified C-terminal H4 peptide HN [38], generating in this case [Thr 89 ]H4-(84–102) as an immediate precursor of rat HN. Next, we sought to verify the expression of H4- v.1 and HN mRNA transcripts in total mRNA prep- arations from various rat tissues and isolated AMs, and determine the complete structures of rat H4-v.1 and HN mRNAs. R. Poirier et al. Histone H4 mRNA variants FEBS Journal 273 (2006) 4360–4373 ª 2006 The Authors Journal compilation ª 2006 FEBS 4361 Localization of H4-v.1 and HN mRNAs Because H4-v.1 mRNA and immunoreactive HN had already been detected within various rat tissues using probes and antibodies that recognized bovine H4-v.1 mRNA and HN, respectively [39,42], initial tests were performed to assess the level of expression of rat H4- v.1 and HN mRNAs in total RNA preparations of various rat tissues by the use of real-time PCR with gene-specific primers (GSPs) designed from the blast information. High levels of both HN and H4-v.1 mRNA transcripts were observed in the liver, adrenal glands, thymus and bone marrow (Fig. 1A). However, the HN mRNA transcript was more widely distri- buted, being also abundant in endocrine, neuroendo- crine and central nervous system tissues such as the pituitaries, the spinal cord and the brain. Both mRNAs were also detected in AMs, and their levels were compared with that of core H4s (Fig. 1B). HN mRNA was markedly increased by incubation of AMs in the presence of interferon gamma (IFN-c) (8.97- fold), whereas H4-v.1 was significantly stimulated by both lipopolysaccharide (LPS) (2.67-fold) and IFN-c (3.48-fold). In contrast, the level of H4s mRNA, although 4.7 and 6.2 times higher than those of control H4-v.1 and HN mRNAs, respectively, was not signifi- cantly affected by incubation of AMs with LPS or IFN-c. Sequence determination of H4-v.1 and HN cDNAs Determination of the length and sequences of H4-v.1 and HN mRNAs was accomplished by 3¢RACE and 5¢RACE experiments using a Marathon-Ready TM rat spleen cDNA library. The 5¢RACE and 3 ¢RACE ampli- cons of H4-v.1 and HN were designed to overlap one another, resulting in complete cDNA structure amplifi- cation. The 5¢RACE and 3¢RACE amplicons were inser- ted into the TOPO cloning vector (Invitrogen) and sequenced. Complete H4-v.1 and HN cDNAs (Gen- Bank accession numbers: AY936209 and AY936210, respectively) were compared with their respective sequences within the rat genome. H4-v.1 and HN cDNA sequences were considered to be accurate if three separ- ate sets of sequenced 5¢RACE and 3¢RACE amplicon clones and the corresponding genome sequences in the NCBI genome database could be matched. The H4-v.1 cDNA transcript contained a short 5¢UTR (19 bp), an ORF corresponding to unmodified whole histone H4, a 3147 bp missing intron, and a relatively long 3¢UTR (965 bp) ending with a polyadenylation signal (AAT AAA) and an auxiliary mRNA-processing facilitator- like element (AAAGAT) (Fig. 2A; AY936209). On the other hand, the HN cDNA transcript contained a relatively long 5¢UTR (253 bp), a short ORF coding for MDVVYTLKRQGRTLYGFGG as an immediate Fig. 1. Relative abundance of H4-v.1 and HN mRNA transcripts in various rat tissues (A) and isolated AMs (B). (A) Total RNA was extracted from rat tissues (three pools of two animals) and the amounts of H4-v.1 and HN mRNAs were determined by real-time PCR with gene-specific primers as described in Experimental proce- dures. The relative abundance of the cDNA amplicons was meas- ured in comparison with GAPDH, using the lung as a comparative tissue for calculation in the equation: mRNA ¼ 2 ) [(Ct mRNA test tissue ) Ct GAPDH test tissue) ) (Ct mRNA comparative tis- sue ) Ct GAPDH comparative tissue)]. In (B), freshly isolated AMs (three preparations from two animals each) were incubated for 24 h in the absence or presence of LPS (1 lgÆmL )1 ) or IFN-c (100 UÆmL )1 ) prior to total RNA extraction and measurement of H4- v.1 and HN mRNA transcripts in comparison with core H4s. Results represent the mean ± SEM of three duplicated sets of experi- ments. Statistical significance was determined using one-way ana- lysis of variance followed by a Bonferonni comparison test. (A) *P 6 0.05 as compared with heart H4-v.1 cDNA amplicon; § P 6 0.05 as compared with heart HH cDNA amplicon; † P 6 0.05 as compared with H4-v.1 cDNA amplicon in the same tissue. (B) § P 6 0.05 as compared with control H4-v.1 mRNA; *P 6 0.05 as compared with control HN mRNA. Histone H4 mRNA variants R. Poirier et al. 4362 FEBS Journal 273 (2006) 4360–4373 ª 2006 The Authors Journal compilation ª 2006 FEBS precursor of rat HN (VVYTLKRQGRTLYGF, the portion of the peptide homologous to bovine HN [38]) and a relatively long 3¢UTR (273 bp) ending with a non- canonical polyadenylation signal (TATAAA) and an auxiliary mRNA-processing facilitator-like element (AAAGAT) (Fig. 2B; AY936209). Comparisons of H4-v.1 and HN cDNAs with core histone H4 cDNAs A comparison of H4-v.1 and HN cDNAs with core germinal (H4g) and somatic (H4s) histone H4 genes showed 80–92% homology in a region corresponding to the ORF of core histone H4 genes (Fig. 3). The nucleotide substitutions in H4-v.1 did not affect the highly conserved amino acid structure of the whole histone H4 protein or the alternative initiation transla- tion product H4-(84–102) (Fig. 2A). On the other hand, the HN cDNA shared a high degree of homol- ogy with the histone H4 coding region (Fig. 3), but lacked the first ATG codon necessary to translate the whole histone H4 protein and contained a modified codon (ACT coding for Thr instead of Ala) in the alternative ORF sequence to code for [Thr 89 ]H4-(84– 102) (Fig. 2B). Comparison of the structures of the proximal 3¢UTR of HN cDNA with those of H4g and H4s revealed a GC-rich stem–loop structure followed clo- sely by a purine-rich region similar to the histone con- sensus palindromic and purine-rich sequences of H4g and H4s (Table 1). However, the stem–loop structure found in HN was considered to be noncanonical, being distinct from that found in other histone genes [48]. No comparable stem–loop structure was found in the proximal 3¢UTR of H4-v.1. Interestingly, the 3¢UTR of both H4-v.1 and HN cDNAs contained an ATTT repeat element (14 and 4 repeats, respectively; AY936209 and AY936210) that is known to play a post-transcriptional role in the synthesis of cytokines in lymphoid cells [49]. Finally, the distal 3¢UTR of H4-v.1 and HN cDNAs, but not H4g or H4s, con- tained a polyadenylation signal characteristic of his- tone cDNA variants (Table 1). Comparison of upstream genome sequences of H4-v.1 and HN genes with those of H4s and H4g indicated a region similar to the site II cell cycle regula- tory domain of the replication-dependent histone H4 genes (Table 1) comprising a TATA box-like motif, a histone H4-specific GGTCCG element, and a motif homologous to the human histone H4 gene cell cycle Fig. 2. Schematic representation of H4-v.1 (A) and HN (B) genes in the rat genome. Rat cDNA structures were determined by combined 5¢RACE and 3¢RACE with a Mara- thon-Ready rat spleen cDNA library as des- cribed in Experimental procedures. The complete sequences of H4-v.1 and HN cDNAs were submitted to the NIH GenBank and have been given the accession numbers AY936209 and AY936210, respectively. The ORFs of the H4-v.1 and HN genes encode complete histone H4 protein and the H4 C-terminal peptide MDVVYTLKRQGRTLYGFGG, respectively (Fig. 3). The 5¢UTR, ORF, 3¢UTR, stem–loop sequence and polyadenylation signals are located as indicated in the schemes. Both genes are preceded by the gene-specific promoters as described in Table 1 and as illustrated. The H4-v.1 gene contains a 3.5 kb intron and does not contain the canonical histone stem–loop sequence 30 nucleotides downstream of the ORF. The HN gene does not contain an intron and its stem–loop sequence downstream of the ORF is distinct from that of histone H4 genes [48]. R. Poirier et al. Histone H4 mRNA variants FEBS Journal 273 (2006) 4360–4373 ª 2006 The Authors Journal compilation ª 2006 FEBS 4363 control motif (5¢-CTTTCGGTTTT-3¢) [46]. Other potential transcriptional regulatory binding motifs close to the site II regulatory domain of the H4-v.1 gene included the mitogen-activated protein (MAP) kinase transcription factor ELK-1 ( )44 taagacGGAActgcttt )28 ), a MAP kinase substrate transcription factor involved in cell growth [50] and the cyclic AMP response elements CREB ( )88 tccgccTGACgctccctgttt )69 ) and CREB-P1 ( )152 ttgctcttACATgaactgaaa )132 ), two tran- scription factors involved in the regulation of metabolic and neuronal activities [51] (Fig. 2). Sequences upstream of the HN gene comprised the Elk-1 motif ( )33 gtacacGGAAgttttag )17 ) [52], the GC box motif ( )121 aaatgaGGCGgagcaa )107 ), a specific stimulating protein 1 (SP1)-binding site that can modulate the action of the nuclear factor kappa B (NF-jB) DNA site [52] and the NF-jB motif ( )181 tgGGGAaaacccc ag )167 ), a transcription factor involved in the matur- ation of immune cells and inflammation processes [53] (Fig. 2). None of these sequences found upstream of either the H4-v.1 gene or HN gene was observed upstream of the replication-dependent H4g (NCBI #m27433) and H4s (NCBI #x13554) genes. Transcription ⁄ translation in an in vitro wheat germ lysate system The presence in H4-v.1 cDNA of both initial and alternative ATG codons should allow its translation into both whole histone H4 and the C-terminal peptide Fig. 3. Comparison between the structures of the ORF of H4 somatic (H4s), H4 germi- nal (H4g) and H4-v.1 cDNAs and corres- ponding 5¢UTR and ORF in HN cDNA. Analyses were done with the BLAST 2 sequences of NCBI. Start and stop codons are indicated by bold letters. H4s (accession number x13554) is 84% and 80% homolog- ous with H4-v.1 and HN, respectively, and 84% homologous with H4g (accession num- ber m27433). H4g is 92% and 88% homol- ogous with H4-v.1 and HN, respectively, whereas H4-v.1 is 89% homologous with HN. H4s, H4g and H4-v.1 cDNA sequences code for unmodified whole histone H4. HN cDNA does not contain the initial ATG codon found in H4s, H4g and H4-v.1 cDNAs, thus giving rise to a translation product (pro-HN) of 19 amino acids corres- ponding to the alternative translation prod- uct in the other genes with a modification at position 89 (T instead of A). The initial M (M 0 ) in the translated H4 protein is cleaved to give rise to a protein of 102 amino acids [48]. Histone H4 mRNA variants R. Poirier et al. 4364 FEBS Journal 273 (2006) 4360–4373 ª 2006 The Authors Journal compilation ª 2006 FEBS H4-(84–102) (Fig. 3) [41]. The presence in HN cDNA of only the alternative H4 ATG codon should allow its translation only into [Thr89]H4-(84–102) as an immediate precursor of rat HN. Experiments per- formed using pCMVTnT vector constructs with the in vitro wheat germ lysate-coupled transcription ⁄ trans- lation system (TnT, Promega) indicated that the H4-v.1 construct synthesizes two radiolabeled (Met35) protein ⁄ peptide products, one comigrating on SDS gel electrophoresis with whole histone H4 (11.4 kDa), and the other comigrating with synthetic H4-(84–102) (approximately 2 kDa) (Fig. 4A). On the other hand, the HN construct produced only one radiolabeled compound, which migrated within the expected molecular mass range of [Thr89]H4-(84–102) (approxi- mately 2 kDa) (Fig. 4A). The empty pCMVTnT vector did not produce any radiolabeled protein ⁄ peptide product, whereas the Promega luciferase control plas- mid showed multiple radiolabeled products, the major band corresponding to luciferase (60 kDa) (Fig. 4A). Inhibition of LPS-induced prostaglandin E 2 (PGE 2 ) release from cultured AMs Prostaglandins are known to play an important role in inflammation and pain. Rat AMs stimulated with LPS (1 lgÆmL )1 ), the archetypal bacterial antigen, produced significant amounts of PGE 2 (171.3 ± 23 pgÆmL )1 compared to 76 ± 8.9 pgÆmL )1 for unstimulated cells). As shown in Fig. 4B, LPS-stimulated release of PGE 2 from primary cultures of rat AMs was reduced to 49.8% and 46.3% of the control value in the presence of 10 )8 m synthetic H4-(86–100) and [Thr89]H4-(86– 100) (rat HN), respectively. Incubation of AMs in the presence of the transcription ⁄ translation HN-pCMV- TnT and H4-v.1-pCMVTnT products (20 lL) also reduced LPS-evoked release of PGE 2 to 54.5% and 49.4% of the control value, respectively (Fig. 4B). In contrast, the transcription ⁄ translation product (20 lL) of the empty pCMVTnT plasmid had no significant effect on PGE 2 release. Inhibition of LPS-induced rat interleukin-8 (IL-8) (GRO/CINC-1) The CXC chemokine IL-8 is a potent neutrophil chem- otactic and activating agent. As IL-8 and its rat ana- log, GRO ⁄ CINC-1, are reported to be produced by human and rat AMs [68], we next investigated whether the synthetic translation products of H4-v.1 and HN mRNAs also modulated the secretion of this inflam- matory cytokine. Rat AMs incubated with LPS (1 lgÆmL )1 ) for 4 h released into the culture medium significant amounts of GRO ⁄ CINC-1 (4023 ± 325 pgÆmL )1 ), whereas GRO⁄ CINC-1 was undetectable in culture supernatants of unstimulated AMs. Incubation of AMs with 10 )8 m H4-(86–100) and [Thr89]H4-(86– 100) significantly decreased LPS-induced GRO ⁄ CINC- 1 secretion (to 58.3% and 62.5% of the control value, respectively) (Fig. 5). AM survival following treatment with H4-v.1 and HN products To verify whether inhibition of LPS-induced PGE 2 and GRO ⁄ CINC-1 by H4-v.1 and HN gene products Table 1. Analysis of 3¢UTR palindromic and purine-rich sequences a (A) and upstream histone H4-like site II regulatory domain b (B) in rat H4g (m27433), H4s (x13554), H4-v.1 (AY936209) and HN (AY936210). Sequence homologies were determined as indicated in Experimental pro- cedures. Gene Sequence characteristics Poly(A) signal c Cap site A H4g 35 GGCCCTTTTCAGGGCCACCCACGAACTCATTCAAAGGG 72 None H4s 56 GGCCCTTTTCAGGGCCCCCAAACTATCCAAAAGGAG 91 None H4-v.1 None AATAAA HN 55 CCACACCATCAGGCTGTGGATACATAGATAAGGCAACATGG 95 TATAAA B H4g –66 CGCCTGTGGTCTTCAATCAGGTCCGCAGAAGGTCTATTTAAA )25 *CTTTT H4s –63 TCCCTGCTGTTTTCAAACAGGTCCGCTCCCAGGAAATATAAGC )21 *CTGTA H4-v.1 )80 CGCTCCCTGTTTTCACTCCGGTCCGCAAGTTCCATATAAGA )40 *GAGCA HN )72 CACTTGAAGTTCTCAACCAGGTCCGATAAGAGTGTATACTT -34 *TGGAA a Underlined sequences represent consensus stem–loop sequences for histone H4, the underlined bold letters indicate a noncanonical stem–loop structure, and bold italic letters indicate purine-rich sequences. Superscript positive numbers indicate the position downstream of the stop codon. TAAA repeat elements [49] are also present in H4-v.1 and HN 3¢UTRs. b TATA-box sequences are indicated in bold letters, histone H4 subtype-specific GGTCCG elements are underlined and in bold letters, and interferon regulatory factor recognition motifs are underlined. Superscript negative numbers indicate the position upstream of the cap site (*). c The polyadenylation signals were preceded by an auxiliary mRNA-processing facilitator-like element (AAAGAT). R. Poirier et al. Histone H4 mRNA variants FEBS Journal 273 (2006) 4360–4373 ª 2006 The Authors Journal compilation ª 2006 FEBS 4365 is related to possible cytotoxic effects on rat AMs, the percentage of living cells was determined on the basis of the cytoplasmic esterase conversion of calcein-AM to the green fluorescent product calcein by living cells. Exposure to 10 )8 m H4-(86–100) or [Thr89]-H4-(86– 100) or to 20 lL of the transcription ⁄ translation prod- ucts of either H4-v.1pCMVTnT, HNpCMVTnT or control pCMVTnT for up to 24 h had no effect on the percentage of viable green fluorescent cells, indicating no loss of AM membrane integrity (Fig. 6). A B Fig. 4. Electrophoretic gel separation of coupled transcription ⁄ trans- lation (TnT) products with H4-v.1 and HN cDNA constructs (A) and inhibition of LPS-induced PGE 2 release (B). (A) Biosynthesis experi- ments with Promega luciferase control DNA plasmid, HN-pCMV- TnT, H4-v.1-pCMVTnT and emptied pCMVTnT plasmids were performed, and the radioactive products were separated by gel electrophoresis as described in Experimental procedures. Arrows show the molecular mass of [ 35 S]Met-labeled protein and peptide products as determined by comparison with the electrophoretic pattern of a Mark 12 molecular weight ladder. (B) Biosynthetic products (20 lL of reaction samples) obtained in parallel experi- ments with unlabeled Met were incubated with primary cultures of rat AMs as described in Experimental procedures, and their ability to inhibit LPS-evoked PGE 2 secretion was compared with those of synthetic H4-(86–100) and [Thr89]H4-(86–100) at 10 )8 M. *P 6 0.05 is considered significant as compared with control. Fig. 5. Inhibition of LPS-induced chemokine secretion. (A) Rat AMs were stimulated for 4 h with LPS (1 lgÆmL )1 ) in the presence and absence of synthetic [Thr89] H4-(86–100)) and H4-(86–100) at 10 )8 M as described in Experimental procedures, and secretion of IL-8 (GRO ⁄ CINC-1) was measured in culture supernatants. Results are means ± SEM of three experiments (*significantly different from control at P 6 0.05). Fig. 6. The percentage of live ⁄ dead cells following treatment of AMs with TnT products of H4-v.1 and HN cDNA constructs (20 lL reaction samples) and corresponding synthetic peptide products (10 )8 M) was determined as described in Experimental procedures by assessing the number of living cells, which take up calcein and convert it to F-calcein (green fluorescence), and dead cells, which take up ethidium bromide homodimer (red fluorescence). Histone H4 mRNA variants R. Poirier et al. 4366 FEBS Journal 273 (2006) 4360–4373 ª 2006 The Authors Journal compilation ª 2006 FEBS Discussion Each class of histones contains its gene variants. Bovine H4-v.1, the first reported example of a histone H4 mRNA variant in mammals [41], contains the palin- dromic and purine-rich sequences typical of cell cycle- dependent histone mRNAs with a 1.3 kb downstream extension that terminates with a polyadenylated track characteristically found in cell cycle-independent his- tone mRNAs. The present results indicate that rat H4- v.1 cDNA differs somewhat from bovine H4-v.1 cDNA by the absence of the consensus palindromic and pur- ine-rich sequences and the excision of an intron, two characteristics of the Drosophila replication-independ- ent histone H4 cDNA [54]. Yet, like bovine H4-v.1 and all other histone cDNA variants, rat H4-v.1 cDNA contains a 3¢UTR extension that terminates with a polyadenylation signal (AATAAA). On the other hand, HN cDNA, the second polyadenylated histone H4-rela- ted cDNA observed in rat, contains noncanonical pal- indromic and purine-rich sequences in its proximal 3¢UTR and a noncanonical TATAAA polyadenylation signal [55] in its distal 3¢UTR (Table 1; Fig. 2). Inter- estingly, whereas the presence or absence of intronic and palindromic sequences varies among subtypes of replication-independent histone mRNAs [56], the pres- ence of a polyadenylation signal is typical of all replica- tion-independent histone mRNAs [2], suggesting that the expression of both H4-v.1 and HN genes may be independent of the cell cycle. Even though the promoter regions of core histone H4 genes are evolutionarily divergent among verteb- rate species [47], they all contain an upstream region named site II, consisting of the cell cycle control ele- ment, H4 gene subtype element, and TATA box [45,46]. The site II region is considered to play a key role in the cell cycle dependency of expression [46,57]. On the other hand, histone gene variants are not expected to be regulated by the same factors that regulate the expression of core histone genes, because variant transcripts accumulate preferentially in nondi- viding and terminally differentiated cells [3]. Analysis of genomic sequences upstream of the H4-v.1 and HN genes indicated that both sequences contain a region similar to the histone H4 site II cell cycle regu- latory domain (Table 1) [46]. This finding suggests that this region may not be the sole determinant for cell cycle dependency of histone H4 gene expression. Further analyses of the upstream genome region proximal of the H4-v.1 and HN genes indicated the presence of specific transcription regulator-binding sequences that were not present upstream of either somatic or germinal core histone H4 genes (Fig. 2). Among these specific regulatory factor-binding motifs, the CREB and NF-jB sites may play an important role in the tissue-specific expression of H4-v.1 and HN mRNAs, respectively. In this regard, marked release of immunoreactive HN from perfused bovine adrenal glands has been observed when the glands are stimulated with carbamylcholine [39]. As carbachol was shown to be a potent activator of NF-jB in iso- lated canine gastric parietal cells [58], we may hypo- thesize that the NF-jB-binding element located 167 nucleotides upstream of the rat HN cDNA cap site has a role to play in the production and release of HN. Interestingly, in contrast with granule prestored enkephalins and catecholamines, which were suc- cinctly and rapidly released after carbamylcholine sti- mulation, the release of HN started only 30 min after the beginning of the stimulation and lasted for more than 1 h, thus allowing sufficient time for transcrip- tion factor activation, HN mRNA formation and translation [39]. Post-transcriptional control of cell cycle-dependent histone mRNAs is monitored by the stem–loop struc- ture present within their 3¢UTRs [20]. H4-v.1 mRNA does not contain a stem–loop sequence, whereas HN possesses a stem–loop sequence that differs from the histone stem–loop consensus sequence observed in H4g or H4s (Table 1). However, as both the H4-v.1 and HN genes are polyadenylated, their post-transcrip- tional maturation and processing are expected to be regulated like those of polyadenylated mRNAs [47]. Interestingly, the 3¢UTRs in H4-v.1 and HN mRNAs also contain TAAA repeat elements (14 and 4 repeats, respectively) that are known to be involved in the post-transcriptional regulation of IL-2 in lymphoid cells [49]. The presence of this repeat element in H4- v.1 and HN mRNAs may explain the particularly high abundance of these genes in lymphoid tissues (Fig. 1) along with OGP and HN [39,42]. In this regard, H4- v.1 and HN cDNA transcripts, as well as C-terminal H4 and HN related-peptides, were shown to be present in nonreplicating and terminally differentiated rat AMs (Fig. 1B) [42]. Synthesis, storage, processing and release of C-terminal histone H4 and related peptides were suggested to follow the same route as cytokines [42], which are stored in microvesicles and processed and released via a lysosomal pathway [59]. The distinct induction of the expression of the H4-v.1 and HN genes by the immunostimulants LPS and IFN-c in AMs suggests that the two genes may have distinct and ⁄ or complementary functions in response to immu- nostimulants, whereas the noninduction of core H4s by the same agents concurs with its known cell replica- tion dependency (Fig. 1B). R. Poirier et al. Histone H4 mRNA variants FEBS Journal 273 (2006) 4360–4373 ª 2006 The Authors Journal compilation ª 2006 FEBS 4367 Examination of the ORF of rat H4-v.1, H4s and H4g cDNAs revealed the presence of two ATG initi- ation codons that allow their translation into whole histone H4 and the alternative C-terminal fragment H4-(84–102) (Fig. 3). Interestingly, the 5¢UTR of rat H4.v-1 cDNA (19 nucleotides; Fig. 2) was much shor- ter than those of H4s and H4g cDNAs (> 100 nucleo- tides). The short 5¢UTR in rat H4-v.1 mRNA may enhance leaky ribosomal scanning, as the first ATG codon could be too close to the 5¢ end to be recog- nized efficiently [60]. Such a possibility is supported by a previous observation indicating that LPS stimulation of the expression of H4-v.1 mRNA in rat AMs is accompanied by an increase in the cell contents of the short peptides OGP and H4-(86–100), but not of whole histone H4 or total H4 mRNA [42]. The in vitro biosynthesis experiments with the H4-v.1 cDNA con- struct also indicate that at least part of the first ATG codon may be skipped to produce the C-terminal pep- tide H4-(84–102) (Fig. 4A). The relatively high level of production of complete histone H4 as compared with H4-(84–102) by the H4-v.1 cDNA construct may be due to the elongation of the 5¢UTR in the cDNA con- struct by 5¢RACE Marathon adapter and 5¢ b-globin leader sequences in the pCMVTnT vector. On the other hand, Bab et al. [61] used a histone H4-CAT reporter fused cDNA vector engineered to produce a polyadenylated histone H4 mRNA. The recombinant construct produced different ratios of whole histone H4-CAT and H4-(84–102)-CAT, depending upon the cell type in which the vector was expressed. Further investigation is required to clarify whether, in the in vivo situation, rat H4-v.1 synthesizes both whole his- tone H4 and H4-(84–102) or mainly H4-(84–102), as suggested by our previous experiments with LPS-sti- mulated rat AMs [42]. Use of the coupled transcription ⁄ translation system with the HN pCMVTnT construct produced a single radiolabeled compound with a molecular mass corres- ponding to that of the expected translation product [Thr89]H4-(84–102). The HN cDNA ORF has the necessary translational start and stop codons to pro- duce this modified H4 C-terminal peptide, but not the first start codon necessary to produce total histone H4 (Figs 2B, 3). The ability of HN mRNA to translate a small peptide allows the messenger to be considered as a minigene. Minigenes are well recognized for their role in the regulation of gene expression [62–65]. With- out including the small interfering RNAs (siRNAs), which affect gene expression but cannot be considered as true genes, due to their lack of an ORF, there exist at least two types of minigene that can affect transcrip- tional or post-transcriptional gene expressions. For instance, Tenson et al. [65] reported that the transla- tion of a minigene with an ORF coding for a peptide of eight amino acids or fewer inhibits protein synthesis by a phenomenon of ‘dropping off’ of the peptide from ribosomes under a form that is still attached to the tRNA corresponding to its C-terminal amino acid, thus creating a shortage of this tRNA for translation. Other minigenes selectively inhibit the translation of the functional downstream cistron [63,64]. As the bio- synthetic products of H4-v.1 and HN cDNA constructs display anti-inflammatory effects in isolated AMs com- parable to those of the synthetic H4 C-terminal pep- tides H4-(86–100) and [Thr89]H4-(86–100) (Figs 4B, 5), the question of whether the expression of these H4 mRNA variants can affect some transcriptional or post-transcriptional gene regulatory mechanisms, in addition to producing the extracellularly acting anti- inflammatory peptides, remains to be investigated. In conclusion, a growing body of evidence indicates that various histones or histone-derived products act in an extranuclear and ⁄ or extracellular manner. Such examples include the histones H2A and H2B, which display growth hormone- and prolactin-releasing activity [20–23], and the antimicrobial histone-H2A peptide buforin I, produced by the action of pepsin within gastric gland cells of the vertebrate stomach [28,29]. In addition, the histone H4-derived peptides HN and OGP have been shown to display antinoci- ceptive and osteogenic activities, respectively [32–37]; whereas synthetic C-terminal histone H4 peptides were reported to serve as potent epitopes for antigen- presenting cells in in vitro models of T-lymphocyte activation [25]. In the present study, we further dem- onstrate that the C-terminal histone H4-related pep- tides transcribed from H4-v.1 and HN genes significantly inhibit the LPS-evoked release of PGE 2 and IL-8 (GRO ⁄ CINC-1), two potent proinflammato- ry mediators produced by activated macrophages. The particular interest in the effects of H4-(86–100) and [Thr89]H4-(86–100) (or rat HN) on AM GRO ⁄ CINC- 1 secretion derives from the knowledge that IL-8 pro- duction represents one of the primary responses of macrophages to inflammation, and that such an effect lasts as long as inflammation persists [69]. IL-8 not only serves to attract inflammatory cells to a site of inflammation and keep them there, but also stimulates neutrophils to a higher activation state. Its release from macrophages is evoked by LPS and cytokines such as IL-1, and its high plasma level is associated with various human inflammatory diseases [68,69]. Therefore, the presence of the H4-v.1 and HN genes in tissues (thymus, bone marrow) wherein immune reactions are known to take place and the potent Histone H4 mRNA variants R. Poirier et al. 4368 FEBS Journal 273 (2006) 4360–4373 ª 2006 The Authors Journal compilation ª 2006 FEBS inhibitory effects of their translation products on macrophage proinflammatory functions suggest that histone H4 mRNA variants may have an important role in the physiology and ⁄ or physiopathology of inflammation. Experimental procedures Computer-assisted analysis of genomic sequence The National Center for Biotechnology Information (NCBI; http://www.ncbi.nlm.nih.gov/BLAST/) and The Institute for Genomic Research (TIgr; http://www.tigt.org/ tdb/tgi/) blast programs were used to gather information regarding possible H4-related sequences carrying a down- stream polyadenylation signal or coding for modified H4 proteins. GSPs were made with the aid of the primer3: www primer tool (http://biotools.umassmed.edu/bioapps/ primer3_http://www.cgi). Computer-assisted analysis of potential upstream transcription factor-binding sites of the H4-v.1 and HN genes was done using the matinspector program (http://www.genomatix.de/cgi-bin/./eldorado/ main.pl). Comparison of upstream sequences with homol- ogy to the site II cell cycle regulatory domain of vertebrate H4 genes was done using the lalign program (http:// www2.igh.cnrs.fr/bin/lalign-guess.cgi). Analysis of potential palindromic sequences within the 3¢UTRs of the mRNAs was done with the aid of the mfold program (http:// biotools.idtdna.com/Analyzer/). Real-time RT-PCR Reverse transcription was performed on 250 ng of rat (Sprague-Dawley) tissue total RNA preparations (RNeasy Mini Kit for total RNA isolation; Qiagen Mississauga, Canada), pretreated with amplification grade DNase 1 (Invitrogen, Burlington, Canada), using database-deduced gene-specific H4-v.1 (5¢-ccagggttttgtttgtttttg-3¢), HN (5¢-ca cagcctgatggtgtggattggtg-3¢) and GAPDH (5¢-aggtcaat gaaggggtcgttg-3¢) antisense primers and 4 U of omniscript reverse transcriptase (Qiagen) (where U is defined as enzyme activity which incorporates 1 nmol TTP into acid- insoluble products in 10 min at 37 °C with poly A template RNA and oligo-dT 12–18 primer). Real-time PCR was per- formed using a standard Quantitect TM SYBR R Green PCR kit (Qiagen) protocol on an Applied Biosystems (Foster City, CA) 7900HT Sequence Detection System. PCR amplifications (40 cycles) were performed using designed rat H4-v.1 (sense, 5¢-ggcggctaagaaacaaagtg-3¢; antisense, 5¢-gaaaagttgggtggaagcaa-3¢) or rat HN (sense, 5¢-gccat ggatgtggtctatact-3¢; antisense, 5¢-gccgaagccatagagagtg-3¢) primers and the QuantiTect SYBR Green PCR Master Mix (Qiagen). Validation was done with GAPDH using rat-spe- cific GAPDH (sense, 5¢-aatggtgaaggtcggtgtgaac-3¢; anti- sense, 5¢-aggtcaatgaaggggtcgttg-3¢) primers. The relative quantification of mRNA transcripts was carried out by the comparative Ct (cycle threshold) method, the theoretical basis of which has previously been described in detail [66]. Amplicons were cloned into the pCR 4-TOPO vector 2.0 using the TOPO TA cloning kit for sequencing (Invitro- gen), transformed, plated as outlined within Invitrogen’s TOPO TA kit manual, and sequenced with a DNA seq- uencer (AIB automatic sequencer; Biotechnology Research Institute, BMI Department, University of Ottawa). 5¢RACE and 3¢RACE of rat H4-v.1 and HN mRNAs in a rat spleen cDNA library Full-length rat (Sprague-Dawley) H4-v.1 and HN cDNAs were obtained using 0.5 ng of a Marathon-Ready TM spleen cDNA library (BD Biosciences Clontech, Paolo Alto, CA) by two successive rounds of PCR with outside and nested gene-specific H4-v.1 and HN primers and the adapter sequence-specific amplification primers (APs) supplied with the Marathon-Ready spleen cDNA kit (AP1 and nested AP2 primers). The initial H4-v.1 5¢RACE and 3¢RACE reaction round used an H4-v.1 5¢-GSP (5¢-H4-v.1 GSP1: 5¢-tatagacatgcctgtagtatctgaacc-3¢) coupled with the adapter primer AP1 (5¢-ccatcctaatacgactcactatagggc-3¢), and an H4- v.1 3¢-GSP (3¢-H4-v.1 GSP1: 5¢-ctacacggagcacgccaag-3¢) coupled with AP1. The initial HN 5¢RACE and 3¢RACE reaction round used an HN 5¢-GSP (5¢-HN GSP1: 5 ¢-aga ggtcctgagttcaattgct-3¢) coupled with AP1, and an HN 3¢- GSP (3¢-HN GSP1: 5¢-ctaagcgcccaccgcaaagtcttg-3¢) coupled with AP1. First-round RACE PCR reactions included 2.5 U (where U is defined as enzyme activity which incor- porates 10 nmol dNTPs into acid-insoluble amplicon in 30 min at 72 °C) of pfuUltra Hotstart DNA polymerase (Stratagene, La Jolla, CA) and were performed in accord- ance with Stratagene’s PCR protocol, using 30 cycles of amplification. Nested 5¢-H4-v.1 or 3¢-H4-v.1 RACE PCR reactions were conducted using a 1 : 100 dilution of the first PCR reactions with either the H4-v.1 5¢-nested GSP (5¢-H4-v.1 GSP2: 5¢-ccagggttttgtttgtttttg-3¢) coupled with AP2 (5¢-actcactatagggctcgagcggc-3¢), or the H4-v.1 3¢-nested GSP (3¢-H4-v.1 GSP2: 5¢-ccaagactaataaaataaacctgaagg-3¢) coupled with AP2. Nested 5¢ and 3¢ HN RACE reactions were conducted as above, with an HN 5¢-nested GSP (5¢- HN GSP2: 5¢-tggcgcttgagagtatagacc-3¢ ) coupled with AP2, and an HN 3¢-nested GSP (3¢-HN GSP2: 5¢-ggatgtggtcta- tactctcaagc-3¢) coupled with AP2. The second-round nested RACE PCR reaction included 2.5 U of HotstarTaq DNA polymerase (Qiagen), and was performed in accordance with the manufacturer’s PCR protocol, with 25 cycles of amplification. Amplicons were cloned into the pCR 4- TOPO vector 2.0 using the TOPO TA cloning kit for sequencing (Invitrogen), transformed and plated as outlined in Invitrogen’s TOPO TA kit manual, and sequenced. R. Poirier et al. Histone H4 mRNA variants FEBS Journal 273 (2006) 4360–4373 ª 2006 The Authors Journal compilation ª 2006 FEBS 4369 [...]... expression of FEBS Journal 273 (2006) 4360–4373 ª 2006 The Authors Journal compilation ª 2006 FEBS R Poirier et al 43 44 45 46 47 48 49 50 51 52 53 54 55 the histone H4 mRNA variant H4- v.1 and the levels of histone H4- (86–100) and H4- (89–102) (OGP) in various rat tissues and alveolar macrophages Peptides 26, 1503– 1511 Franklin SG & Zweidler A (1977) Non-allelic variants of histones 2A, 2B, and 3 in... instructions IL-8 (GRO ⁄ CINC-1) secretion from rat AMs For the evaluation of the effects of synthetic HN- and H4related peptides on IL-8, the prototype of CXC chemokines, AMs (0.2 · 106) were cultured in 0.2 mL of tissue culture medium for 4 h with and without LPS in the presence and absence of various concentrations of H4- and HN-related peptides as indicated Release of rat IL-8, i.e growth-related oncogene.. .Histone H4 mRNA variants R Poirier et al Insertion of H4- v.1 and HN into pCMvTnT vector H4- v.1 and HN amplicons containing the Marathon adapter sequence and their respective 5¢UTR segments, ORFs and a small amount of their respective 3¢UTRs were created following the second-round nested RACE PCR method as described above, including either the 5¢ -H4- v.1 GSP2 (5¢-ccagggttttgtttgtttttg-3¢)... and characterization of the Drosophila histone H4 replacement gene FEBS Lett 388, 219–222 Silver Key SC & Pagano JS (1997) A noncanonical poly(A) signal, UAUAAA, and flanking elements in Epstein–Barr virus DNA polymerase mRNA function in cleavage and polyadenylation assays Virology 234, 147–159 Histone H4 mRNA variants 56 Alvelo-Ceron D, Niu L & Collart DG (2000) Growth regulation of human variant histone. .. given standard laboratory chow and water ad libitum, and were used within 2 weeks Approval was obtained from the Animal Care and Use Committee of the University of Ottawa for all procedures Isolation of rat AMs and gene transcript measurements AMs were recovered from normal rats by bronchoalveolar lavage (BAL) as described previously [67] Briefly, the lungs were lavaged with seven 7 mL aliquots of sterile... for his judicious advice in the design and engineering of gene-specific cDNA constructs for in vitro transcription ⁄ translation experi- Histone H4 mRNA variants ments This work was supported by the Canadian Institutes of Health Research (CIHR) References 1 Zhao J (2004) Coordination of DNA synthesis and histone gene expression during normal cell cycle progression and after DNA damage Cell Cycle 3, e112–e114... chamber, and viability (98–100%) was determined by trypan blue exclusion Differential analysis of lavage cells made by cytocentrifuge smears (Shandon, Pittsburgh, PA; 2.5 · 104 cells) and stained with Wright-Giemsa indicated that the BAL cell population was essentially composed of AMs (99%) in normal rats For measurement of H4- v.1 and HN gene transcripts, AMs (1 · 106 cells) were incubated in 1 mL of tissue... (5¢-tat gtatccacagcctgatggtg-3¢) The H4- v.1 or HN amplicons were further cloned into the pCR 4-TOPO vector 2.0 and transformed as described above Six micrograms of the HN or H4- v.1 vectors were digested with EcoR1 (Invitrogen) and run on a 1.5% agarose gel, and the H4- v.1 or HN segments, including EcoR1 arms, were isolated using a gel extraction kit (Qiagen) The H4- v.1-EcoR1 and HN-EcoR1 segments were ligated... Differential expression of human replacement and cell cycle dependent H3 histone genes Gene 312, 135–143 Pauli U, Chrysogelos S, Stein G, Stein J & Nick H (1987) Protein–DNA interactions in vivo upstream of a cell cycle-regulated human H4 histone gene Science 236, 1308–1311 Xie R, van Wijnen AJ, van Der Meijden C, Luong MX, Stein JL & Stein GS (2001) The cell cycle control element of histone H4 gene transcription... Maltais LJ (2002) The human and mouse replication-dependent histone genes Genomics 80, 487–497 3 Schumperli D (1986) Cell-cycle regulation of histone gene expression Cell 45, 471–472 4 Wu RS & Bonner WM (1981) Separation of basal histone synthesis from S-phase histone synthesis in dividing cells Cell 27, 321–330 ´ 5 Dryhurst D, Thambirajah AA & Ausio J (2004) New twists on H2A.Z: a histone variant with a . Characterization, localization and possible anti-inflammatory function of rat histone H4 mRNA variants Rene ´ Poirier, Irma Lemaire and Simon. of the H4 mRNA variants. The results con- firm the existence and illustrate the structures of H4- v.1 and HN genes, two polyadenylated histone H4 mRNA variants