GTC
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE ADVANCED SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Genes to Cells (2007) 12, 105-117. doi:10.1111/j.1365-2443.2006.01035.x
© 2007 Blackwell Publishing or its licensors
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Gotoh, I.
Right arrow Articles by Seiki, M.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Gotoh, I.
Right arrow Articles by Seiki, M.

Regulated nucleo-cytoplasmic shuttling of human aci-reductone dioxygenase (hADI1) and its potential role in mRNA processing

Isamu Gotoh, Takamasa Uekitaa and Motoharu Seiki*

Division of Cancer Cell Research, Institute of Medical Science, University of Tokyo, Minato-ku, Tokyo, 108-8639, Japan


   Abstract
Top
 Abstract
Introduction
 Results
 Discussion
 Experimental procedures
 References
 
Bacterial aci-reductone dioxygenase (ARD), a member of the cupin superfamily, has evolutionarily primitive protein folding and functions in the methionine recycling pathway. Recently, a human ARD orthologue (human ADI1, hADI1) has been identified and exhibits functions other than ARD activity. The hADI1 localizes mainly to the cytoplasm, but a substantial fraction is nuclear, suggesting functions in both cellular compartments. In this study, we report that nucleo-cytoplasmic transport of hADI1 is regulated by a non-canonical nuclear export signal (NES) located in the N-terminal region of hADI1. The NES is composed of multiple basic amino-acid residues instead of the canonical leucine-rich sequence. Nuclear export of hADI1 was not mediated by CRM1, a major transporter that binds to leucine-rich NES. Substitution of the basic residues with alanines abolished NES activity. Mutant hADI1 accumulated in the nucleus and formed speckles frequently observed with splicing factors and some transcription factors. Indeed, hADI1 specifically co-localized with the splicing factor U1-70K to the nucleus but not with another splicing factor, SC35. U1-70K over-expression induced nuclear accumulation of hADI1. Nuclear hADI1 expression significantly altered the splicing pattern of the adenovirus E1A mini-gene, which generates multiple alternatively spliced transcripts. Thus, hADI1 may have acquired a novel role in nuclear mRNA processing possibly by modulating U1-70K-related functions, an activity negatively regulated by a non-classical NES sequence.


   Introduction
Top
 Abstract
 Introduction
Results
 Discussion
 Experimental procedures
 References
 
The cupin domain, a small thermostable protein structure called the ß-barrel fold, is composed of two conserved ß-strands separated by a variable loop region (Dunwell et al. 2001, 2004). Proteins exhibiting this domain are present in archaea, bacteria and eukarya, and form a large cupin superfamily. The cupin domain exhibits conserved structures but has variable biochemical activity, which is likely due to differences in the amino acid composition of the variable loop. Major members are metalloenzymes, including dioxygenases, hydroxylases, decarboxylases, super oxide dismutases and isomerases. Others have not yet been proved to have enzymatic activity, and some of them have lost metal-binding ligands during evolution. These include: (i) small compound binding regulatory proteins such as auxin binding protein (ABP), sucrose binding protein (SBP) and AraC transcription factor, (ii) structural proteins such as seed storage globulins and centromeric protein CENP-C, and (iii) jumonji C (JmjC)-domain-containing transcription factors. The JmjC domain is a cupin-like structure (Clissold & Ponting 2001) recently suggested to have histone demethylase activity and a function in chromatin remodeling (Tsukada et al. 2006).

Previously, we isolated by yeast two-hybrid screening a cupin protein that binds the cytoplasmic tail of membrane-type 1 matrix metalloproteinase (MT1-MMP) and named it MT1-MMP cytoplasmic tail binding protein 1 (MTCBP1) (Uekita et al. 2004). MTCBP1 is a ubiquitous protein composed of 179 amino acids. Based on its protease activity, MT1-MMP is a potent pericellular modulator that regulates cell migration, invasion and proliferation (Itoh & Seiki 2006). Over-expression of MTCBP1 inhibits MT1-MMP-dependent cell migration and invasion (Uekita et al. 2004). Expression of MTCBP1 is down-regulated in many human tumor cell lines, and it may therefore have an anti-tumor activity by inhibiting MT1-MMP activity. Among cupin family proteins, MTCBP1 shows significant homology to aci-reductone dioxygenase (ARD), a Klebsiella pneumoniae gene product (Wray & Abeles 1995; Pochapsky et al. 2002) that functions in a methionine recycling pathway known as the 5'-methylthioadenosine (MTA) pathway (Sekowska et al. 2004). Indeed, MTCBP1 was demonstrated to have ARD activity in yeast (Hirano et al. 2005), and disruption of the orthologous yeast gene (yARD) abrogated cells’ ability to use the MTA as a sulfur source for cell growth. Re-expression of yARD or MTCBP1 in the mutant cells rescued MTA-dependent cell growth, indicating that MTCBP1 and yARD can function in that pathway. Based on this result, we renamed MTCBP1 human ARD (hARD). However, HUGO (the Human Genome Organization) Gene Nomenclature Committee recently designated eukaryotic ARD as ADI1 (aci-reductone dioxygenase 1), and we use this name for hARD/MTCBP-1.

Recently, homologues of bacterial ARD exhibiting multiple functions have been identified in eukaryotic species (Yeh et al. 2001; Oram et al. 2004; Lin et al. 2005; Sauter et al. 2005). Yeh et al. (2001) isolated a 5'-truncated complementary DNA (cDNA) of human ADI1 (hADI1) as Sip-L (submergence-induced protein-like). Sip-L over-expression allowed replication of the hepatitis C virus (HCV) in otherwise non-permissive HEK293 cells, although the mechanism remains unknown. Oram et al. (2004) isolated a rat homologue called ALP1 (ARD-like protein 1), which is highly expressed in prostate and induced by androgen. ALP1 expression is down-regulated in rat prostate tumors, as is that of hADI1 in human cancer cell lines. Thus, ADI1 appears to have acquired multiple functions during evolution, presumably employing the flexible interface of the cupin domain. In addition, eukaryotic ADI1s may have nuclear roles, because we observed in a previous study that yeast ADI1 (yADI1) and hADI1 are both nuclear and cytoplasmic (Uekita et al. 2004; Hirano et al. 2005).

In this study, we identify a novel sequence element functioning as a nuclear export signal (NES) in the N-terminal region of hADI1. This NES is rich in basic residues, differing from the classical leucine-rich NES. Consistent with this difference, Leptomycin B, an inhibitor of the major nuclear export receptor CRM1, did not inhibit nuclear export mediated by the NES, suggesting that hADI1 is excluded from the nucleus by a novel export pathway. Disruption of NES activity by mutations caused nuclear accumulation of hADI1. In the nucleus, hADI1 co-localized with an mRNA splicing factor U1-70K, and U1-70K over-expression induced nuclear accumulation of hADI1. Nuclear expression of hADI1 altered the splicing pattern of mRNA transcribed from an adenovirus-derived test gene, and this activity did not require hADI1 enzymatic activity. These results suggest that hADI1 modulates mRNA processing directly or indirectly. This activity appears to be negatively regulated by nuclear export via a novel NES. The apparent multifunctional nature of eukaryotic ADI1 proteins raises the possibility that ARD/ADI1 is an appropriate protein model for structure–function analysis of the cupin domain.


   Results
Top
 Abstract
 Introduction
 Results
Discussion
 Experimental procedures
 References
 
Identification of a hADI1 sequence exhibiting NES activity

We previously observed that yeast and human ADI1 proteins are both cytoplasmic and nuclear (Fig. 1A) (Uekita et al. 2004; Hirano et al. 2005). In a previous study, we also found that deletion of the N-terminal region of yADI1 altered localization such that it was primarily nuclear (Hirano et al. 2005). Thus, putative nuclear exclusion signals likely reside in the deleted part. Initial efforts to identify such signals using a deletion strategy were unsuccessful, possibly due to protein misfolding. However, we observed sequences fitting the consensus of a leucine-rich NES (Fig. 1B, seq. 1) and a highly basic NLS (Fig. 1B, seq. 2). To determine whether these two sequences mediated nucleo-cytoplasmic transport, synthetic peptides for seq. 1 and seq. 2 were conjugated to ovalbumin (OV) (Fig. 1C, OV), a 45 kDa protein, and microinjected into the nucleus or cytoplasm of cultured cells together with fluorescein isothiocyanate (FITC)-labeled BSA as an injection marker (Fig. 1C, BSA). The seq. 1-conjugate did not show NES or NLS activity in our assay condition (Fig. 1C, left panels). The seq. 2-conjugate also did not show predicted NLS activity (Fig. 1C, right lower panels). However, contrary to expectations, the seq. 2-conjugate injected into the nucleus moved to the cytoplasm, indicating NES activity (Fig. 1C, right upper panels). A peptide corresponding to amino acid 41–61 which contains seq. 2 at its C-terminus retained significant NES activity (Fig. 2A), while a peptide of amino acid 41–51 lacking seq. 2 did not (Fig. 2A), suggesting that the observed NES activity was surely attributed to seq. 2 but not to the artificial sequence at the junction of conjugation with OV.


Figure 1
View larger version (53K):
[in this window]
[in a new window]

  		Figure 1  Identification of a sequence with NES activity in hADI1. (A) Subcellular localization of hADI1. Cos7 and HT1080 cells were transfected with an expression vector encoding hADI1 and cultured for 24 h. Cells were fixed and stained with anti-hADI1 antisera. hADI1 appeared mostly cytoplasmic, although it was also detected in the nucleus in both cell lines. Scale bar is 20 µm. (B) Amino-acid sequence of hADI1 is presented separately depending on the encoding exons indicated at the left. Exon 3 encodes the entire cupin domain, where residues binding to the metal co-factor are denoted by asterisks. Sequence 1 (seq. 1), which matches the consensus of a leucine-rich NES (characteristic hydrophobic residues are marked by asterisks), and Sequence 2 (seq. 2), containing a basic-residue cluster (basic residues marked by asterisks) characteristic of an NLS, is indicated by upper lines. (C) Peptides seq. 1 and seq. 2 were conjugated with ovalbumin (OV) and microinjected into the nucleus (nuc) or cytoplasm (cyto) of Cos7 cells together with the injection marker FITC-BSA (BSA). After 45 min incubation, localization of the conjugates was visualized by immunostaining with anti-ovalbumin antibody. Scale bar is 20 µm.

 

Figure 2
View larger version (50K):
[in this window]
[in a new window]

  		Figure 2  Nuclear export mediated by seq. 2 does not require CRM1. (A) Peptides corresponding to amino-acid 41–61, 41–51 or 52–61 (= seq. 2) of hADI1 were conjugated to ovalbumin (OV) and NES activity was analyzed as in Fig. 1C. Mutant seq. 2 with basic residues (marked by asterisks) changed to alanines (underlined, seq. 2(4A)) was also analyzed. Scale bar is 20 µm. (B) Nuclear export of seq. 2-conjugate was tested at 0 °C or 37 °C for 15 min and 60 min. Scale bar is 20 µm. (C) NES activity of MEK-NES (ALQKKLEELELDE) (Fukuda et al. 1996) or seq. 2-NES was tested in the absence (–) or the presence (+) of LMB added to the culture at 10 ng/mL 30 min before injection. Scale bar is 20 µm.

 
Replacement of the basic residues of seq. 2 with alanine almost completely abolished NES activity (Fig. 2A, seq. 2(4A)). Export of the seq. 2-conjugate from the nucleus at 37 °C was evident 15 min after injection (Fig. 2B). In contrast, the rate of nuclear export was greatly reduced at 0 °C (Fig. 2B), suggesting that it was energy-dependent. In most cases of protein export from the nucleus, CRM1 acts as a major transporter by binding to leucine-rich classical NES sequences (Yoneda 2000; Fried & Kutay 2003). LMB, which binds CRM1, is a specific inhibitor of classical NES-mediated export (Yoneda 2000; Fried & Kutay 2003). Indeed, LMB blocked nuclear export mediated by an NES derived from MAP kinase kinase (MAPKK/MEK), as shown in Fig. 2C (left panels). However, it did not block seq. 2-mediated nuclear export in the same assay conditions (Fig. 2C, right panels). These results clearly indicate that the transporter responsible for the nuclear exclusion mediated by the basic-rich seq. 2 is not CRM1.

To evaluate seq. 2-mediated NES activity in the context of entire protein, four basic residues in hADI1 were mutated to alanine (4A mutant) as shown in Fig. 2A and the mutant protein was expressed in Cos7 cells. Different subcellular localization of mutant and wild-type proteins were not greatly evident by immunostaining (Fig. 3A), but biochemical fractionation showed nuclear accumulation of mutant protein compared to the wild type (Fig. 3B). We also evaluated the NES function of seq. 2 in relation to an exogenous NLS derived from the SV40 virus large-T antigen. Fusion of that NLS-hADI1 (WT-NLS) enhanced nuclear localization of the chimeric protein, although a substantial amount was retained in the cytoplasm (Fig. 3C,D). Mutation of seq. 2 in hADI1-NLS (4A-NLS) largely prevented its cytoplasmic localization (Fig. 3C,D). Thus, seq. 2 counteracts authentic NLS activity, indicating that seq. 2 in ADI1 functions as an NES.


Figure 3
View larger version (53K):
[in this window]
[in a new window]

  		Figure 3  NES activity of seq. 2 in hADI1. (A) Localization of wild-type hADI1 (WT) and seq. 2 mutant (4A) in Cos7 cells. In the 4A mutant, four basic residues of seq. 2 were replaced with alanine as in Fig. 2A. Proteins were expressed in Cos7 cells and stained with anti-hADI1 antisera. Scale bar is 25 µm. (B) Cos7 cells expressing indicated proteins were fractionated as cytoplasmic (C) and nuclear (N) fractions and analyzed by Western blotting using anti-hADI1 antisera, anti-actin or anti-lamin A/C antibodies. Molecular weight in kDa is shown at right. (C) Localization of hADI1-NLS (WT-NLS) and hADI1(4A)-NLS (4A-NLS) in Cos7 cells examined by immunostaining as in (A). Scale bar is 25 µm. (D) Cells in (C) were fractionated and proteins were detected by Western blotting as in (B). (E) Cos7 cells were transfected with the indicated pEGFP-hADI1 constructs (illustrated at the right). Fusion proteins were expressed for 30 h and EGFP signals were observed. Expression of intact fusion proteins was confirmed by Western blotting with anti-GFP antibody (data not shown). Cells expressing EGFP-{Delta}N were treated with LMB at 10 ng/mL for 3 h. Scale bar is 20 µm.

 
As mentioned previously, serial deletions from the N-terminus frequently prevented expression of the mutant ADI1 proteins. However, a mutant having substitution of the N-terminal 51 amino acids with enhanced green fluorescent protein (EGFP) (EGFP-{Delta}N) was expressed at significant levels. To our surprise, the substitution augmented localization of the mutant protein mostly to the cytoplasm (Fig. 3E, EGFP-{Delta}N). Introduction of the 4A mutation of seq. 2 into this construct prevented nuclear exclusion, often showing predominant nuclear accumulation of the mutant protein (Fig. 3E, EGFP-{Delta}N/4A). Fusion of an additional seq. 2 sequence to the N terminus re-directed the protein to the cytoplasm (Fig. 3E, EGFP-{Delta}N/4A+seq.2). LMB did not prevent cytoplasmic localization of EGFP-{Delta}N (Fig. 3E, EGFP-{Delta}N+LMB). These results further confirm that the short linear stretch of seq. 2 is sufficient to function as n NES in hADI1 and that the activity is not affected by its location in the protein. In addition, nuclear accumulation of EGFP-{Delta}N/4A indicates that hADI1 has a specific mechanism to localize to the nucleus. Thus, the subcellular distribution of hADI1 appears to be controlled by at least two cis-acting elements, an NES represented by seq. 2 and an undefined element with nuclear localization activity. The N-terminal 51 amino acids also appear to regulate the nucleo-cytoplasmic transport of ADI1 by an unknown mechanism.

Nuclear co-localization of hADI1 and splicing factor U1-70K

Mutant hADI1 (EGFP-{Delta}N/4A) accumulates in the nucleus in speckle-like structures (Fig. 3E) that are frequently observed with splicing factors and some transcription factors (Lamond & Spector 2003). Interestingly, an interaction between yeast ADI1 and the splicing factor SNP1 has been detected in a high throughput two-hybrid screening (Fromont-Racine et al. 1997). Thus, we asked whether hADI1 interacts with U1-70K, the human SNP1 homologue, using nuclear derivatives of hADI1 (hADI1 (4A)–NLS and EGFP-{Delta}N/4A). These mutant hADI1 proteins accumulated in the nucleus forming clear speckle-like microdomains, and their signals overlapped with U1-70K signals (Fig. 4A). Another splicing factor, SC35, also formed nuclear speckles, but they were clearly different from those of mutant hADI1 protein (Fig. 4A). Thus, hADI1 likely interacts with U1-70K either directly or as a specific component in U1-70K speckles. Speckles of wild-type hADI1 are not evident under our culture conditions, but some weak signals appear to co-localize with U1-70K in the nucleus (Fig. 4B).


Figure 4
View larger version (44K):
[in this window]
[in a new window]

  		Figure 4  Co-localization of nuclear hADI1 with U1-70K. (A) Co-localization of hADI1 derivatives with U1-70K. hADI1(4A)-NLS and EGFP-{Delta}N/4A were expressed in Cos7 cells together with FLAG-tagged U1-70K (U1-70K-F). hADI1(4A)-NLS and U1-70K-F were visualized with anti-hADI1 antisera and anti-FLAG antibody (mouse), respectively. EGFP-{Delta}N/4A was detected directly by the EGFP signal. Cos7 cells expressing hADI1(4A)-NLS were also stained with anti-SC35 ascites. Scale bar is 5 µm. (B) hADI1 and U1-70K-F were expressed in Cos7 cells either as a single protein (left columns) or together (middle columns), and localization of the proteins was visualized by immunostaining. Images in the insets in the middle column are magnified and shown in the right columns with contrast optimized to enable visualization of the partial co-localization of the two proteins (indicated by white arrowheads). Scale bars in the left and right columns are 20 and 5 µm, respectively. (C) hADI1 was expressed in Cos7 cells together with increasing amounts of U1-70K-F (left columns) or FLAG-tagged MTAP-NLS (MTAP-NLS-F, right columns). Cells were fractionated into nuclear (N) and cytoplasmic (C) fractions and proteins detected by Western blot analysis. U1-70K-F, MTAP-NLS-F, hADI1, actin and lamin A/C were detected using specific antibodies.

 
To determine whether interaction between hADI1 and U1-70K was direct, a pull-down assay was performed, but no precipitates containing both hADI1 and U1-70K were obtained (data not shown). Then, we asked whether both proteins partitioned similarly by subcellular fractionation. Cytoplasm and nuclear fractions of cells expressing increasing amounts of U1-70K were prepared and the amount of hADI1 in each fraction was examined by Western blotting (Fig. 4C, left panels). Actin was used as a cytoplasmic marker and lamin A/C as a nuclear marker. In the absence of exogenous U1-70K, hADI1 was primarily cytoplasmic. Exogenous U1-70K increased the amount of hADI1 that accumulated in the nucleus in a dose-dependent manner. The amount of actin in the nuclear fraction was not affected by U1-70K, suggesting that the observed increase is specific to hADI1 and not due to contamination by cytoplasmic proteins. Thus, ADI1 appears to interact with a nuclear complex containing U1-70K in a dose-dependent manner. As a control for over-expression of a nuclear protein, MTA phosphorylase (MTAP), an enzyme in the MTA pathway (Della Ragione et al. 1986), was fused with an NLS derived from SV40 virus large-T antigen (MTAP-NLS) and expressed in Cos7 cells. Nuclear accumulation of ADI1 was not induced by MTAP-NLS expression (Fig. 4C, right panels).

Effect of hADI1 on mRNA processing

Based on its localization with U1-70K-containing speckles, we hypothesized that hADI1 may affect mRNA splicing. Thus, we used an adenovirus E1A mini-gene construct (Hallier et al. 1998; Ohkura et al. 2005) as a reporter to monitor mRNA splicing. The test gene yields five alternatively splicing variants corresponding to original gene products with co-sedimentation constants of 13S, 12S, 11S, 10S and 9S (Fig. 5A). The major forms detected in Cos7 cells by reverse transcriptase polymerase chain reaction (RT-PCR) were 13S and 12S (Fig. 5B). Expression of hnRNP A1, a known regulator of mRNA splicing, increased levels of the 9S form (Fig. 5B) in a dose-dependent manner, to a degree similar to that reported previously (Caceres et al. 1994). hADI1 expression also increased the 9S form, although less potently than hnRNP A1 (Fig. 5B, see hADI1 and hnRNP A1). However, expression of hADI1-NLS localizing in the nucleus (Fig. 5D) enhanced levels of the 9S spliced form (Fig. 5B), while hADI1-NES localizing out of the nucleus (Fig. 5D) did not induce any increase of 9S form at all, suggesting that a nuclear population, rather than a cytoplasmic one, of hADI1 is responsible for the effects observed. These results were quantified in Fig. 5C, which shows a correlation between decreased 13S and increased 9S.


Figure 5
View larger version (37K):
[in this window]
[in a new window]

  		Figure 5  hADI1 induces alternative mRNA splicing in the nucleus. (A) Schematic representation of alternatively spliced variants predicted from the E1A mini-gene construct. PCR primers to amplify reverse-transcribed products are also indicated (RP67; forward primer, E1A2; reverse primer). Open and closed arrowheads indicate splice donor and acceptor sites, respectively. (B) The mini-gene (0.25 µg) was transfected into Cos7 cells together with increasing amounts (0.75, 1.25 and 1.75 µg) of expression plasmids encoding indicated splicing modulators. Thirty hours after transfection, transcripts from the mini-gene were analyzed by RT-PCR using specific primers. A representative image of RT-PCR products analyzed on agarose gels is presented (upper panel). Co-expressed FLAG-tagged proteins are indicated at the top. Positions of each RT-PCR product are indicated at left. The 11S transcript was not visible in this condition. Co-expressed proteins were detected by Western blot with an anti-FLAG antibody (rabbit) to show comparable expression of proteins (lower panels). Molecular weight in kDa is shown at right. Actin is the loading control. (C) Band intensity of each variant on the gels was traced and quantified by NIH image. The percentages of 13S (upper panel) and 9S (lower panel) to all the variants produced (13S, 12S, 10S and 9S) were calculated (% isoform). The mean value of three independent experiments is presented with standard deviations (SDs) and analyzed by Student's t test. (D) Localization of proteins expressed in Cos7 cells was examined by immunostaining using an anti-FLAG antibody (mouse). Scale bar is 25 µm.

 
hADI1 is a dioxygenase acting in the MTA pathway (Wray & Abeles 1995). Glutamic acid (E) in the catalytic cupin domain is required for enzyme activity (Pochapsky et al. 2002; Pochapsky et al. 2006), as its substitution with alanine (A) abolishes that activity (Hirano et al. 2005). To determine whether hADI1 dioxygenase activity is required to modulate mRNA splicing, an E-to-A (EA) mutation was introduced into hADI1-NLS. Both hADI1-NLS and its EA mutant (hADI1(EA)-NLS) enhanced splicing to the 9S form to a similar extent (Fig. 6A,B). Thus, splicing modulation by hADI1 is not mediated by dioxygenase activity. We also tested another metabolic enzyme in the MTA pathway, MTAP. Fusion with an NLS directed MTAP to the nucleus (MTAP-NLS) (Figs 4C and 6C) but did not cause significant changes in the splicing pattern (Fig. 6A,B). Expression and appropriate subcellular localization of expressed proteins were confirmed (Fig. 6A, lower panels and Fig. 6C). Thus, the activity of hADI1 on alternative splicing is independent of dioxygenase activity and the MTA cycle.


Figure 6
View larger version (48K):
[in this window]
[in a new window]

  		Figure 6  Splicing modulator activity of hADI1 is independent of its function in the MTA pathway. (A) Effect of hADI1 mutants on splicing was tested as in Fig. 5 and a representative agarose gel analyzing RT-PCR products is presented (upper panel). Increasing levels of expressed FLAG-tagged proteins are indicated at the top. Positions of each RT-PCR product are indicated at left. Comparable expression of tested proteins was confirmed by Western blot with an anti-FLAG antibody (rabbit) (lower panel). Actin is the loading control. (B) Band intensity of each variant was quantified, and the percentages of 13S (upper panel) and 9S (lower panel) to all the variants produced is presented as the mean (n = 3)+SD. (C) Localization of expressed proteins in Cos7 cells was examined by immunostaining using an anti-FLAG antibody (mouse). Scale bar is 25 µm.

 

   Discussion
Top
 Abstract
 Introduction
 Results
 Discussion
Experimental procedures
 References
 
Regulated nucleo-cytoplasmic transport of hADI1

Although hADI1 is a relatively small protein whose apparent molecular weight is 18 kDa (Figs 3B and 4C), its translocation between the cytoplasm and the nucleus appears to be regulated by specific mechanisms. Firstly, we identified a basic amino-acid stretch in the N-terminal region (seq. 2) as an NES element, whose basic residues were essential for NES activity (Fig. 2A). Mutation of that sequence in hADI1 increased nuclear levels of the mutant protein (Fig. 3B). Structure modeling of hADI1 in silico based on Klebsiella ARD data (Pochapsky et al. 2002; Pochapsky et al. 2006) revealed that seq. 2 forms a single {alpha}-helix at the surface of the protein folding (data not shown) and therefore appears to be suitable as a binding site for a putative receptor protein. However, as expected from its atypical sequence, the transporter responsible for the NES is not CRM1, which exports proteins having a typical leucine-rich NES element (Kutay & Guttinger 2005), because LMB, a specific inhibitor for CRM1, did not block seq. 2-mediated export (Fig. 2C).

To date some CRM1-independent NES sequences have been reported (Lischka et al. 2001; Saporita et al. 2003; Thakurta et al. 2004) but there is no clear consensus among them. However, a recent study by Mingot et al. (2004) reporting RanBP16 as Exportin 7 (Exp7) is interesting in this regard. Exp7 can export p50RhoGAP and 14-3-3{delta} from the nucleus through a CRM1-independent pathway. NES sequences in target proteins were identified as a folding motif that contains a basic amino-acid stretch; however, a short linear sequence was insufficient for high affinity binding to Exp7. Exp7 was also suggested to export several additional cargoes having this new class of NES. Since seq. 2 has a basic amino-acid stretch, it resembles an Exp7-dependent NES although it can minimally function as a short linear sequence. Sequences closely matching seq. 2 are seen in several human proteins such as C/EBPß and HES5, suggesting that they may also function as an NES in these proteins.

Although hADI1 contains an NES sequence, significant levels of hADI1 are found in the nucleus. Deletion of the N-terminal 51 amino acids (hADI1-{Delta}N) prevented nuclear localization of the mutant protein and induced its clear cytoplasmic localization (Fig. 3E), suggesting that the deleted sequence negatively regulates NES function. In addition, the 4A mutation of seq. 2 in hADI1-{Delta}N promoted nuclear accumulation of the mutant protein (Fig. 3E), indicating that a nuclear localizing activity exists within the protein though it is dominantly overcome by the NES activity (Fig. 3E). The nuclear localizing activity of hADI1 seems to be also weakened by the N-terminal 51 amino-acid region because the 4A mutation of seq. 2 in the context of hADI1 did not cause such strong nuclear accumulation as hADI1-{Delta}N/4A showed (Fig. 3A,E). Thus, it is possible to postulate that the N-terminal 51 amino acids functions as an inhibitory domain which primarily blocks active shuttling of hADI1 between the cytoplasm and the nucleus through the nuclear pore, while basal level of NES activity can be detected even in the presence of the region because NES disruption induced partial nuclear accumulation of hADI1 (Fig. 3B).

Possible nuclear function of hADI1

The structure of bacterial ARD protein and its activity in the MTA pathway are conserved from bacteria to mammals (Dunwell et al. 2004). In this study, we demonstrated that this ancient metabolic enzyme also localizes to the nucleus in higher eukaryotes and is concentrated in specific nuclear microdomains. Thus, eukaryotic ADI1 may have evolved novel nuclear functions. As the possible association between yADI1 and SNP1 was suggested by systematic yeast two-hybrid analysis (Fromont-Racine et al. 1997), hADI1 co-localized with U1-70K to the same nuclear speckles. Enforced expression of U1-70K induced nuclear accumulation of hADI1 in a dose-dependent manner (Fig. 4C). Co-localization with U1-70K was specific because hADI1 did not co-localize with speckles containing another splicing factor SC35 while the interaction with U1-70K seems not direct.

U1-70K is a component of U1-snRNP whose RNA component hybridizes to the 5' splicing donor site of pre-mRNA. U1-snRNP forms an early complex of the spliceosome together with U2-snRNP, which binds the branch point in the intron. Finally, after spliceosome maturation, the lariat form of the intron is released. Since hADI1 interacts with U1-70K possibly via an associating protein, it may alter function of the U1-snRNP complex and eventually affect mRNA splicing. Indeed, nuclear over-expression of hADI1 altered the splicing pattern of mRNA transcribed from an adenovirus mini-gene, but another metabolic enzyme in the MTA pathway did not (Figs 5 and 6). The effect of hADI1 on gene expression may be relevant to the previous observation that an N-terminally truncated form of hADI1 was identified as a hepatic factor called Sip-L. Sip-L supports replication of HCV in an otherwise non-permissive cell line (Yeh et al. 2001). The HCV genome is composed of a single-stranded RNA having positive polarity, and viral products are translated as a single polypeptide that is proteolytically processed into smaller peptides (Gale & Foy 2005). Although RNA genome replication of HCV does not require splicing events, nuclear hADI1 may affect host factors required for the replication.

Molecular evolution from bacterial ARD to eukaryotic ADI1

ARD was originally characterized as a bacterial enzyme acting in the MTA pathway to regenerate methionine (Wray & Abeles 1995). ARD mediates two different reactions with a single substrate, depending on the metal co-factor, Fe2+ or Ni2, acting as "one protein/two enzymes" (Dai et al. 1999). Fe-ARD produces a Met precursor (on pathway), while Ni-ARD produces methylthiopropionic acid and carbon monoxide (CO) (off pathway). In eukaryotic cells, ADI1 appears to have acquired a nuclear function such as modulation of mRNA splicing. In vertebrates, ADI1 has further acquired the ability to regulate ECM remodeling by binding the cytoplasmic tail of MT1-MMP (Uekita et al. 2004). To carry out the nuclear function in a regulated manner, ADI1 also appears to have acquired cis elements regulating nucleo-cytoplasmic transport. Thus, several ADI1 functions have evolved in a single protein. Although there are many cupin family proteins, ARD/ADI1 may be an interesting model protein to study how the variability of the cupin domain correlates with novel functions during evolution.


   Experimental procedures
Top
 Abstract
 Introduction
 Results
 Discussion
 Experimental procedures
References
 
cDNA construction

cDNA encoding hADI1 (Uekita et al. 2004) was subcloned into a mammalian expression vector, pSG5 (Stratagene), and a C-terminal FLAG tag was added to each cDNA by PCR. cDNA clones encoding human U1-70K and human MTA phosphorylase (MTAP) were obtained by PCR from a cDNA mixture of HEK293 cells, tagged with FLAG, and subcloned into pSG5. To produce hADI1 EA and 4A mutants, mutagenesis was carried out by PCR amplification of pSG5-hADI1-FLAG with mutagenic primer pairs carrying E94A or K54AR56AR57AR59A mutations of hADI1. hADI1-{Delta}N was also obtained by PCR with a sense primer corresponding the amino terminus of the hADI1 {Delta}1–51 mutant, and the product was subcloned into pEGFP-C3 (Clontech). The addition of "sequence 2" (seq. 2, Fig. 1B) to the N-terminus of hADI1 {Delta}N/4A was carried out by PCR with a sense primer containing a coding sequence for seq. 2, and the product was subcloned into pEGFP-C3. A nuclear localization signal (NLS) from the SV40 virus large-T antigen (PPKKKRKVED) and an NES from RanBP1 (KVAEKLEALSVR) (Richards et al. 1997) were carboxyl-terminally added to each protein by PCR with antisense primers containing their corresponding sequences. pCMV/Taq2B-FLAG hnRNP A1 was kindly provided by Dr Naganari Ohkura at the National Cancer Center Research Institute (Tokyo, Japan).

Cell culture and transfection

Cos7 and HT1080 cells were cultured in Dulbecco's modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and antibiotics. Cells were transfected using Fugene 6 (Roche) according to the manufacturer's instructions. Leptomycin B (LMB) was purchased from Sigma.

Peptide conjugation and microinjection

Peptides were synthesized with a cysteine residue (C) at the amino terminus for coupling with its mercapto group and conjugated with OV as described (Fukuda et al. 1996). Briefly, OV (3 mg/mL) was mixed with sulfosuccinimidyl 4-(N-maleimidomethyl) cyclohexane-1-carboylate (sulfo-SMCC, 5 mg/mL, PIERCE) in phosphate-buffered saline (PBS, pH 8.0), incubated at room temperature for 1 h and charged onto a BioGel-10G column (Bio-Rad) equilibrated with PBS (pH 7.0) to remove excess sulfo-SMCC. Sulfo-SMCC-activated OV was then mixed with synthesized peptides (3 mg/mL) in PBS (pH 7.0), incubated at room temperature for 3 h and charged onto a BioGel-10G column (Bio-Rad) to remove excess peptides. Eluted conjugates were dialyzed against injection buffer (10 mM HEPES, pH 7.3 and 120 mM KCl), and concentrated by ultrafiltration with Ultrafree-15 Biomax-10K filter (Millipore). An aliquot was subjected to sodium dodecyl sulfate-containing polyacrylamido gel electrophoresis (SDS-PAGE) followed by Coomassie Brilliant Blue (CBB) staining. Conjugations were confirmed by an increase in molecular weight relative to OV, and the number of peptides conjugated was estimated to be 9 per OV molecule, on average, based on the apparent molecular weights of the conjugates.

Conjugates (2 mg/mL) were mixed with an injection marker, FITC-labeled bovine serum albumin (FITC-BSA) (1 mg/mL), and the mixtures were filtered through a Millex®-GV filter unit (0.22 µm pore, Millipore). Mixtures were then injected to cells using an injection apparatus (Eppendorf) and the cells were cultured for 45 min. Injected conjugates were visualized by staining with anti-OV rabbit polyclonal antibody (1 µg/mL, ICN), followed by Cy3-conjugated anti-rabbit immunogloblin G (IgG) antibody (1 µg/mL, Jackson ImmunoResearch) (see below).

Cell staining

Cos7 cells (1 x 105/35 mM dish) were transfected with expression vectors, and proteins were expressed for 24 h. Cells were fixed with 3.7% formaldehyde at room temperature for 40 min, washed 3 times with PBS and permeabilized with 0.5% Triton X-100 in PBS for 1 min. Cells were then washed with PBS for 3 min twice and exposed to primary antibodies at room temperature for 1 h. Antibodies used were: anti-hADI1 rabbit polyclonal antisera (1:1000) (Uekita et al. 2004), anti-FLAG mouse monoclonal antibody (M2, 1 µg/mL, Sigma), anti-OV rabbit polyclonal antibody (1 µg/mL, ICN) and anti-SC35 mouse monoclonal ascites (1:2000, Sigma). Cells were then washed with PBS for 5 min 3 times, followed by incubation with secondary antibodies (Cy3 (Jackson ImmunoResearch)- or Alexa488 (Molecular Probe)- labeled anti-mouse or rabbit IgG antibodies, 1 µg/mL for all) at room temperature for 30 min. Cells were washed with PBS for 5 min 3 times, rinsed with Milli Q water and mounted on glass slides with MOWIOL (Calbiochem). Stained cells were observed with x40 or x60 objective lenses using an IX70 inverted fluorescent microscope (Olympus) and analyzed with MetaMorph software (version 4.6, Universal Imaging Corporation).

Subcellular fractionation and preparation of total cell lysates

Cos7 cells (1 x 106 cells / 100 mM dish) were transfected with the expression vectors, and proteins were expressed for 24–48 h. Cells were washed with ice-cold PBS twice and scraped with a rubber policeman in A buffer (10 mM HEPES–KOH, pH 7.9, 10 mM KCl, 0.1 mM ethyleneglycol bis(2-aminoethylether)tetraacetic acid (EGTA), 1 mM dithiothreitol (DTT) and protease inhibitor cocktail set III (1:1000, Roche), 700 µL/dish) (Dignam et al. 1983). Cells were rotated for 15 min at 4 °C, and NP-40 was added to 0.5%. Cells were then centrifuged at 3000 rpm for 5 min at 4 °C (Nikolaev et al. 2003). The resulting supernatants constituted the cytoplasmic fractions. Pellets, designated the nuclear fractions, were washed with A buffer once, combined with 175 µL of 2x Laemmli's sample buffer (25 mM Tris–HCl, pH 6.8, 0.8% SDS, 4% glycerol, 0.0004% bromophenol blue and 10% ß-mercaptoethanol) and boiled for 5 min. For total cell lysates (TCL), 1x Laemmli's sample buffer was added to cells (100 µL/35 mM dish), and cells were scraped with a rubber policeman and boiled for 5 min.

Western blotting

Denatured samples were subjected to SDS-PAGE, transferred to PVDF membranes (Millipore), and incubated with primary antibodies as follows: anti-hADI1 rabbit polyclonal antisera (1:1000) (Uekita et al. 2004), anti-FLAG rabbit polyclonal antibody (1 µg/mL, Sigma), anti-actin mouse monoclonal antibody (1 µg/mL, CHEMICON), anti-lamin A/C mouse monoclonal antibody (1 µg/mL, Santa Cruz Biotechnology), anti-GFP rabbit polyclonal antibody (1 µg/mL, Molecular Probe). Then, after washing with tris-buffered saline (TBS), blots were incubated with horseradish peroxidase (HRP)-conjugated secondary antibodies against rabbit or mouse IgG (1:10 000, Amersham Biosciences), and protein bands were visualized with enhanced chemiluminescence (ECL-plus, Amersham Biosciences).

In vivo splicing assay

Cos7 cells (0.4 x 105/60 mM dish) were transfected with expression vectors together with pCS3-MT-E1A (Hallier et al. 1998; Ohkura et al. 2005). Proteins were expressed for 30 h, and total RNAs were extracted using the RNeasy mini kit (Qiagen). cDNAs were synthesized by superscript II (Invitrogen) and treated with RNase H at 37 °C for 20 min. PCR was carried out with the primer-pair RR67 (5'-GAGCTTGGGCGACCTCA-3') and E1A2 (5'-TCTAGACACAGGTGATGTCG-3') for 35 cycles, and the products subjected to agarose-gel electrophoresis (AGEP) followed by gel staining with ethidium bromide (EtBr). The gel was scanned, band densities were measured using NIH image, and values were analyzed using Excel software (Microsoft).


   Acknowledgements
 
We are grateful to Dr Naganari Ohkura at the National Cancer Center Research Institute (Tokyo, Japan) for providing us with pCMV/Taq2B -FLAG hnRNP A1. This work was supported by a grant-in-aid for Cancer Research and a Grant-in-Aid for Young Scientists from the Ministry of Education, Culture, Sports, Science and Technology of Japan.


   Footnotes
 
Communicated by: Yoshikazu Nakamura

aPresent address: Growth Factor Division, National Cancer Center Research Institute, Chuo-ku, Tokyo, 104-0045, Japan Back

* Correspondence: E-mail: mseiki{at}ims.u-tokyo.ac.jp


   References
Top
 Abstract
 Introduction
 Results
 Discussion
 Experimental procedures
 References
 
Caceres, J.F., Stamm, S., Helfman, D.M. & Krainer, A.R. (1994) Regulation of alternative splicing in vivo by overexpression of antagonistic splicing factors. Science 265, 1706–1709.[Abstract/Free Full Text]

Clissold, P.M. & Ponting, C.P. (2001) JmjC: cupin metalloenzyme-like domains in jumonji, hairless and phospholipase A2beta. Trends Biochem. Sci. 26, 7–9.[CrossRef][Medline]

Dai, Y., Wensink, P.C. & Abeles, R.H. (1999) One protein, two enzymes. J. Biol. Chem. 274, 1193–1195.[Abstract/Free Full Text]

Della Ragione, F., Carteni-Farina, M., Gragnaniello, V., Schettino, M.I. & Zappia, V. (1986) Purification and characterization of 5'-deoxy-5'-methylthioadenosine phosphorylase from human placenta. J. Biol. Chem. 261, 12324–12329.[Abstract/Free Full Text]

Dignam, J.D., Lebovitz, R.M. & Roeder, R.G. (1983) Accurate transcription initiation by RNA polymerase II in a soluble extract from isolated mammalian nuclei. Nucleic Acids Res. 11, 1475–1489.[Abstract/Free Full Text]

Dunwell, J.M., Culham, A., Carter, C.E., Sosa-Aguirre, C.R. & Goodenough, P.W. (2001) Evolution of functional diversity in the cupin superfamily. Trends Biochem. Sci. 26, 740–746.[CrossRef][Medline]

Dunwell, J.M., Purvis, A. & Khuri, S. (2004) Cupins: the most functionally diverse protein superfamily? Phytochemistry 65, 7–17.[CrossRef][Medline]

Fried, H. & Kutay, U. (2003) Nucleocytoplasmic transport: taking an inventory. Cell. Mol. Life Sci. 60, 1659–1688.[CrossRef][Medline]

Fromont-Racine, M., Rain, J.C. & Legrain, P. (1997) Toward a functional analysis of the yeast genome through exhaustive two-hybrid screens. Nat. Genet. 16, 277–282.[CrossRef][Medline]

Fukuda, M., Gotoh, I., Gotoh, Y. & Nishida, E. (1996) Cytoplasmic localization of mitogen-activated protein kinase kinase directed by its NH2-terminal, leucine-rich short amino acid sequence, which acts as a nuclear export signal. J. Biol. Chem. 271, 20024–20028.[Abstract/Free Full Text]

Gale, M., Jr. & Foy, E.M. (2005) Evasion of intracellular host defence by hepatitis C virus. Nature 436, 939–945.[CrossRef][Medline]

Hallier, M., Lerga, A., Barnache, S., Tavitian, A. & Moreau-Gachelin, F. (1998) The transcription factor Spi-1/PU.1 interacts with the potential splicing factor TLS. J. Biol. Chem. 273, 4838–4842.[Abstract/Free Full Text]

Hirano, W., Gotoh, I., Uekita, T. & Seiki, M. (2005) Membrane-type 1 matrix metalloproteinase cytoplasmic tail binding protein-1 (MTCBP-1) acts as an eukaryotic aci-reductone dioxygenase (ARD) in the methionine salvage pathway. Genes Cells 10, 565–574.[Abstract/Free Full Text]

Itoh, Y. & Seiki, M. (2006) MT1-MMP: a potent modifier of pericellular microenvironment. J. Cell. Physiol. 206, 1–8.[CrossRef][Medline]

Kutay, U. & Guttinger, S. (2005) Leucine-rich nuclear-export signals: born to be weak. Trends Cell Biol. 15, 121–124.[CrossRef][Medline]

Lamond, A.I. & Spector, D.L. (2003) Nuclear speckles: a model for nuclear organelles. Nat. Rev. Mol. Cell Biol. 4, 605–612.[CrossRef][Medline]

Lin, T., He, X., Yang, L., Shou, H. & Wu, P. (2005) Identification and characterization of a novel water-deficit-suppressed gene OsARD encoding an aci-reductone-dioxygenase-like protein in rice. Gene 360, 27–34.[CrossRef][Medline]

Lischka, P., Rosorius, O., Trommer, E. & Stamminger, T. (2001) A novel transferable nuclear export signal mediates CRM1-independent nucleocytoplasmic shuttling of the human cytomegalovirus transactivator protein pUL69. EMBO J. 20, 7271–7283.[CrossRef][Medline]

Mingot, J.M., Bohnsack, M.T., Jakle, U. & Gorlich, D. (2004) Exportin 7 defines a novel general nuclear export pathway. EMBO J. 23, 3227–3236.[CrossRef][Medline]

Nikolaev, A.Y., Li, M., Puskas, N., Qin, J. & Gu, W. (2003) Parc: a cytoplasmic anchor for p53. Cell 112, 29–40.[CrossRef][Medline]

Ohkura, N., Takahashi, M., Yaguchi, H., Nagamura, Y. & Tsukada, T. (2005) Coactivator-associated arginine methyltransferase 1, CARM1, affects pre-mRNA splicing in an isoform-specific manner. J. Biol. Chem. 280, 28927–28935.[Abstract/Free Full Text]

Oram, S., Jiang, F., Cai, X., Haleem, R., Dincer, Z. & Wang, Z. (2004) Identification and characterization of an androgen-responsive gene encoding an aci-reductone dioxygenase-like protein in the rat prostate. Endocrinology 145, 1933–1942.[Abstract/Free Full Text]

Pochapsky, T.C., Pochapsky, S.S., Ju, T., Mo, H., Al-Mjeni, F. & Maroney, M.J. (2002) Modeling and experiment yields the structure of acireductone dioxygenase from Klebsiella pneumoniae. Nat. Struct. Biol. 9, 966–972.

Pochapsky, T.C., Pochapsky, S.S., Ju, T., Hoefler, C. & Liang, J. (2006) A refined model for the structure of acireductone dioxygenase from Klebsiella ATCC 8724 incorporating residual dipolar couplings. J. Biomol. NMR 34, 117–127.[CrossRef][Medline]

Richards, S.A., Carey, K.L. & Macara, I.G. (1997) Requirement of guanosine triphosphate-bound ran for signal-mediated nuclear protein export. Science 276, 1842–1844.[Abstract/Free Full Text]

Saporita, A.J., Zhang, Q., Navai, N., Dincer, Z., Hahn, J., Cai, X. & Wang, Z. (2003) Identification and characterization of a ligand-regulated nuclear export signal in androgen receptor. J. Biol. Chem. 278, 41998–42005.[Abstract/Free Full Text]

Sauter, M., Lorbiecke, R., Ouyang, B., Pochapsky, T.C. & Rzewuski, G. (2005) The immediate-early ethylene response gene OsARD1 encodes an acireductone dioxygenase involved in recycling of the ethylene precursor S-adenosylmethionine. Plant J. 44, 718–729.[CrossRef][Medline]

Sekowska, A., Denervaud, V., Ashida, H., Michoud, K., Haas, D., Yokota, A. & Danchin, A. (2004) Bacterial variations on the methionine salvage pathway. BMC Microbiol. 4, 9.[CrossRef][Medline]

Thakurta, A.G., Gopal, G., Yoon, J.H., Saha, T. & Dhar, R. (2004) Conserved nuclear export sequences in Schizosaccharomyces pombe Mex67 and human TAP function in mRNA export by direct nuclear pore interactions. J. Biol. Chem. 279, 17434–17442.[Abstract/Free Full Text]

Tsukada, Y., Fang, J., Erdjument-Bromage, H., Warren, M.E., Borchers, C.H., Tempst, P. & Zhang, Y. (2006) Histone demethylation by a family of JmjC domain-containing proteins. Nature 439, 811–816.[CrossRef][Medline]

Uekita, T., Gotoh, I., Kinoshita, T., Itoh, Y., Sato, H., Shiomi, T., Okada, Y. & Seiki, M. (2004) Membrane-type 1 matrix metalloproteinase cytoplasmic tail-binding protein-1 is a new member of the Cupin superfamily. A possible multifunctional protein acting as an invasion suppressor down-regulated in tumors. J. Biol. Chem. 279, 12734–12743.[Abstract/Free Full Text]

Wray, J.W. & Abeles, R.H. (1995) The methionine salvage pathway in Klebsiella pneumoniae and rat liver. Identification and characterization of two novel dioxygenases. J. Biol. Chem. 270, 3147–3153.[Abstract/Free Full Text]

Yeh, C.T., Lai, H.Y., Chen, T.C., Chu, C.M. & Liaw, Y.F. (2001) Identification of a hepatic factor capable of supporting hepatitis C virus replication in a nonpermissive cell line. J. Virol. 75, 11017–11024.[Abstract/Free Full Text]

Yoneda, Y. (2000) Nucleocytoplasmic protein traffic and its significance to cell function. Genes Cells 5, 777–787.[Abstract]

Received: 28 August 2006
Accepted: 10 October 2006





This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Gotoh, I.
Right arrow Articles by Seiki, M.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Gotoh, I.
Right arrow Articles by Seiki, M.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE ADVANCED SEARCH TABLE OF CONTENTS