| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* Department of Molecular Pharmacology, Albert Einstein College of Medicine, Bronx, New York, USA; and
Institut de Biotecnologia I de Biomedicina and Department Bioquimica i Biologia Molecular, Universitat Autonoma de Barcelona, Bellaterra (Barcelona), Spain
1Correspondence: Department of Molecular Pharmacology, Albert Einstein College of Medicine, 1300 Morris Park Ave., Bronx, NY 10461, USA. E-mail: fricker{at}aecom.yu.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: posttranslational processing • Nna1 • Purkinje cell degeneration • CCP
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Most CPs are present in the secretory pathway of various cell types, functioning either within the cell by processing molecules that are secreted or functioning after secretion by processing extracellular peptides and proteins. In addition to these roles, CPs are involved in the processing of alpha-tubulin (2
, 3)
. A ligase attaches Tyr to the C terminus of alpha-tubulin (4
, 5)
, and an unidentified CP removes this Tyr. The role of the tyrosinylation and detyrosinylation of tubulin is likely to contribute to microtubule stability and/or function. In various human carcinomas, the extent of tubulin processing correlates with tumor progression (6
, 7)
. Thus, inhibitors of the tubulin CP may have value as therapeutics in cancer treatment. Despite considerable interest, little progress has been made in the purification of the tubulin tyrosine CP due to its instability. The enzymatic properties of the tubulin tyrosine CP are distinct from all other known CPs (2
, 3)
. Furthermore, the previously identified CP activities are not generally thought to be cytosolic, and therefore, these proteins would be unable to function as tubulin-processing enzymes.
Most mammalian metalloCPs have been traditionally divided into two subfamilies based largely on amino acid sequence similarity (8
, 9)
. The A/B subfamily (M14A) includes the pancreatic digestive enzymes CPA1, CPA2, and CPB, the mast cell CPA, a plasma CPB that functions in clot lysis, and four additional genes that are expressed in a limited number of tissues (10
11
12
13
14
15
16)
. The N/E subfamily (M14B) includes CPN, a serum enzyme, CPE, a neuroendocrine enzyme that activates peptides, CPD, a broadly expressed enzyme that functions in the processing of proteins within the trans Golgi network, CPM, a broadly distributed enzyme that functions on the cell surface, and CPZ, an enzyme present in the extracellular matrix (ECM) that regulates the activity of specific forms of Wnt (17
18
19
20
21)
. In addition to these active members of the N/E subfamily, there are three members of the N/E family (CPX1, CPX2, and ACLP/AEBP1) that do not appear to encode active CP enzymes based on the absence of critical active site residues as well as on direct biochemical analyses (22
23
24
25)
.
Recently, a screen for mRNAs up-regulated by axonal regeneration identified a gene product with homology to the CP family; this gene product was named Nna1 (26)
. However, the initial alignments of Nna1 with the enzymatically active CPs did not support a role for Nna1 as an active enzyme due to the absence of critical active site residues (26)
, although a recent study showed an alignment more consistent with an active enzyme (27)
. Because Nna1 was found to be both cytosolic and nuclear when expressed as a fusion protein with green fluorescent protein (GFP), Nna1 was proposed to function in the processing of nuclear factors needed for neuronal regeneration and survival (26)
. Support for a role in neuronal survival was provided by the discovery that Purkinje cell degeneration (pcd) mice have a defect in the gene encoding Nna1 (28)
. These mice, which resulted from a spontaneous mutation, have an altered gait due to the degeneration of Purkinje cells (29)
. The pcd mice also show degeneration of olfactory bulb mitral cells (30)
, retinal photoreceptor cells (31
, 32)
, and other defects such as sterility (29)
and abnormal sperm shape (33)
. Interestingly, the Purkinje cells in these mice die around the third postnatal week whereas mitral and retinal cells degenerate slowly around 1 yr of age. Altogether, five variants of the pcd mutation are known. The 3J mutation represents a deletion of several exons of the Nna1 gene, while the others are more subtle mutations that decrease mRNA levels (28
, 34)
. Recently, Wang and colleagues (27)
found that the degeneration of Purkinje cells in pcd3J mice could be rescued by expression of wild-type (WT) Nna1 but not by a mutant lacking two putative substrate-binding residues, suggesting that CP activity is essential for the function of this protein.
The previous studies have left many unanswered questions. First, why is cell death in the pcd mice limited to Purkinje cells of young mice and a few additional cell types in older mice? Nna1 is broadly distributed in the central nervous system (CNS) and also present in other tissues (26
, 28)
; therefore, the absence of this protein would be expected to have an effect on more cell types. Second, what is the function of Nna1? To address these questions, we first searched the mouse genome for additional Nna1-like sequences. In a related study that is published as an accompanying paper, genomes from a variety of organisms were also searched for Nna1-like sequences (35)
. From these searches, it was clear that five additional Nna1-like genes existed in mouse and human genomes. Alignment of these various sequences revealed several highly conserved residues, including critical active site residues that are necessary for the catalytic activity of A/B and N/E subfamily CPs. Structural modeling based on these alignments resulted in structures that share the basic CP catalytic domain structure of both A/B and N/E CPs. The distribution of the mRNAs encoding the various Nna1 homologs in mouse tissues suggests a broad function with some redundancy, thus explaining why only a few cell types die in the pcd mice. The absence of a signal peptide from all forms of the five novel Nna1 homologs suggested a cytosolic distribution; this was confirmed by expression in cell lines and immunofluorescence analysis. Because the name Nna1 was based on the previous study of nuclear localization in neurons and from the present data, it is clear that neither Nna1 nor the other family members are restricted to neurons or the nucleus, we named this subfamily cytosolic carboxypeptidase (CCP). The new members are numbered from CCP2 through CCP6, and we refer to Nna1 as Nna1/CCP1. One function of these family members appears to be tubulin processing, based on the absence of C-terminally-truncated alpha tubulin forms in the mitral cells of pcd mice.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
.
Bioinformatics and cDNA sequence analysis
To identify genes and gene products with homology to Nna1/CCP1, the mouse genome was searched with mouse Nna1/CCP1 protein sequence using tblastn (http://www.ncbi.nlm.nih.gov/). The nonredundant (NR) and expressed sequence tag (EST) GenBank databases were searched using mouse Nna1/CCP1, and the fragments were identified from the genome search. Full-length cDNA sequences of the various splice forms were compiled from the NR and EST databases. Several cDNA clones with incomplete sequences in GenBank were purchased (Invitrogen, Carlsbad, CA) and sequenced in both directions; these included clone 3994516 (CCP2), 6813901 (CCP5), and 6446357, 5024144, and 6742069 (CCP6). For CCP4, the longest cDNA sequences in the NR and EST databases appeared to encode only a portion of the CP domain, whereas the genome showed the presence of additional potential exons with homology to the CP domain of Nna1/CCP1. To confirm this, reverse-transcription polymerase chain reaction (RT-PCR) was performed and the resulting PCR product was sequenced (described below).
Alignment of CCPs and modeling
Amino acid alignments of the six mouse CCPs were performed using DNA-Star MegAlign version 5.05 using gap penalty and gap wt penalty values of 9. For this analysis, a representative amino acid sequence of each mammalian metalloCP was included.
Structural modeling was performed on mouse CCP1 using a CPA/B structural alignment extracted from HOMSTRAD database (37
, 38)
, which includes pancreatic CPA from Bos taurus (2ctc) and Sus scrofa (1pca) and the pancreatic CPB from Sus scrofa (1nsa) and CPT from Thermoactinomyces vulgaris (1obr). These structures were checked as templates against mouse CCP1 and used for modeling the latter protein sequence with MODELLER 8v2 modeling suite environment (39
, 40)
Batches of at least 500 submits were performed until models with good geometry, and suitable fold requirements were retrieved and verified with PROCHECK (41)
, VERIFY3D (42)
, and JOY (38)
.
The CCP1 derived alignment was transferred to the other candidate CCPs, and they were subsequently modeled following the same procedures and crosschecked against each other and their templates to ensure their quality. Swiss-PDB Viewer (43)
, Jmol (www.jmol.org), and PyMOL (44)
were used for visually checking the quality of the final models and to determine the spatial orientation of certain key residues (discussed below). PyMOL was also used for generating surface representations of the models.
RNA extraction and RT-PCR
Total RNA from different mouse brain regions and organs were obtained using an RNA easy kit (Qiagen, Valencia, CA, USA). RT-PCR was performed using the OneStep RT-PCR kit (Qiagen) amplifying an appropriate sequence of CCP mRNA using the forward and reverse primers from Supplemental Table 1. Briefly, 100 ng of total RNA was mixed with primers, dNTPs, and an enzyme combination of Omniscript reverse transcriptase, Sensiscript reverse transcriptase, and HotStarTaq DNA polymerase. The reverse transcription was carried out for 30 min at 50°C followed by a denaturation step at 95°C for 15 min and by 30–40 PCR cycles (40 s at 94°C, 0.5–1 min at 55–64°C, and 40 s 72°C). The resulting cDNA was separated in a 1% agarose gel. Most PCR products were subcloned into the pCRII-TOPO vector (Invitrogen) and sequenced.
In situ hybridization
Adult C57B6 mice (7- to 8-mo-old, male) were anesthetized in a CO2 chamber. Brains were dissected out and frozen in isopentane (Sigma, St. Louis, MO, USA) at –50°C and stored at –80°C. Brain coronal and sagital sections (14 µm thick) were thaw-mounted onto Superfrost/Plus slides (Fisher Scientific, Pittsburgh, PA, USA) and stored at –80°C until use.
To generate the probes, RT-PCR was used to produce cDNA fragments that corresponded to regions with low homology to other CCPs or to other sequences present in the GenBank database (see Supplemental Table 1). PCR products were subcloned into pCRII-TOPO (Invitrogen) and sequenced. The 35S-labeled antisense RNA probes were transcribed from linearized templates with SP6 RNA polymerase (Invitrogen), along with 35S-uridine triphosphate (UTP; 1250 Ci/mmol; PerkinElmer Life Sciences Inc., Boston, MA, USA). Labeled sense-strand probes were transcribed with T7 RNA polymerase (Invitrogen). Unincorporated nucleotides were removed by chromatography with Sephadex G-50 (Amersham Biosciences, Piscataway, NJ, USA).
In situ hybridization (IHS) analysis was performed as described previously (45)
. Briefly, brain sections were fixed with 4% paraformaldehyde in 0.1 M sodium phosphate buffer at 4°C, acetylated with 0.25% acetic anhydride in 0.1 M triethanolamine, and dehydrated in ethanol. Hybridization was performed using 107 cpm of probe per ml of buffer in a humidified chamber at 55°C for 18–20 h. After being rinsed, the slides were incubated at 37°C for 30 min with RNase A (Qiagen) and washed twice for 15 min in 2x standard saline citrate [saline-sodium citrate (SSC)] at 37°C, incubated for 30 min at 60°C in 0.5x SSC, washed for 30 min at 65°C in 0.1x SSC, and rinsed for 5 min in 0.1x SSC at room temperature. They were dehydrated in ethanol and air dried.
Autoradiograms were generated by apposition of labeled sections to X-Omat Blue X-1 film (Eastman Kodak Co, Rochester, NY, USA) for 5–22 days sat room temperature. The slides were then dipped in Hypercoat LM-1 nuclear emulsion (Amersham Biosciences) and stored for several weeks at 4°C. The emulsion was developed in D19 (Eastman Kodak Co.), rinsed in water, and fixed in Kodak Polymax Fixer (Eastman Kodak Co.). Sections were then counterstained lightly with Mayer’s hematoxylin, dehydrated in a graded ethanol series, defatted in xylene, mounted in cytoseal (Stephens Scientific, Riverdale, NJ, USA), coverslipped, and analyzed under a Zeiss Axiophot microscope.
Immunohistochemistry
WT (10 wk old) and pcd mice (10 and 14 wk old) were anesthetized with ether and perfused with 4% paraformaldehyde. Brains were removed and postfixed for 2 h, placed into 30% sucrose Tris-buffered saline solution overnight at 4°C, and frozen in isopentane (Sigma) at –50°C and stored at –80°C. Brain coronal and sagital sections (14 µm thick) were serially cut on a sliding microtome at –20°C, thaw mounted onto superfrost/Plus slides (Fisher Scientific, Pittsburgh, PA, USA), and stored at –80°C until use. Sections were air-dried for 2 h. After being washed in PBS containing 0.2% Tween 20, the sections were blocked to prevent nonspecific binding of antibody (Ab) with 10% normal goat serum in PBS containing 0.5% Triton X-100 for 2 h at room temperature and incubated overnight at 4°C with primary Ab. Antibodies and dilutions included rabbit anti-Glu-tubulin (1:250, Chemicon, Temecula, CA, USA), rat anti-Tyr-tubulin (1:100, Chemicon), and rabbit anti-CPD C-terminal antisera AE160 (1:1000). After being washed extensively to remove unbound primary Ab, biotinylated secondary antibodies were applied. After 2 h of incubation at room temperature, the sections were washed and immunostained using a fluorescent avidin kit (Vector Laboratories, Inc., Burlingame, CA, USA) according to the manufacturer’s instructions. In some experiments, sections were stained with Mayer’s hematoxylin.
Expression of CCPs in cell lines
Full-length human CCP1 cDNA was obtained from Invitrogen (clone 30344122). The 4 kb CCP1 cDNA was subcloned into pcDNA3 lacking the KpnI site. Then, oligonucleotides introducing the sequence MYPYDVPDYA were ligated into the KpnI site of CCP1, which is located in the C-terminal region following the CP domain. Mouse CCP2 cDNA in the vector pCMV-SPORT6 was purchased from Invitrogen (clone 3994516). This clone lacked the initiation ATG and six N-terminal amino acids. Oligonucleotides containing an ATG initiation site encoding the sequence MYPYDVPDYAVDPRVR were inserted into the XmaI-SalI sites located upstream of the CCP2 N terminus. Full-length mouse CCP5 cDNA was obtained from Invitrogen (clone 6813901). The 3.3 kb CCP5 cDNA insert was subcloned into pcDNA3. Oligonucleotides encoding the sequence YPYDVPDYAGRAS were ligated into the AscI site in position 1730 which is located immediately after the CP domain. Full-length CCP6 was obtained from Invitrogen (clone 6446357). The 1.9 kb SmaI-XhoI CCP6 cDNA fragment was inserted into pcDNA3 lacking an XbaI site. Oligonucleotides encoding the sequence YPYDVPDYALD were ligated into the XbaI site of CCP6 located in the C-terminal region (following the CP domain). For every plasmid, the sequence of the entire coding region was verified with a combination of vector and internal primers.
Cell culture transfection and analysis
Neuro2A cells, AtT20 cells, NIH3T3 cells, and COS7 cells were grown in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% FBS. PC12 cells were grown in DMEM supplemented with 5% FBS and 10% heat inactivated horse serum. Transfection of CCP plasmids into Neuro2A cells used the Trans IT-Neural transfection reagent (Mirus Madison, WI, USA) and transfection of NIH3T3, COS7, AtT20, and PC12 cells used Lipofectamine 2000 (Invitrogen). Cells were used 24 h after transfection.
For immunofluorescence analysis, Neuro2A, NIH3T3, COS7, AtT20, and PC12 cells expressing different CCP constructs were rinsed with cold PBS, fixed with 4% paraformaldehyde in PBS at room temperature for 15 min, washed, permeabilized with 0.1% Triton X-100 in PBS for 15 min, blocked in PBS containing 5% BSA for 1 h, and incubated with antisera that recognize the hemagglutinin (HA) epitope (YPYDVPDYA): mouse anti-HA primary antibodies (1:2500, Sigma) or mouse HA.11 monoclonal antibody (mAb; 1:1000, Covance) in PBS containing 5% BSA. After being washed, cells were incubated with CY-2 conjugated secondary anti-mouse Ab (1:100, Jackson ImmunoResearch Laboratories), washed, and the coverslips were mounted with Vectashield mounting medium for fluorescence with 4'-6-diamidino-2-phenylindole [4',6'-diam idino-2-phenylidole (DAPI), Vector Laboratories, Inc.].
For Western blot analysis, Neuro2A, NIH3T3, and COS7 cells transiently expressing the CCPs were extracted in buffer containing 1% SDS and subjected to denaturing PAGE followed by transfer to nitrocellulose membranes. The HA epitope was detected with mouse mAb HA.11 (1:1000, Covance) and an IgG-peroxidase conjugate (1:3000, Amersham Biosciences), using the enhanced chemiluminescence (ECL) reaction (Pierce, Rockford, IL, USA). In some experiments, NIH3T3 cells expressing CCP2 were treated with either 10 nM phorbol 12-miristate-13-acetate (Sigma) for 1 h, or 10 µM forskolin (Sigma) for 1 h, or subjected to serum deprivation for 1 h or heat shock for 1 h at 43°C followed by incubation for 1 h or 6 h at 37°C. In addition, Neuro2A cells expressing different CCP constructs were treated with 5 µM A23187 (Calbiochem, San Diego, CA, USA) for 20 min or 0.5 µM thapsigargin (Sigma) for 1.5 h. Cells were analyzed by Western blotting as described above.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
). The length of the CCP2–6 genes varies from 18,000 nucleotides for mouse CCP5 to over 1,200,000 nucleotides for mouse CCP6 (Fig. 1)
. Interestingly, the human gene for CCP6 is also over 1,200,000 nucleotides, and the introns are generally similar in length between human and mouse genes. The length of the cDNA for each of the CCPs ranged from 1881 bp for CCP6 to 4369 bp for CCP2 (Supplemental Figures S1-S5). Representative sequences of the splice forms of each mouse CCP have been deposited in GenBank (accession numbers DQ867026 through DQ867040).
|
For most of the CCPs, a large number of cDNA clones were found in both the NR and EST databases, and except for CCP4 (discussed below) many of these had long open reading frames that included the CP domain. A comparison of the mouse cDNA and genomic sequences revealed a large number of alternatively spliced forms; these were investigated by sequencing cDNA clones and RT-PCR products (Supplemental Figures S1-S6). For CCP4, the longest cDNA sequence in the databases corresponded to exons 1 through 13, which only encodes a portion of the CP domain. However, the mouse genome contains two additional regions with homology to the remainder of the CP domain (exons 14 and 15, Fig. 1
). RT-PCR and sequence analysis were used to confirm that CCP4 mRNA exists with these additional two exons (Supplemental Figure 3).
Alignments and modeling of the CCPs
The amino acid sequences of the CCPs have similarities to the 300-residue CP catalytic domain of A/B and N/E family CPs (Fig. 2
). In addition, all six mouse CCPs have a 120–150 residue N-terminal domain that is moderately conserved among CCPs (Fig. 2)
. This N-terminal domain begins with the FEGSN sequence and ends with the YPYTYS sequence (Fig. 3
). The length and location of the 120–150 residue N-terminal domain in the CCPs are similar to the pro domain within A/B family CPs, although there is no substantial amino acid sequence similarity (Fig. 3)
. Nna1/CCP1 has a long N-terminal extension of over 500 amino acids, while CCP5 has only 10–15 amino acids preceding the N-terminal domain (Fig. 2)
. There is also considerable variability in the length of the C-terminal extension following the catalytic CP domain (Fig. 2)
. The various CCPs show no sequence similarity among their N-terminal or C-terminal extensions.
|
|
Within the N-terminal and the CP domains, most of the CCPs require similar short gaps or inserts (<20 residues) to align both the sequences and the predicted secondary structure elements with other CPs (Figs. 2
and 3)
. The exception is CCP5, which contains several large inserts relative to the other CPs (Fig. 3)
. Structural modeling suggests that mouse CCP catalytic domains display the typical fold containing an alpha/beta/alpha sandwich structure with an antiparallel beta-sheet of eight strands. The residues involved in the coordination of the active site zinc share the same conformation described in other metalloCPs of the A/B and N/E subfamilies. The Zn atom is held in place by penta-coordination with two His residues, one Glu residue, and a water molecule (data not shown). A comparison of the structure of mouse CCP1 and bovine CPA gave a root mean square of 1.602 and a DRMS of 1.191, indicating that the two structures are quite similar.
There are several motifs found in all CCPs, some of which correspond to active site or important structural residues of A/B and N/E family CPs (described below). Others, such as the FESGN motif, the WFYF motif, and the YPYTYS motif found in the N-terminal domain of the CCPs, have no counterpart in other family CPs (Fig. 3)
. Of these regions, the FESGN corresponds to the beginning of the conserved N-terminal domain. Interestingly, the YPYTYS sequence corresponds to the junction between the N-terminal domain and the CP domain, which is located 
70 residues upstream of the HxxE sequence (discussed below).
Motifs common to the CCPs and other CP families include the zinc binding residues His69, Glu72, and His196 (by convention, the numbering system of bovine CPA and CPB is used, which assigns residue 1 as the N terminus after pro peptide removal). Other conserved residues include Arg127, which in the A/B and N/E family CPs binds to the carbonyl bond at the cleavage site, and Glu270, which transfers a proton from the incoming water molecule to the leaving amine. Another conserved motif is the NPDG sequence located upstream of Arg127; this NPDG sequence is highly conserved in all CPs and presumably plays a role in the structure or interactions of the catalytic domain. Based on the presence of the active site residues that are critical for catalytic activity and the modeling results that suggest these residues are properly located in the 3-dimensional structure to perform a catalytic role, it is likely that the CCPs encode functional peptidases. This proposal is supported by experimental results in the accompanying paper, in which a C. elegans CCP homologue was shown to be enzymatically active (35)
.
Analysis of the residues in the putative substrate binding site allows for predictions of the cleavage specificity of the CCPs. The presence of an Asn and Arg in the positions equivalent to 144–145 of CPA/B implies that the CCPs will cleave C-terminal residues (in CPA and CPB, Arg145 binds to the carboxylate of the C terminus of the substrate). Another critical substrate-binding site in CPs is residue 255. It is not possible to accurately predict the residue in the position equivalent to 255 for any of the CCPs due to limited sequence similarity with A/B or N/E CPs in this region. For Nna1/CCP1 and CCP4, the residue in position 255 is likely to be an Ala, and for CCP2, CCP3, CCP5, and possibly CCP6, a Gly. If so, then these enzymes would presumably cleave C-terminal hydrophobic and/or basic amino acids, based on the broad substrate specificity of an insect CP with a Ser in position 255 (46)
. For CCP6, it is also possible that position 255 is occupied by an Arg (Fig. 3
, bottom). This would imply that CCP6 cleaves acidic C-terminal residues. In the modeling studies, the difference between models was small since the particular region involved does not offer a high local stability within the model.
Distribution of the CCPs
The tissue distribution of the various CCPs was examined by RT-PCR using the minimal number of cycles necessary to produce detectable signals for most tissues without saturating the signal for those tissues with higher amounts of the CCP mRNA (Fig. 4
). These experiments were replicated several times with different amounts of starting RNA and/or numbers of cycles, with comparable results. Nna1/CCP1 RNA is broadly distributed among tissues, whereas the other CCPs show more striking variations in distribution. CCP2 RNA is abundant in testis, detectable in brain, eye, muscle, lung, pancreas, intestine, stomach, pituitary, spleen, adrenal, and kidney, and very low or not detectable in the other tissues examined (Fig. 4)
. CCP3 RNA is abundant in testis, pituitary, and kidney, moderate in brain, eye, fat, lung, pancreas, stomach, and adrenal, and low or undetectable in the other tissues examined. CCP4 RNA is abundant in eye, muscle, and testis, present in brain and pituitary at lower levels, and not detectable in the other tissues examined (Fig. 4)
. However, when RT-PCR was performed with oligomers to another part of the CCP4 mRNA (exons 11–15), there was no signal detected for pituitary and the signal for testis was very weak (Supplemental Figure 6). Thus, the forms of CCP4 mRNA in these tissues represent the shorter forms detected in the bioinformatics analysis, which lack the CP domain. CCP5 is abundant in testis, moderate in pituitary, brain, eye, and kidney, and lower in all other tissues examined (Fig. 4)
. CCP6 mRNA is abundant in testis, pituitary and brain, moderate in eye, stomach, adrenal and kidney, and much lower or undetectable in the other tissues examined (Fig. 4)
. For most of the probes and tissues, a single band of PCR product was detected, corresponding to the predicted size of the cDNA; this was confirmed by sequencing. However, for some probes and tissues (CCP3 with lung and several other tissues; CCP4 with pituitary and testis; CCP5 with several tissues) another major band was detected. Although these were not sequenced, the size of these other bands corresponds to the predicted size of the mRNA with an unspliced intron (CCPs 3 and 5) or to a missing exon (CCP4); this is consistent with the variability of splicing seen during the cDNA cloning, bioinformatic analysis, and RT-PCR described above.
|
ISH was used to examine the distribution of CCP mRNA in brain. The probes used for ISH corresponded to regions N-terminal to the CP domain that showed no significant nucleotide sequence similarity among CCPs or to other sequences in GenBank. Nna1/CCP1 was included as a positive control; the distribution of this mRNA in mouse brain has previously been reported (26
, 28)
. The other CCP mRNAs produced much weaker signals than Nna1/CCP1, requiring substantially longer exposure times to produce signals of comparable strength (Fig. 5
). Nna1/CCP1 mRNA was broadly expressed throughout the mouse brain, with slightly higher signals in cerebellum and hippocampus (Fig. 5)
, as previously reported (28)
. CCP2 also showed higher levels in the cerebellum than other areas (Fig. 5
, middle panels). Localized expression of CCP2 mRNA was detected in cells adjacent to the lateral ventricle, the third ventricle, and the dorsal third ventricle such as the paraventricular thalamic nucleus. Similarly, CCP3 mRNA was found in the cells adjacent to these ventricles and in the cerebellum (Fig. 5)
. CCP4 mRNA was broadly distributed at a low level throughout the brain, with no distinctive pattern (Fig. 5)
. CCP5 mRNA showed a slight enrichment in the cerebellum and olfactory bulb and in a region near the paraventricular thalamic nucleus (Fig. 5)
. CCP6 mRNA was broadly distributed, although levels in the cerebellum and olfactory bulb were lower than in most other brain regions (Fig. 5)
.
|
Because pcd mice lack Nna1/CCP1 expression and show selective loss of Purkinje and mitral cells, the levels of all CCP mRNA were examined in more detail in the cerebellum and olfactory bulb. Nna1/CCP1 mRNA is abundant in the Purkinje cells as well as granular cells (Fig. 6
), as previously reported (28)
. CCP2 mRNA is also present in granular cells but is not detectable in Purkinje cells (Fig. 6)
. CCP3 mRNA is present at low levels throughout the various layers of the cerebellum but is not enriched in Purkinje cells (Fig. 6)
. CCP5 mRNA is present in granular cells at low levels, and not detectable in the Purkinje cells (Fig. 6)
. CCP6 mRNA is detectable in the Purkinje cells and in selected cells within the granular layer (Fig. 6)
.
|
Nna1/CCP1 mRNA is expressed in the mitral cells of the olfactory bulb and also in cells within the granular layer (Fig. 7
), consistent with a previous report (28)
. Low levels of CCP2 mRNA are detectable in granular and mitral cells (Fig. 7)
. CCP3 mRNA is detectable in the mitral cell layer of the accessory olfactory bulb but not in mitral cells of the olfactory bulb (Fig. 7)
. CCP5 mRNA is present at a low level throughout the olfactory bulb (Fig. 7)
. CCP6 mRNA is slightly enriched in the mitral cells, and also shows diffuse expression in the granular layer (Fig. 7)
.
|
Expression of CCPs in cell lines
To investigate the subcellular localization of several of the CCPs, we introduced small HA tags into the protein either on the N terminus (for mouse CCP2) or immediately following the CP domain (mouse CCP5 and CCP6). In addition, human CCP1 was included as a control, with the HA tag introduced immediately following the CP domain (Fig. 8
, top). On transient expression, all of the CCPs examined showed strong cytoplasmic expression in both Neuro2A cells (Fig. 8)
and several other cell types (not shown). In addition to the cytoplasmic staining, a small amount of nuclear staining was detected for each of the constructs (Fig. 8)
. The nuclear expression was most pronounced for CCP5 and was especially strong in the NIH3T3 cell line (data not shown).
|
Western blots of cells transfected with CCP1 show a band at 130 kDa, consistent with the size of the full length protein (Fig. 8
, top). No epitope-tagged protein was detected at 40 kDa, the predicted size of the CP domain and attached C-terminal region if the N-terminal domain was cleaved at the YPYTYS sequence (Fig. 8)
. Similary, CCP5 showed a band of 90 kDa, CCP6 a band of 45 kDa, and CCP2 a band of 87 kDa (Fig. 8)
. Except for CCP6, which has a predicted size of 54 kDa, all of these sizes match the predicted size of the intact full-length form. In all cases, there were no detectable signals of the size predicted for the form in which the N-terminal domain was cleaved from the CP domain in either Neuro2A cells (Fig. 8)
or NIH3T3 cells (data not shown). To investigate the possibility that cleavage at the YPYTY sequence is dependent on a specific stimulus, cells transfected with various epitope-tagged CCPs were treated with a calcium ionophore, a phorbol ester, or forskolin, or subjected to serum deprivation or heat shock. Western blot analysis showed only full-length forms after these treatments (Fig. 8
, and data not shown).
Studies on pcd mice
The hypothesis that Nna1/CCP1 functions as a tubulin tyrosine CP was tested by examining the forms of tubulin in pcd mice. For these studies, we used the 3J mutant, which was previously found to result from the deletion of exons 6–8 and the introduction of a stop codon prior to the CP domain (28)
. Previous studies characterized the cell loss of the original pcd mutant, renamed the pcd1J mouse, and found that although Purkinje cells died within several weeks of birth, mitral cell degeneration occurred slowly over 1 yr (29
, 30)
. Because we found that mitral cells express predominantly Nna1/CCP1, with low or undetectable levels of CCP2–6, this cell type provided the ideal test of the role of Nna1/CCP1 in tubulin processing. First, it was necessary to examine whether mitral cells survived for several months in the pcd3J mutant, as reported for the pcd1J mutant (29
, 30)
. Histochemical analysis showed the presence of mitral cells in olfactory bulb of pcd3J mice that appeared generally similar to WT mice (Fig. 9
). Sections of olfactory bulb were stained with an antiserum to CPD; this protein was previously found to be expressed in mitral cells (47)
. The pcd3J mice showed CPD-expressing mitral cells at 10 wk, but by 14 wk there was some degeneration of these CPD-expressing cells (Fig. 9)
. Thus, the severity of the 3J line is greater than that of the 1J mutant, as predicted from the complete absence of Nna1/CCP1 encoding the active CP in the 3J line (28)
.
|
WT mice contain both Tyr- and Glu-tubulin in the mitral cells (Fig. 10
, top). In other cell layers of WT mice, Tyr tubulin is more abundant in the central granular layer than in the glomerular layer, whereas Glu tubulin is more abundant in the glomerular layer than in the granular layer (Fig. 10)
. The olfactory bulb of pcd mice also shows higher levels of Tyr tubulin in the granular layer and higher levels of Glu tubulin in the glomerular layer (Fig. 10
, bottom). However, whereas the mitral cells show abundant expression of Tyr tubulin, these cells do not contain detectable levels of Glu tubulin (Fig. 10
, bottom). This is consistent with the hypothesis that Nna1/CCP1 is a tubulin tyrosine CP.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, which is published as an accompanying paper. Together, these two studies support the hypothesis that Nna1/CCP1 and the related CCP genes encode active enzymes that function in the processing of cytosolic proteins such as tubulin. Modeling and analysis of the residues in the putative substrate binding pocket suggests that Nna1/CCP1, CCP2, CCP3, CCP4, and CCP5 have similar specificities toward bulky C-terminal residues, such as hydrophobic and/or basic amino acids. CCP6 either has a similar specificity as the other CCPs or else it cleaves acidic residues; modeling could not resolve these two possibilities.
All of the CCPs found in the present study lack an N-terminal signal peptide sequence, in contrast to the previously characterized CPs of the A/B and N/E subfamilies, which function either in the secretory pathway or following secretion. The cytosolic distribution of CCP2, 5, and 6 observed in the present study is consistent with the absence of a signal peptide. Previously, human Nna1/CCP1 was reported to show a cytosolic and nuclear distribution when expressed in primary cortical neurons as a fusion protein with GFP (26)
. Using HA-tagged Nna1/CCP1, we confirmed the cytosolic expression but not the nuclear expression; the difference may be due to the smaller HA epitope tag or the cell types used in each study. In the present study, several cell lines were used and all showed Nna1/CCP1 to be predominantly localized to the cytoplasm. However, CCP5 showed a nuclear-like distribution in each cell line tested, with some lines showing a more pronounced nuclear staining than others. The deduced amino acid sequence of CCP5 contains several stretches of three or more basic residues in the C-terminal region; these may function as nuclear targeting signals.
In addition to the CP domain that is conserved in all mouse CCPs and all other members of the M14 CP family, there is also an N-terminal domain immediately adjacent to the CP domain that is highly conserved among mouse CCPs but not with other M14 subfamily CPs. The other subfamilies of M14 CPs each contain a conserved domain of about the same length (100 amino acids) that functions in the folding of the CP domain, although there is no sequence similarity between subfamilies. In the case of the A/B subfamily of M14 CPs, the N-terminal pro domain functions both in folding of the CP domain and in maintaining the enzyme in an inactive state until cleavage by an endopeptidase. N/E subfamily members do not have this pro domain and instead have a domain immediately to the C-terminus of the CP domain; this C-terminal domain has structural but not amino acid sequence homology to transthyretin. Unlike A/B CPs, the transthyretin-like domain of N/E CPs does not need to be cleaved before the enzyme is fully active. Although the CCPs appear to be more similar to the A/B subfamily by the presence of an N-terminal conserved domain and the absence of a transthyretin-like C-terminal domain, the amino acid sequence similarity is too low between the pro region of CPA or CPB and the N-terminal domain of CCPs to permit modeling. Also, it is not clear if the N-terminal domain of the CCPs is removed like the pro domain of A/B CPs or if it remains attached like N/E CPs. In the present study, CCP proteins were not found to be processed when expressed in Neuro2A or NIH3T3 cells. In the accompanying paper, Rodriguez de la Vega et al. (35)
show that a C. elegans CCP homologue is enzymatically active toward small synthetic substrates with the N-terminal region attached to the CP domain. Thus, it appears that activity toward small peptides does not require cleavage of the CCP, although it is possible that cleavage could increase the CP activity or alter the substrate specificity. If the CCPs are cleaved between the conserved N-terminal domain and the CP domain, the cleavage site is presumably within the YPYTY sequence, which is one of the most highly conserved motifs within the different mouse CCPs and is also present in CCP homologs from many diverse organisms (35)
. Alternatively, if the YPYTY sequence at the junction of the N-terminal domain and the CP domain is not a cleavage site, this sequence may play a role in the regulation of the protein, perhaps by allowing the N-terminal domain to move away from the active site and allow substrates to enter. Mouse Nna1/CCP1 was proposed to contain an ATP-GTP binding motif (26)
. Although this sequence is not conserved in other members of the mouse CCP family or in a C. elegans homologue, the cleavage of small peptide substrates by the C. elegans enzyme is stimulated by ATP or ADP (35)
. This stimulation may reflect the movement of an inhibitory domain away from the active site.
There are several potential roles for cytosolic CPs. It is possible that these enzymes function in protein degradation, after the breakdown of proteins by proteasomes. The proteasomes typically degrade proteins into small peptides, and further hydrolysis is required to generate free amino acids that can be recycled into proteins. The CCP family of enzymes may play this important role. Alternatively, the CCPs may be involved in the selective processing of proteins, such as the removal of Tyr and Glu from the C terminus of alpha tubulin. Our finding that the mitral cells of pcd mice have high levels of alpha tubulin containing the C-terminal Tyr and low levels of the detyrosinylated form is consistent with the proposal that Nna1/CCP1 is a tubulin tyrosine CP. CCPs 2–5 may also contribute to tubulin processing; this could explain the observation that nonmitral cells in the olfactory bulb of pcd mice contain Glu-tubulin despite the absence of Nna1/CCP1. These other cell types express CCP2–5, as do the non-Purkinje cells in the cerebellum. Because mitral and Purkinje cells have high levels of Nna1/CCP1 but low or undetectable levels of CCP2–5, these other enzymes would not be available to compensate for the defective Nna1/CCP1. Although mitral and Purkinje cells do have CCP6, this protein may be unable to compensate for the absence of CCP1 in the pcd mice if the substrate binding pocket of CCP6 has an Arg in position 255; this would confer specificity for acidic C-terminal residues. Further studies are needed to examine the enzymatic properties of the various mouse CCPs. In the accompanying paper, the enzymatic properties of a related Nna1-like protein from C. elegans were examined and the substrate specificity was found to be quite broad, suggesting a role in processing a number of proteins or peptides (35)
. It is conceivable that the CCPs process a number of proteins by removing C-terminal amino acids, either native, or added by a ligase. The ligase that adds Tyr to the C-terminus of tubulin is a member of a gene family with 12 other members (48)
. One of these has been found to be responsible for the addition of polyglutamates to proteins (48)
. The functions of the other members of the ligase gene family are not presently known, and it is possible that they add a variety of amino acids to different cytosolic proteins. The CCPs would thus function in the removal of these residues, analogous to the kinase/phosphatase system that regulates the function of many cytosolic proteins. Importance of the amino acid addition/removal cycle is evident from the finding that mice lacking the Tyr ligase die as neonates. Although pcd mice are viable, this is presumably due to the redundancy of the CCPs, as discussed above, and cell types with predominantly Nna1/CCP1 and low levels of CCP2–6 degenerate within 3–14 wk of age. Further studies to explore the role of the CCPs can be accomplished by comparative proteomics of WT and pcd mice; these studies are currently in progress.
|
ACKNOWLEDGMENTS |
|---|
Received for publication September 25, 2006. Accepted for publication October 25, 2006.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  X. Zhang, F. Tan, Y. Zhang, and R. A. Skidgel Carboxypeptidase M and Kinin B1 Receptors Interact to Facilitate Efficient B1 Signaling from B2 Agonists J. Biol. Chem., March 21, 2008; 283(12): 7994 - 8004. [Abstract] [Full Text] [PDF]
|
