Pfam and PROSITE (pattern/profile) (http://motif.genome.jp/), CLUSTAL W (ver. 1.81) Multiple Sequence Alignments (http://align.genome.jp/), SOSUI (ver. 1.11) (http://bp.nuap.nagoya-u.ac.jp/sosui/) and PSORT II and iPSORT (http://psort.ims.u-tokyo.ac.jp/) predictions were performed using web-based free software. The GenBank accession numbers that were used were ARMCX1 (NM_016608), ARMCX2 (NM_014782), ARMCX3 (NM_016607), ARMCX4 (AK292543), ARMCX5 (NM_022838), ARMCX6 (NM_019007), armadillo repeat containing 10 (ARMC10) (alias SVH) (NM_031905), G-protein coupled receptor-associated sorting protein 1 (GPRASP1) (NM_014710) and GPRASP2 (NM_001004051).

Human cervical carcinoma (HeLa) cells were cultured in Earle's modified Eagle's medium (MEM) supplemented with 10% fetal bovine serum, 1% non-essential amino acids and antibiotic-antimycotics (all from Invitrogen, Carlsbad, CA, USA). The ARMCX6 EST (expressed sequence tag) cDNA clone (GenBank no. BC007677; IMAGE 3609980; Open Biosystems, Huntsville, AL, USA) was used as a template to amplify the ARMCX6-coding region using a polymerase chain reaction (PCR), and the coding region was ligated into the respective enzyme EcoRI and BamHI sites of the pEGFP-N2 vector (Clontech, Mountain View, CA, USA). The PCR primer set was as follows (underlines indicate the flanking enzyme site): CGGAATTCGCCACCATGGGCCGGGCTCGGGAAGTG and CGGGATCCATGGGGCAGGGGTTTCCAG. The plasmid sequence was verified by sequencing at the Takara facility (Mie, Japan). The cells were transfected using Lipofectamine Plus (Invitrogen) or the Neon Transfection System (Invitrogen) according to the manufacturer's instructions. Briefly, Neon electroporation was performed in a 24-well plate with a pulse voltage of 1,200 and a pulse width of 40. After 24–48 h of transfection, the cells were washed using phosphate-buffered saline (PBS) and photographed using a fluorescent microscope DM IRB (Leica, Wetzlar, Germany), equipped with a digital single-lens reflex camera (Camedia E-20; Olympus, Tokyo, Japan).

The total RNA samples were prepared using an RNeasy Mini-Spin Column (Qiagen, Valencia, CA, USA) in accordance with the manufacturer's instructions. Prior to RT-PCR, total RNA was treated with deoxyribonuclease I (Invitrogen) according to the manufacturer's instructions. One or two micrograms of total RNA was reverse transcribed using High-Capacity cDNA Reverse Transcription Kits (Applied Biosystems, Foster City, CA, USA). The PCR was carried out in 25 μl of a mix consisting of 1X buffer, 200 μM of dNTPs, 400 nM of primers, 1 mM of MgSO₄, 5% DMSO and 1 unit of KOD plus DNA polymerase (Toyobo, Osaka, Japan). Hot-start PCR was then performed as follows: denaturation for 3 min at 94°C, followed by ad libitum cycles at 94°C for 15 sec, a gene-specific annealing temperature [ARMCX3, 56°C; ARMCX4, 56°C; ARMCX5, 57°C; ARMCX6, 54°C; ARMC10, 67°C; glyceraldehyde-3-phosphate dehydrogenase (GAPDH), 60°C] for 30 sec and 68°C for 30 sec, followed by an extension step of 3 min at 68°C. The PCR results were verified by varying the number of PCR cycles for each cDNA and set of primers. The target gene primer pairs were as follows: ARMCX3, GGGCTGTCCAGA AACGGGCT and CCCTGAGCAGTTCCCTAGTC for 540 bp; ARMCX4, GCAAGAAGTGGGCCTAGGGC and TTCCCT GGGTACTGCCAAGG for 420 bp; ARMCX5, GGCCTAAT CCGAAGGCCTGC and CCAGCTGCACAGGGGAGTTC for 360 bp; ARMCX6, GGGCTCAATCCAGGACCACA and GTTCACTATCCATCAGGCGC for 350 bp; ARMC10, TCAC CTGCCAACCTGACCAT and GCGCTATCTCAGCTCAC TGC for 460 bp; and GAPDH (NM_002046), CCATGGC AAATTCCATGGCA and GTCCTTCCACGATACCAAAG for 365 bp. The amplified products were separated on 1.0% agarose gels and visualized under ultraviolet transillumination. For the cDNA panel analysis, 2.5 μl of cDNA purchased from Clontech was used (Human MTC panels I and II cDNA panel). The GAPDH primer in the kit was used as a control to amplify 983 bp.

The cells were harvested and lysed in modified RIPA lysis buffer [50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1% Nonidet P40, 0.5% sodium deoxycholate and 1 mM EDTA] containing a protease inhibitor cocktail (Sigma, St. Louis, MO, USA) for 20 min on ice. The cell lysates were centrifuged, and the protein concentrations were determined using the Bio-Rad Protein Assay Kit (Bio-Rad Laboratories, Hercules, CA, USA). The protein lysates were loaded onto each lane of a gel. Before performing sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), the reaction was terminated by the addition of Laemmli sample buffer containing 100 mM of dithiothreitol (DTT). Equal amounts of cellular protein were electrophoresed on NuPAGE 4–12% Bis-Tris gel with MES running buffer (Invitrogen) and transferred to a Hybond-PVDF membrane (GE Healthcare, Piscataway, NJ, USA). The membrane was first blocked using PBS containing 0.1% Tween-20 and 5% non-fat dried milk, then incubated with GFP (JL-8; Clontech) and GAPDH (Applied Biosystems) antibodies. Alkaline phosphatase (AP)-labeled secondary antibodies were purchased from Promega (Madison, WI, USA). A Western Blue-stabilized substrate was used to detect the signals, according to the manufacturer's protocol (Promega).

To determine the functional aspects of ARMCX6, we predicted the motifs or domains of the amino acids (aa) composing ARMCX6 using a Pfam and PROSITE (pattern and profile) search. Pfam revealed that a protein of unknown function (referred to as DUF634) is a characteristic feature of the carboxyl (C)-terminal of the ARMCX6 protein (9–300 aa of full-length 300 aa). In addition to ARMCX6, DUF634 was observed in other proteins including ARMCX1 (184–447 aa of full-length 453 aa), ARMCX2 (365–626 aa of full-length 632 aa), ARMCX3 (100–363 aa of full-length 379 aa), ARMCX5 (288–552 aa of full-length 558 aa) and ARMC10 (74–337 aa of full-length 343 aa). Among the ARMCXs, DUF634 was not identified in ARMCX4. DUF634 was also recognized in the C-terminus of GPRASP1 (1,132–1,384 aa of full-length 1,395 aa) and GPRASP2 (574–836 aa of full-length 838 aa). In contrast, Arm (Armadillo/β-catenin-like repeat) was exclusively predicted in ARMCX1 (239–276 aa), ARMCX2 (419–457 aa), ARMCX3 (155–192 aa) and ARMC10 (127–167 aa). Based on PROSITE, ARM_REPEAT (Armadillo/plakoglobin ARM repeat motif) was only noted in ARMCX1 (247–284 aa) and ARMCX3 (163–200 aa).

Next, we performed CLUSTAL W multiple sequence alignments using full-length ARMCX1 (453 aa), ARMCX2 (632 aa), ARMCX3 (379 aa), ARMCX4 (348 aa), ARMCX5 (558 aa), ARMCX6 (300 aa) and ARMC10 (343 aa). GPRASP1 and GPRASP2 apparently had longer coding regions compared to ARMCXs and ARMC10; therefore, these two proteins were excluded from further analysis. Consistent with the DUF634 domain, the C-terminals of each of the proteins were similar (data not shown), suggesting the evolutionarily conserved functional importance of this region. Previously, ARMCX1, ARMCX2 and ARMCX3 were predicted to possess similar N-terminal transmembrane domain and armadillo repeats (3). Based on the alignment of ARMCX1-6 and ARMC10, conserved amino acids were detected in the N-terminal transmembrane domain (data not shown); however, SOSUI prediction revealed that one trans-membrane helix was identified in the N-terminal of ARMCX1 (5–27 aa), ARMCX2 (5–27 aa) and ARMCX3 (7–29 aa) as well as the central region of ARMCX4 (96–118 aa) but not in any region of ARMCX5 or ARMCX6. SOSUI also predicted that ARMCX5 and ARMCX6 may be soluble proteins. On the other hand, iPSORT predicted that the N-terminal (1–30 aa) of ARMCX6 as well as ARMCX1–3 and ARMC10 may be a signal peptide. Apparently, ARMCX4 and ARMCX5 do not possess a signal peptide.

To predict the subcellular localization of ARMCXs, PSORT II prediction was performed. ARMCX6 had a 44.4% probability of localizing in an extracellular compartment, including the cell wall, whereas the other ARMCXs did not have a high probability of being localized extracellularly (Table I). Based on published information, PSORT II prediction is a reliable bioinformatics tool. Indeed, ARMCX3 has been shown to be an integral membrane protein of the mitochondrial outer membrane (4). In addition to the ARMCXs, ARMC10 also possesses a transmembrane domain and has been shown to localize in the endoplasmic reticulum (10). PSORT II predicted that ARMC10 is localized in vacuolar regions (22.2%), extracellular regions including the cell wall (22.2%), the Golgi apparatus (22.2%), mitochondrial regions (22.2%) and the endoplasmic reticulum (11.1%). These cellular regions are functionally proximal to the endoplasmic reticulum, as they all are membrane-enclosed organelles. To determine the cellular localization of ARMCX6, we constructed a C-terminal green fluorescent protein (GFP)-tagged ARMCX6 expression plasmid. Fluorescent microscopy revealed a few GFP-positive cells in a GFP-tagged ARMCX6-transfected well, whereas abundant GFP-positive cells were detected in a control GFP-transfected well (data not shown). In the GFP-tagged ARMCX6-transfected well, the GFP-positive cells were all round in shape and had shrunken, but the control GFP-transfected cells showed a typical extended shape (Fig. 1A, middle panel vs. upper panel). Exceptionally, GFP-positive intact cells were detected in the GFP-tagged ARMCX6-transfected well. In it, GFP-tagged ARMCX6 was diffusely spread throughout the cells, with the probable exception of the nuclei (Fig. 1A, lower panel).

After 48 h of transfection, using an anti-GFP antibody, an ARMCX6-GFP fusion protein was detected as a faint band with an estimated molecular weight of ∼60 kDa, compared with the abundantly expressed 27-kDa GFP (Fig. 1B, upper panel). GAPDH was consistently detected in all lanes (Fig. 1B, lower panel). The ARMCX6 protein seems to be unstable, since another expression system FLAG-tagged ARMCX6 again produced a faint band (data not shown). Furthermore, we constructed N-terminal deleted FLAG-tagged ARMCX6 constructs to exclude the signal peptide but failed to detect a reasonable protein level using immunoblotting (data not shown). Taken together, these results suggest that ARMCX6 is not a secreted protein and cellular toxicity is induced by ectopic ARMCX6 expression.

Next, we focused on the transcriptional regulation of ARMCXs, the chromosomal localizations of which are uniquely mapped to the same region of Xq. ARMC10 is localized at 7q22.1 but was included in this investigation. To date, the mRNA distributions of ARMCX1 and ARMCX2 in human tissues have been reported (3), therefore ARMCX1 and ARMCX2 were omitted from the present study. The primer set used to detect ARMCX3–6 and ARMC10 was able to specifically amplify their mRNAs in the case of RT plus- but not RT minus-derived cDNAs (Fig. 2A). The expression levels of ARMCX3–6 and ARMC10 were examined using a multiple tissue-derived cDNA pool that included 16 adult tissues. ARMCX6 mRNA was predominantly detected in the pancreas and spleen, moderately detected in the liver, skeletal muscle, kidney, thymus, prostate, testis, ovary and small intestine, and scarcely detected in the placenta and leukocytes; ARMCX6 mRNA was not detected in the heart, brain, or colon (Fig. 2B). In contrast to ARMCX6, ARMCX3–5 and ARMC10 showed a relatively ubiquitous expression pattern (Fig. 2B).