- Email: [email protected]
Localization of N-Terminal Sequences in Human AMP Deaminase Isoforms That Influence Contractile Protein Binding
Biochemical and Biophysical Research Communications 285, 489 – 495 (2001) doi:10.1006/bbrc.2001.5180, available online at http://www.idealibrary.com on
Localization of N-Terminal Sequences in Human AMP Deaminase Isoforms That Influence Contractile Protein Binding Donna K. Mahnke-Zizelman and Richard L. Sabina 1 Department of Biochemistry, Medical College of Wisconsin, Milwaukee, Wisconsin 53226
Received June 6, 2001
The reversible association of AMP deaminase (AMPD, EC 3.5.4.6) with elements of the contractile apparatus is an identified mechanism of enzyme regulation in mammalian skeletal muscle. All three members of the human AMPD multigene family contain coding information for polypeptides with divergent N-terminal and conserved C-terminal domains. In this study, serial N-terminal deletion mutants of up to 111 (AMPD1), 214 (AMPD2), and 126 (AMPD3) residues have been constructed without significant alteration of catalytic function or protein solubility. The entire sets of active enzymes are used to extend our understanding of the contractile protein binding of AMPD. Analysis of the most truncated active enzymes demonstrates that all three isoforms can associate with skeletal muscle actomyosin and suggests that a primary binding domain is located within the C-terminal 635– 640 residues of each polypeptide. However, discrete stretches of N-terminal sequence alter this behavior. Residues 54 – 83 in the AMPD1 polypeptide contribute to a high actomyosin binding capacity of both isoform M spliceoforms, although the exon 2ⴚ enzyme exhibits significantly greater association compared to its exon 2ⴙ counterpart. Conversely, residues 129 –183 in the AMPD2 polypeptide reduce actomyosin binding of isoform L. In addition, residues 1– 48 in the AMPD3 polypeptide dramatically suppress contractile protein binding of isoform E, thus allowing this enzyme to participate in other intracellular interactions. © 2001 Academic Press Key Words: AMP deaminase isoforms; contractile protein binding; N-terminal sequence.
AMP deaminase (AMPD, EC 3.5.4.6) is a highly regulated enzyme catalyzing a branchpoint reaction in the adenylate catabolic pathway. AMPD competes with cy1 To whom correspondence and reprint requests should be addressed at Department of Biochemistry, Medical College of Wisconsin, 8701 Watertown Plank Road, Milwaukee, WI 53226. Fax: 414456-6510. E-mail: [email protected].
tosolic 5⬘-nucleotidase (AMP-preferring) for available AMP, thus serving to regulate the intracellular production of adenosine, an important signaling compound in a variety of physiological settings. Three AMPD isoforms are encoded by transcripts produced from a multigene family in mammalian species (1–10), and alignments across predicted primary amino acid sequences reveal conserved C-terminal and divergent N-terminal domains (5, 6). Conversely, alignments across rat and human sequences demonstrate cross-species conservation of the entire AMPD1 (3) and AMPD3 (10) polypeptide, suggesting an operative selective pressure to maintain isoform-specific divergent sequences. Each mammalian AMPD gene produces multiple transcripts that differ at, or near, their 5⬘-ends (5, 10 –12), and confer additional variation to each divergent N-terminal region. Consequently, gene-specific divergent sequences are variable in length and consist of 200 –300 residues. Each AMPD isoform exhibits different physical, catalytic, and regulatory properties (13–16). Putative catalytic residues reside in the conserved C-terminal region of the polypeptide, for example, the consecutive aspartate residues contained within the AMPD signature sequence, SLSTDDP. Evidence in support of catalytic aspartate residues in AMPD are the similar role they assume in the adenosine deaminase signature sequence, SLNTDDP (17, 18), and the reported lack of catalytic activity in a rat AMPD1 recombinant enzyme with a glycine substitution at the second aspartate residue (19). Conversely, it has been more difficult to ascribe functional significance to divergent sequences in mammalian AMPD isoforms, particularly extreme N-terminal residues, due to limited proteolysis that typically accompanies purification and subsequent storage of the enzyme (20, 21). Deletions of nearly 100 residues from the AMPD1 and AMPD3 polypeptides produce enzymes that retain robust catalytic activity (16), demonstrating that extreme N-terminal sequence is dispensable to the functional conformation of these two isoforms.
489
0006-291X/01 $35.00 Copyright © 2001 by Academic Press All rights of reproduction in any form reserved.
Vol. 285, No. 2, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
Although not essential for catalytic activity, limited data indicate that extreme N-terminal sequences play a role in determining isoform-specific properties of the enzyme. For example, the N-terminal 95 amino acids of the rabbit muscle AMPD1 polypeptide facilitate ATP regulation under mildly acidic conditions typical of exercising skeletal muscle (22). In addition, corresponding regions in the human AMPD1 and AMPD3 polypeptides dramatically influence the contractile protein binding properties of these activities (16, 23). This study has been designed to further our understanding of isoform specificity and the structural basis for contractile protein binding of human AMPD isoforms. MATERIALS AND METHODS Materials. Grace’s insect cell culture medium and fetal bovine serum were purchased from Life Technologies, Inc. Phosphocellulose resin (P-11) was supplied by Whatman Ltd. Disposable glass columns (1.5 ⫻ 30 cm), a protein assay kit, and a silver stain kit were purchased from Bio-Rad. A BCA protein assay kit was acquired from Pierce. All biochemicals and other reagents were of the highest quality available. Construction of baculoviral expression plasmids. AMPD1 wild type (exon 2⫹), ⌬L96AMPD1, AMPD3 wild type (1b), and ⌬M90AMPD3 constructs have been described previously (16). The ⌬E129AMPD2 enzyme was produced from a human AMPD2 cDNA (6), which was subsequently discovered to contain intron 2 sequence at its 5⬘ end encoding the anomalous N-terminal extension, MLTFLPSPQ (12). The 5⬘ end of the AMPD1 wild-type (exon 2⫺) cDNA was synthesized using RT-PCR from poly(A) ⫹ mRNA prepared from approximately 500 g of pooled total cellular RNA isolated from available frozen adult human heart samples. The resultant PCR product and the previously described wild type (exon 2⫹) cDNA was digested with BamHI and BglII and the former subcloned into the existing 3⬘ end of the latter. All other constructs used in this study were generated by oligonucleotide-directed mutagenesis of existing cDNAs by PCR. Primers were designed to create new 5⬘ ends encoding each N-truncated polypeptide. In some cases, methionine initiation codons were created that resulted in a single residue substitution (AMPD1, ⌬L84M; AMPD2, ⌬L215M; AMPD3, ⌬L20M, ⌬I49M, and ⌬E65M). In other cases, the introduction of an NcoI restriction endonuclease site, which contains an ATG sequence, resulted in a substitution at the second residue (⌬M54AMPD1, Q55E; ⌬D184AMPD2, F185E; ⌬M127AMPD3, P128A). Oligonucleotides previously designed to sequence wild-type cDNAs were used as 3⬘primers, and each was selected based on a location 3⬘ to a unique restriction site. This facilitated the subsequent cassette subcloning of the PCR product into the existing 3⬘ end of the wild-type sequence. All subcloned PCR products were sequenced in order to ensure fidelity of the mutagenic procedure. AMP deaminase assay. AMP deaminase assays were performed and analyzed by anion-exchange HPLC as described previously (16). Units are defined as micromoles of AMP deaminated per minute. Expression and purification of AMPD recombinant enzymes. Human AMPD recombinant enzymes were expressed in Spodoptera frugiperda (Sf 9) insect cells, 96 –120 h postinfection with recombinant baculovirus (m.o.i. ⫽ 5) as previously described (14). Cells (6 –12 confluent 185-mm 3 flasks) were harvested in ice cold extraction buffer [20 mM potassium phosphate, pH 6.7 (AMPD1 and AMPD2) or 7.0 (AMPD3), containing 100 mM potassium chloride and 0.1% -mercaptoethanol], and disrupted by sonication. Sonicates were clarified by centrifugation at 10,500g for 10 min at 4°C. Crude
extracts were batch adsorbed to phosphocellulose resin equilibrated in extraction buffer for 30 min at 4°C with rotation. Adsorbed resin was washed twice with ice-cold extraction buffer and poured into a prechilled 1.5 ⫻ 30-cm glass column. Protein was eluted with a 100-ml linear gradient of 0.1–2.0 M potassium chloride. Fractions (2 ml) were collected and assayed for AMPD activity and protein. Aliquots of these partially purified preparations were subsequently used for the analysis of catalytic behaviors and actomyosin-binding capacities. With notable exceptions, recombinant enzymes were further purified. Peak fractions from several phosphocellulose preparations were batch adsorbed a second time to freshly prepared resin equilibrated in 20 mM potassium phosphate, pH 6.7 (AMPD1 and AMPD2) or 7.0 (AMPD3), containing 200 mM potassium chloride and 0.1% -mercaptoethanol. Protein was then eluted with a 100 ml linear gradient of 0.02– 0.45 M potassium phosphate. Fractions were collected and assayed for AMPD activity and protein. Peak fractions were concentrated by ammonium sulfate precipitation [26 –30% (w/v) cut] and resuspended in 50 mM imidazole hydrochloride, pH 6.5 (AMPD1 and AMPD2) or 7.0 (AMPD3), containing 500 mM potassium chloride. Resuspended enzyme was diluted 1:1 with 95% glycerol and frozen at ⫺80°C until further use. Polyacrylamide gel electrophoresis. Approximately 0.9 –1.8 U of each human AMPD recombinant enzyme was fractionated by SDS– polyacrylamide gel electrophoresis (SDS–PAGE) as previously described (16). Nine percent SDS–polyacrylamide gels were fixed and stained with either Coomassie blue or silver nitrate. Western blot analysis. Selected N-truncated human AMPD recombinant enzymes were fractionated by SDS–PAGE, electroblotted onto nitrocellulose membranes, and incubated with rabbit polyclonal antisera raised against immobilized human AMPD1 and AMPD3 recombinant enzymes as previously described (16). Rabbit polyclonal antiserum was raised against the ⌬E129AMPD2 recombinant enzyme for use in this study. Actomyosin binding assay. Actomyosin binding capacities of human AMPD recombinant enzymes were analyzed as previously described (16). Briefly, 0.3 U of each AMPD recombinant enzyme was mixed at room temperature with 60 g of purified adult rat skeletal muscle actomyosin prepared free of endogenous AMPD activity. The mixtures were diluted to ⬍100 mM salt concentration and actomyosin was precipitated by centrifugation. Pellets were washed twice and resuspended in high salt buffer and assayed for AMPD activity. Parallel mixtures containing ␥-globulin in place of actomyosin served as nonspecific protein controls and routinely generated ⬍4% of added AMPD activity in resuspended pellets. Computer-assisted statistical analysis. Statistical analysis was accomplished using the Instat program software.
RESULTS Expression and Purification of Wild-Type and N-Truncated Human AMPD Recombinant Enzymes All enzymes were expressed from modified baculoviral genomes in infected Sf 9 cells and isolated by phosphocellulose chromatography. With the exception of the largest N-terminal deletion mutants (⌬M151AMPD1, ⌬M255AMPD2, and ⌬M160AMPD3), relatively high enzyme activities were achieved with all constructs. However, lower levels of expression were consistently observed with other N-truncated AMPD3 variants compared to the wild type enzyme. A single elution from phosphocellulose resin with a potassium chloride gradient accomplished 21- to 64-
490
Vol. 285, No. 2, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 1. SDS–PAGE analysis of purified human AMP deaminase recombinant enzymes. Approximately 0.9 –1.8 units of each enzyme were fractionated on 9% polyacrylamide gels, fixed, and stained with Coomassie blue. Predicted subunit molecular masses (in kDa) are AMPD1: exon 2⫹ (86.5), exon 2⫺ (86.0), ⌬M54 (80.4), ⌬L84M (76.9), ⌬L96 (75.6), ⌬V112M (73.8); AMPD2: ⌬M55 (94.9), ⌬E129 (88.2), ⌬D184 (80.6), ⌬L215M (77.0); AMPD3: 1b (88.8), ⌬L20M (86.5) ⌬I49M (83.3), ⌬E65M (81.4), ⌬M90 (78.5), ⌬M127 (74.8). Migration of molecular weight standards denoted in kilodaltons (kDa). All enzymes were purified by two separate elutions from a phosphocellulose column followed by ammonium sulfate precipitation. AMPD1 spliceoforms were only partially purified owing to poor recoveries from phosphocellulose resin [subunits identified with an asterisk (*) in these two preparations]. Lower bands in the ⌬M55 (AMPD2), ⌬E129 (AMPD2), and 1b (AMPD3) lanes represent N-terminal proteolytic products of these subunits. The lower band in the ⌬I49M (AMPD3) lane represents a contaminating protein.
fold purifications with recoveries ranging from 69 to 98%. Notable exceptions included both wild type AMPD1 spliceoforms, for which only 9- to 12-fold purifications were obtained in association with low recoveries (3–20%). Poor recoveries from phosphocellulose resin precluded any further purification of these two enzymes. Pooled peak fractions from three to four preparations of all other enzymes were passed over phosphocellulose resin a second time with potassium phosphate gradient elution, followed by ammonium sulfate precipitation. Estimated specific activities of the more purified preparations of N-truncated AMPD1 enzymes ranged from 5000 to 7800 U/mg of protein, N-truncated AMPD2 enzymes from 560 to 760 U/mg of protein, and wild-type and N-truncated AMPD3 enzymes from 1300 to 2039 U/mg of protein. An exception was the ⌬I49M derivative of AMPD3 with a specific activity of 550 U/mg of protein. The relatively low specific activity of the ⌬I49M enzyme was attributed to the persistence of a lower molecular mass contaminant in the purified preparation. SDS–PAGE analysis of all active recombinant enzymes revealed observed subunit molecular masses that were consistent with predicted sizes in each case (Fig. 1). Each inactive human AMPD recombinant protein and the corresponding active variant with the next largest N-terminal deletion were evaluated by Western blot analysis. Results presented in Fig. 2 revealed immunoreactive bands of the appropriate size for the ⌬M151AMPD1, ⌬M255AMPD2, and ⌬M160AMPD3 polypeptides in whole cell extracts. However, unlike their active N-truncated counterparts (⌬V112MAMPD1, ⌬L215MAMPD2, and
⌬M127AMPD3, respectively), these inactive enzymes partitioned predominantly into the insoluble fraction of the extract.
FIG. 2. Western blot analysis of human AMP deaminase recombinant enzymes. Whole cell extracts (W) were prepared by sonication of insect cells infected with recombinant baculovirus containing human AMPD constructs. Each extract was partitioned into soluble (S) and insoluble (I) fractions by centrifugation at 10,000g for 10 min. The insoluble pellet was resuspended in nonreducing denaturing buffer. Samples (W, 5 g of total protein; S, volume equivalent to W; I, load based on that amount lost from each whole cell extract after centrifugation) were fractionated by 9% SDS–PAGE, electroblotted onto nitrocellulose membranes, and incubated with polyclonal antisera raised against recombinant AMPD1 (top), AMPD2 (middle), or AMPD3 (bottom) enzymes. Predicted subunit molecular masses (in kDa) of immunoreactive bands are AMPD1: ⌬V112M (73.8), ⌬M151 (69.3); AMPD2: ⌬L215M (77.0), ⌬M255 (72.4); AMPD3: ⌬M127 (74.8), ⌬M160 (71.1). Enzyme activities (in mU) loaded in W and S are listed.
491
Vol. 285, No. 2, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS TABLE 1
Catalytic Properties of Human AMPD Recombinant Enzymes K m (app), mM ⫺ATP
Enzyme
⫹ATP
AMPD1 2⫹ WT 2⫺ WT ⌬M54 ⌬L84M ⌬L96 ⌬V112M
(4)* (4) (3) (3) (3) (3)
4.1 ⫾ 1.1 3.4 ⫾ 1.0 2.2 ⫾ 0.1# 1.7 ⫾ 0.8# 1.3 ⫾ 0.4# ,@ ,ˆ 1.2 ⫾ 0.63 ,@ ,ˆ
1.3 ⫾ 0.3 2.1 ⫾ 0.4# 1.8 ⫾ 1.0 1.5 ⫾ 0.5 1.6 ⫾ 0.8 1.1 ⫾ 0.2@ n H (⫺ATP)
AMPD2 (1B/2) ⌬M55 ⌬E129 ⌬D184 ⌬L215M
(4)* (3)* (3)* (3)*
18.9 ⫾ 12.8 15.2 ⫾ 2.3 14.1 ⫾ 6.2 10.1 ⫾ 2.6
1.6 ⫾ 0.4 2.0 ⫾ 0.8 1.3 ⫾ 0.2 1.2 ⫾ 0.3
2.0 ⫾ 0.9 2.2 ⫾ 0.2 1.6 ⫾ 0.6 2.0 ⫾ 1.3
AMPD3 1b WT ⌬L20M ⌬I49M ⌬E65M ⌬M90 ⌬M127
(5) (3) (3)* (3)* (3) (5)
1.5 ⫾ 1.1 0.8 ⫾ 0.3 1.1 ⫾ 0.1 0.9 ⫾ 0.2 1.0 ⫾ 0.1 1.3 ⫾ 0.6
0.8 ⫾ 0.2 0.8 ⫾ 0.2 0.7 ⫾ 0.1 0.6 ⫾ 0.1 0.8 ⫾ 0.3 1.3 ⫾ 1.0
in Fig. 3 show that both wild-type AMPD1 exon 2 spliceoforms and the ⌬M54 enzyme exhibited high actomyosin-binding capacities (70 –90% bound, Fig. 3A). However, the exon 2⫺ spliceoform had a significantly higher capacity than the other two enzymes. Further N-terminal truncations of the AMPD1 polypeptide produced enzymes (⌬L84M, ⌬L96, and ⌬V112M) with intermediate actomyosin binding capacities (40 –50% bound) that were each significantly lower than all three larger constructs. N-truncated human AMPD2 recombinant enzymes had intermediate (35– 40% bound) to high (75– 80% bound) actomyosin binding capacities (Fig. 3B), with the ⌬M55 and ⌬E129 constructs both exhibiting significantly lower capacities than either the ⌬D184 or ⌬L215M activities. Finally, N-truncations of the AMPD3 polypeptide produced enzymes with intermediate actomyosin binding capacities (40 –50% bound) that were all significantly higher than the low capacities of the wild-type (9 ⫾ 3% bound) and ⌬L20M (16 ⫾ 5% bound) enzymes (Fig. 3C). However, the ⌬M90 variant (27 ⫾ 4% bound) was significantly lower than that of the other N-truncated AMPD3 enzymes with intermediate actomyosin binding capacities.
* Significant differences (P ⬍ 0.05) between ⫺ATP and ⫹ATP values in a two-tailed t test. # Significant difference (P ⬍ 0.05) compared to the exon 2⫹ wild-type enzyme in a two-tailed t test. @ Significant difference (P ⬍ 0.05) compared to the exon 2⫺ wild-type enzyme in a two-tailed t test. ˆ Significant difference (P ⬍ 0.05) compared to the ⌬M54 enzyme in a two-tailed t test.
Catalytic Properties of Human AMPD Recombinant Enzymes In contrast to the dramatic effects produced by the largest N-terminal deletions, Table 1 shows that lesser truncations had little or no effect on the catalytic behavior of each isoform. For example, a gradual increase in affinity for substrate in the absence of ATP (up to 3.4-fold) was observed with sequential N-terminal truncations of the AMPD1 polypeptide. Consistent with the reported behavior of isoform L isolated from human tissues (13), all N-truncated AMPD2 enzymes exhibited a higher K m in the absence of ATP and were allosterically activated by this nucleotide effector, with Hill coefficients of approximately 2. Actomyosin Binding Capacities of Human AMPD Recombinant Enzymes Actomyosin binding capacities of each human AMPD recombinant enzyme were evaluated. Data presented
FIG. 3. In vitro actomyosin binding capacities of human AMPD recombinant enzymes. Graphical presentation of the percentage of AMPD activity recovered in resuspended pellets following coprecipitation with rat skeletal muscle actomyosin (⫾SD). Number of trials using independent preparations of each enzyme are shown in parentheses. (A) AMPD1; (B) AMPD2; (C) AMPD3. Using two-tailed t tests for comparisons within each gene-specific group of enzymes: *P ⬍ 0.05 compared to all other AMPD1 enzymes, except ⌬M54; §P ⬍ 0.05 compared to all other AMPD1 enzymes; †P ⬍ 0.05 compared to ⌬L84M, ⌬L96, and ⌬V112M; £P ⬍ 0.05 compared to ⌬D184 and ⌬L215M; ¥P ⬍ 0.05 compared to ⌬I49M, ⌬E65M, ⌬M90, and ⌬M127; ¶P ⬍ 0.05 compared to ⌬M90 and ⌬M127; ⌺P ⬍ 0.05 compared to ⌬M90; ⍀P ⬍ 0.05 compared to ⌬M127. Substitution of gammaglobulin for actomyosin in each binding assay typically yields 1– 4% of recovered activities in resuspended pellets.
492
Vol. 285, No. 2, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS TABLE 2
Alignment of the Proposed Myosin Binding Region in the Rat (r) AMPD1 Polypeptide with All Corresponding Human (h) AMP Deaminase Sequences a rAMPD1 hAMPD1 hAMPD2 hAMPD3 rAMPD1 hAMPD1 hAMPD2 hAMPD3 rAMPD1 hAMPD1 hAMPD2 hAMPD3
178 178 293 192 230 230 345 244 282 282 397 296
ESFYPVFTPPPKKGEDPFRREDLPANLGYHLKMKGGVIYIYPDEAAASRDEP ----------V--------TDN--E---------D--V-V--N---V-K--ADAPVHPPALEQHPYEHCEPSTM-GD--LG-R-VR--VHV-TRREPDEHCSE -EGL-D-H--PLPQ---YCLD-A-P--D-LVH-Q--ILFV-DNKKMLEHQ-KPYPYPNLDDFLDDMNFLLALIAQGPVKTYTHRRLKFLSSKFQVHQMLNEMD --L------T-----------------------------------------VEL---D-QE-VA-V-V-M---IN--I-SFCY---QY------M-VL----K HSL---D-ETYTV--SHI----TD--T---C----N--E---SL-E-----S ELKELKNNPHRDFYNCRKVDTHIHAAACMNQKHLLRFIKKSYHIDADRVVYS ------------------------------------------Q----------AAQ-KV-------I---------SS------------RAMKRHLEEI-HV -F----S--------V-----------------------HT-QTEP--T-AE
a Stretches of rat AMPD1 sequence exhibiting homology with extended sequences in titin (31) are overlined. Dashes (-) represent identity to the rat AMPD1 sequence.
DISCUSSION AMPD1 (1, 6) and AMPD3 (7, 10) expression is highest in adult skeletal muscle, where it is well established that enzyme activity can associate with elements of the contractile apparatus (24 –29). Moreover, these interactions appear to be important physiological regulators of catalytic activity (30). However, there are conflicting reports with respect to AMPD isoform specificity of contractile protein binding (25, 26). Data presented in this study further our understanding of the structural basis for these interactions and also address the issue of isoform-specificity. Although extreme N-terminal sequences in each human AMPD polypeptide have only subtle effects on the catalytic and regulatory behaviors of the enzyme, discrete stretches of these divergent residues appear to play more dramatic structural roles in contractile protein binding. For example, two separate stretches of N-terminal sequence in the AMPD1 polypeptide have a significant effect on contractile protein binding of isoform M. (i) Exon 2-encoded residues (8-AEEKQ-12) modestly suppress actomyosin binding capacity. The higher actomyosin binding capacity of the human exon 2⫺ spliceoform is consistent with a previous report showing that the rat enzyme was relatively refractive to a variety of experimental conditions that resulted in lower colocalization with myosin heavy chain (MHC) compared to the exon 2⫹ spliceoform (31). (ii) Residues 54 – 83 significantly enhance actomyosin binding capacity. Relatedly, a rat ⌬N65AMPD1 construct exhibited significantly lower colocalization with MHC compared to either wild-type spliceoform (31). These combined data suggest that a secondary contractile protein binding region is located within residues 54 – 65 of the AMPD1 polypeptide. Acting in concert with the previously identified primary contractile protein binding domain contained within residues 178 –333 (31), these two regions
provide a greater ability to associate with elements of the contractile apparatus during exercise. Conversely, stretches of N-terminal sequence in the AMPD2 and AMPD3 polypeptides suppress actomyosin binding of isoforms L and E, respectively. Residues 129 –183 in the AMPD2 polypeptide are associated with an approximate 50% reduction in actomyosin binding capacity, whereas the N-terminal 48 residues of the AMPD3 polypeptide nearly eliminate contractile protein binding. Further deletions up to residue 215 in the AMPD2 polypeptide and residue 127 in the AMPD3 polypeptide result in activities with high and intermediate actomyosin binding capacities, respectively. Whereas extreme N-terminal sequence in each human AMPD isoform exerts unique modulatory effects on isoform-specific actomyosin binding capacities, the combined data also provide compelling evidence for a conserved contractile protein binding region located within the C-terminal catalytic domain of all AMPD polypeptides. Although dramatic reductions in catalytic activity and protein solubility precluded further N-terminal truncations to map this element, a previous study identified a myosin binding domain within residues 178 –333 in the rat AMPD1 polypeptide (31). Table 1 presents an alignment between residues 178 and 333 in the rat sequence with the corresponding regions in all three human polypeptides. Rat and human AMPD1 polypeptides are 91% identical in this region, whereas less conservation is apparent in the AMPD2 (44% identity, residues 293– 448) and AMPD3 (57% identity, residues 192–347) sequences. These results are not surprising since this stretch of sequence spans the boundary between divergent N-terminal and conserved C-terminal amino acids across AMPD isoforms. The divergent part of this domain is proposed to contain important contractile protein binding elements for the AMPD1 isoform based on homology between small stretches of the rat sequence and several extended
493
Vol. 285, No. 2, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
regions of titin [overlined residues in Table 2 (31)]. However, the homology to titin is apparent only in the AMPD1 polypeptide. Therefore, the hypothesis for a conserved AMPD contractile protein binding domain also predicts the involvement of sequences located in the C-terminal portion of this region where identities are higher. The near complete masking of this putative conserved contractile protein binding site by extreme N-terminal sequence in the AMPD3 polypeptide likely facilitates other intracellular interactions involving isoform E, such as its reported association with membranes (32, 33). Finally, data presented in this study may also be used to explain conflicting reports on AMPD isoform specificity of contractile protein binding. Studies also performed in vitro using contractile proteins and AMPD activities isolated from adult rat tissues have reported AMPD1 isoform specificity in one case (26), and binding of all three isoforms in another (25). Consistent with the first report, human AMPD1 exon 2 spliceoforms and the AMPD3 recombinant enzyme exhibit relative high and low actomyosin binding capacities, respectively. However, contractile protein binding is profoundly influenced by divergent N-terminal sequences in these two AMPD isoforms. Moreover, these regions of the AMPD1 and AMPD3 polypeptides are also sensitive to proteolysis during and after purification (16). Genetically engineered enzymes modeled after N-truncated proteolytic products of these human AMPD recombinant enzymes (⌬L84MAMPD1 and ⌬M90AMPD3) both exhibit intermediate actomyosin binding capacities. Therefore, data generated using N-truncated enzymes are consistent with the second report (25). Consequently, it appears reasonable to conclude that contractile protein binding behaviors exhibited by different AMPD activities purified from mammalian tissues would depend on the relative degree of N-terminal proteolysis in the preparation.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
ACKNOWLEDGMENT This work was supported by Public Health Service Grant DK50902 from the National Institutes of Health.
16.
REFERENCES 17. 1. Sabina, R. L., Marquetant, R., Desai, N. M., Kaletha, K., and Holmes, E. W. (1987) Cloning and sequence of rat myoadenylate deaminase cDNA: Evidence for tissue-specific and developmental regulation. J. Biol. Chem. 262, 12397–12400. 2. Sabina, R. L., Morisaki, T., Clarke P., Eddy, R., Shows, T. B., Morton, C. C., and Holmes, E. W. (1990) Characterization of the human and rat myoadenylate deaminase genes. J. Biol. Chem. 265, 9423–9433. 3. Morisaki, T., Sabina, R. L., and Holmes, E. W. (1990) Adenylate deaminase: A multigene family in humans and rats. J. Biol. Chem. 265, 11482–11486. 4. Sabina, R. L., Fishbein, W. N., Pezeshkpour, G., Clarke, P. R. H.,
18.
19.
20.
494
and Holmes, E. W. (1992) Molecular analysis of the myoadenylate deaminase deficiencies. Neurology 42, 170 –177. Mahnke-Zizelman, D. K., and Sabina, R. L. (1992) Cloning of human AMP deaminase isoform E cDNAs: Evidence for a third AMPD gene exhibiting alternatively spliced 5⬘-exons. J. Biol. Chem. 267, 20866 –20877. Bausch-Jurken, M. T., Mahnke-Zizelman, D. K., Morisaki, T., and Sabina, R. L. (1992) Molecular cloning of AMP deaminase isoform L: Sequence and bacterial expression of human AMPD2 cDNA. J. Biol. Chem. 267, 22407–22413. Mahnke-Zizelman, D. K., Eddy, R., Shows, T. B., and Sabina, R. L. (1996) Characterization of the human AMPD3 gene reveals that 5⬘ exon usage is subject to transcriptional control by three tandem promoters and alternative splicing. Biochim. Biophys. Acta 1306, 75–92. Mahnke-Zizelman, D. K., Van den Bergh, F., Bausch-Jurken, M. T., Eddy, R., Sait, S., Shows, T. B., and Sabina, R. L. (1996) Cloning, sequence and characterization if the human AMPD2 gene: Evidence for transcriptional regulation by two closely spaced promoters. Biochim. Biophys. Acta 1308, 122–132. Wang, X., Morisaki, H., Sermsuvitayawong, K., Mineo, I., Toyama K., Ogasawara, N., Mukai, T., and Morisaki T. (1997) Cloning and expression of cDNA encoding heart-type isoform of AMP deaminase. Gene 188, 285–290. Mahnke-Zizelman, D. K., D’Cunha, J., Wojnar, J. M., Brogley, M. A., and Sabina, R. L. (1997) Regulation of rat AMP deaminase 3 (isoform C) by development and skeletal muscle fibre type. Biochem. J. 326, 521–529. Mineo, I., Clarke, P. R. H., Sabina, R. L., and Holmes, E. W. (1990) A novel pathway for alternative splicing: Identification of an RNA intermediate that generates an alternative 5⬘ splice donor site not present in the primary transcript of AMPD1. Mol. Cell. Biol. 10, 5271–5278. Van den Bergh, F., and Sabina, R. L. (1995) Characterization of human AMPD2 gene expression reveals alternative transcripts encoding variable N-terminal extensions of isoform L. Biochem. J. 312, 401– 410. Ogasawara, N., Goto, H., Yamada, Y., and Watanabe, T. (1978) Distribution of AMP-deaminase isozymes in rat tissues. Eur. J. Biochem. 87, 297–304. Ogasawara, N., Goto, H., Yamada, Y., Watanabe, T., and Asano, T. (1982) AMP deaminase isozymes in human tissues. Biochim. Biophys. Acta 714, 298 –306. Ogasawara, N., Goto, H., and Watanabe, T. (1975) Isozymes of rat brain AMP deaminase: Developmental changes and characterization of five forms. FEBS Lett. 58, 245–248. Mahnke-Zizelman, D. K., Tullson, P. C., and Sabina, R. L. (1998) Novel aspects of tetramer assembly and N-terminal domain structure and function are revealed by recombinant expression of human AMP deaminase isoforms. J. Biol. Chem. 273, 35118 – 35125. Chang, Z., Nygaard, P., Chinault, A. C., and Kellems, R. E. (1991) Deduced amino acid sequence of Escherichia coli adenosine deaminase reveals evolutionarily conserved amino acid residues: Implications for catalytic function. Biochemistry 30, 2273– 2280. Wilson, D. K., Rudolph, F. B., and Quiocho, F. A. (1991) Atomic structure of adenosine deaminase complexed with a transitionstate analog: Understanding catalysis and immunodeficiency mutations. Science 252, 1278 –1284. Gross, M., Morisaki, H., Morisaki, H., and Holmes, E. W. (1994) Identification of functional domains in AMPD1 by mutational analysis. Biochem. Biophys. Res. Commun. 205, 1010 –1017. Ranieri-Raggi, M., and Raggi, A. (1980) Effects of storage on
Vol. 285, No. 2, 2001
21.
22.
23.
24.
25.
26.
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
activity and subunit structure of rabbit skeletal-muscle AMP deaminase. Biochem. J. 189, 367–368. Chilson, O. P., Kelly-Chilson, A. E., and Siegel, N. R. (1997) AMP-deaminases from chicken and rabbit muscle: Partial primary sequences of homologous 17-kDa CNBr fragments: Autorecognition by rabbit anti-[chicken AMPD]. Comp. Biochem. Physiol. 116B, 371–377. Ronca, F., Ranieri-Raggi, M., Brown, P. E., Moir, A. J. G., and Raggi, A. (1994) Evidence for species-differentiated regulatory domain within the N-terminal region of skeletal muscle AMP deaminase. Biochim. Biophys. Acta 1209, 123–129. Sims, B., Mahnke-Zizelman, D. K., Profit, A. A., Prestwich, G. D., Sabina, R. L., and Theibert, A. B. (1999) Regulation of AMP deaminase by phosphoinositides. J. Biol. Chem. 274, 25701– 25707. Ashby, B., and Frieden, C. (1977) Interaction of AMPaminohydrolase with myosin and its subfragments. J. Biol. Chem. 252, 1869 –1872. Ogasawara, N., Goto, H., and Yamada, Y. (1978) Effects of various ligands on interaction of AMP deaminase with myosin. Biochim. Biophys. Acta 524, 442– 446. Shiraki, H., Ogawa, H., Matsuda, Y., and Nakagawa, H. (1979) Interaction of rat muscle AMP deaminase with myosin. I. Biochemical study of the interaction of AMP deaminase and myosin in rat muscle. Biochim. Biophys. Acta 566, 335–344.
27. Rundell, K. W., Tullson, P. C., and Terjung, R. L. (1992) AMP deaminase binding in contracting rat skeletal muscle. Am. J. Physiol. 263, C287–C293. 28. Koretz, J. F., Irving, T. C., and Wang, K. (1993) Filamentous aggregates of native titin and binding of C-protein and AMPdeaminase. Arch. Biochem. Biophys. 304, 305–309. 29. Soteriou, A., Gamage, M., and Trinick, J. (1993) A survey of interactions made by the giant protein titin. J. Cell Sci. 104, 119 –123. 30. Rundell, K. W., Tullson, P. C., and Terjung, R. L. (1992) Altered kinetics of AMP deaminase by myosin binding. Am. J. Physiol. 263, C294 –C299. 31. Hisatome, I., Morisaki, T., Kamma, H., Sugama, T., Morisaki, H., Ohtahara, A., and Holmes, E. W. (1998) Control of AMP deaminase I binding to myosin heavy chain. Am. J. Physiol. 275, C870 –C881. 32. Rao, S. N., Hana, L., and Askari, A. (1968) Alkali cation activated AMP deaminase of erythrocytes: Some properties of the membrane bound enzyme. Biochim. Biophys. Acta 151, 651– 654. 33. Pipoly, G. M., Nathans, G. R., Chang, D., and Deuel, T. F. (1979) Regulation of the interaction of purified human erythrocyte AMP deaminase and the human erythrocyte membrane. J. Clin. Invest. 63, 1066 –1076.
495
Localization of N-Terminal Sequences in Human AMP Deaminase Isoforms That Influence Contractile Protein Binding Donna K. Mahnke-Zizelman and Richard L. Sabina 1 Department of Biochemistry, Medical College of Wisconsin, Milwaukee, Wisconsin 53226
Received June 6, 2001
The reversible association of AMP deaminase (AMPD, EC 3.5.4.6) with elements of the contractile apparatus is an identified mechanism of enzyme regulation in mammalian skeletal muscle. All three members of the human AMPD multigene family contain coding information for polypeptides with divergent N-terminal and conserved C-terminal domains. In this study, serial N-terminal deletion mutants of up to 111 (AMPD1), 214 (AMPD2), and 126 (AMPD3) residues have been constructed without significant alteration of catalytic function or protein solubility. The entire sets of active enzymes are used to extend our understanding of the contractile protein binding of AMPD. Analysis of the most truncated active enzymes demonstrates that all three isoforms can associate with skeletal muscle actomyosin and suggests that a primary binding domain is located within the C-terminal 635– 640 residues of each polypeptide. However, discrete stretches of N-terminal sequence alter this behavior. Residues 54 – 83 in the AMPD1 polypeptide contribute to a high actomyosin binding capacity of both isoform M spliceoforms, although the exon 2ⴚ enzyme exhibits significantly greater association compared to its exon 2ⴙ counterpart. Conversely, residues 129 –183 in the AMPD2 polypeptide reduce actomyosin binding of isoform L. In addition, residues 1– 48 in the AMPD3 polypeptide dramatically suppress contractile protein binding of isoform E, thus allowing this enzyme to participate in other intracellular interactions. © 2001 Academic Press Key Words: AMP deaminase isoforms; contractile protein binding; N-terminal sequence.
AMP deaminase (AMPD, EC 3.5.4.6) is a highly regulated enzyme catalyzing a branchpoint reaction in the adenylate catabolic pathway. AMPD competes with cy1 To whom correspondence and reprint requests should be addressed at Department of Biochemistry, Medical College of Wisconsin, 8701 Watertown Plank Road, Milwaukee, WI 53226. Fax: 414456-6510. E-mail: [email protected].
tosolic 5⬘-nucleotidase (AMP-preferring) for available AMP, thus serving to regulate the intracellular production of adenosine, an important signaling compound in a variety of physiological settings. Three AMPD isoforms are encoded by transcripts produced from a multigene family in mammalian species (1–10), and alignments across predicted primary amino acid sequences reveal conserved C-terminal and divergent N-terminal domains (5, 6). Conversely, alignments across rat and human sequences demonstrate cross-species conservation of the entire AMPD1 (3) and AMPD3 (10) polypeptide, suggesting an operative selective pressure to maintain isoform-specific divergent sequences. Each mammalian AMPD gene produces multiple transcripts that differ at, or near, their 5⬘-ends (5, 10 –12), and confer additional variation to each divergent N-terminal region. Consequently, gene-specific divergent sequences are variable in length and consist of 200 –300 residues. Each AMPD isoform exhibits different physical, catalytic, and regulatory properties (13–16). Putative catalytic residues reside in the conserved C-terminal region of the polypeptide, for example, the consecutive aspartate residues contained within the AMPD signature sequence, SLSTDDP. Evidence in support of catalytic aspartate residues in AMPD are the similar role they assume in the adenosine deaminase signature sequence, SLNTDDP (17, 18), and the reported lack of catalytic activity in a rat AMPD1 recombinant enzyme with a glycine substitution at the second aspartate residue (19). Conversely, it has been more difficult to ascribe functional significance to divergent sequences in mammalian AMPD isoforms, particularly extreme N-terminal residues, due to limited proteolysis that typically accompanies purification and subsequent storage of the enzyme (20, 21). Deletions of nearly 100 residues from the AMPD1 and AMPD3 polypeptides produce enzymes that retain robust catalytic activity (16), demonstrating that extreme N-terminal sequence is dispensable to the functional conformation of these two isoforms.
489
0006-291X/01 $35.00 Copyright © 2001 by Academic Press All rights of reproduction in any form reserved.
Vol. 285, No. 2, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
Although not essential for catalytic activity, limited data indicate that extreme N-terminal sequences play a role in determining isoform-specific properties of the enzyme. For example, the N-terminal 95 amino acids of the rabbit muscle AMPD1 polypeptide facilitate ATP regulation under mildly acidic conditions typical of exercising skeletal muscle (22). In addition, corresponding regions in the human AMPD1 and AMPD3 polypeptides dramatically influence the contractile protein binding properties of these activities (16, 23). This study has been designed to further our understanding of isoform specificity and the structural basis for contractile protein binding of human AMPD isoforms. MATERIALS AND METHODS Materials. Grace’s insect cell culture medium and fetal bovine serum were purchased from Life Technologies, Inc. Phosphocellulose resin (P-11) was supplied by Whatman Ltd. Disposable glass columns (1.5 ⫻ 30 cm), a protein assay kit, and a silver stain kit were purchased from Bio-Rad. A BCA protein assay kit was acquired from Pierce. All biochemicals and other reagents were of the highest quality available. Construction of baculoviral expression plasmids. AMPD1 wild type (exon 2⫹), ⌬L96AMPD1, AMPD3 wild type (1b), and ⌬M90AMPD3 constructs have been described previously (16). The ⌬E129AMPD2 enzyme was produced from a human AMPD2 cDNA (6), which was subsequently discovered to contain intron 2 sequence at its 5⬘ end encoding the anomalous N-terminal extension, MLTFLPSPQ (12). The 5⬘ end of the AMPD1 wild-type (exon 2⫺) cDNA was synthesized using RT-PCR from poly(A) ⫹ mRNA prepared from approximately 500 g of pooled total cellular RNA isolated from available frozen adult human heart samples. The resultant PCR product and the previously described wild type (exon 2⫹) cDNA was digested with BamHI and BglII and the former subcloned into the existing 3⬘ end of the latter. All other constructs used in this study were generated by oligonucleotide-directed mutagenesis of existing cDNAs by PCR. Primers were designed to create new 5⬘ ends encoding each N-truncated polypeptide. In some cases, methionine initiation codons were created that resulted in a single residue substitution (AMPD1, ⌬L84M; AMPD2, ⌬L215M; AMPD3, ⌬L20M, ⌬I49M, and ⌬E65M). In other cases, the introduction of an NcoI restriction endonuclease site, which contains an ATG sequence, resulted in a substitution at the second residue (⌬M54AMPD1, Q55E; ⌬D184AMPD2, F185E; ⌬M127AMPD3, P128A). Oligonucleotides previously designed to sequence wild-type cDNAs were used as 3⬘primers, and each was selected based on a location 3⬘ to a unique restriction site. This facilitated the subsequent cassette subcloning of the PCR product into the existing 3⬘ end of the wild-type sequence. All subcloned PCR products were sequenced in order to ensure fidelity of the mutagenic procedure. AMP deaminase assay. AMP deaminase assays were performed and analyzed by anion-exchange HPLC as described previously (16). Units are defined as micromoles of AMP deaminated per minute. Expression and purification of AMPD recombinant enzymes. Human AMPD recombinant enzymes were expressed in Spodoptera frugiperda (Sf 9) insect cells, 96 –120 h postinfection with recombinant baculovirus (m.o.i. ⫽ 5) as previously described (14). Cells (6 –12 confluent 185-mm 3 flasks) were harvested in ice cold extraction buffer [20 mM potassium phosphate, pH 6.7 (AMPD1 and AMPD2) or 7.0 (AMPD3), containing 100 mM potassium chloride and 0.1% -mercaptoethanol], and disrupted by sonication. Sonicates were clarified by centrifugation at 10,500g for 10 min at 4°C. Crude
extracts were batch adsorbed to phosphocellulose resin equilibrated in extraction buffer for 30 min at 4°C with rotation. Adsorbed resin was washed twice with ice-cold extraction buffer and poured into a prechilled 1.5 ⫻ 30-cm glass column. Protein was eluted with a 100-ml linear gradient of 0.1–2.0 M potassium chloride. Fractions (2 ml) were collected and assayed for AMPD activity and protein. Aliquots of these partially purified preparations were subsequently used for the analysis of catalytic behaviors and actomyosin-binding capacities. With notable exceptions, recombinant enzymes were further purified. Peak fractions from several phosphocellulose preparations were batch adsorbed a second time to freshly prepared resin equilibrated in 20 mM potassium phosphate, pH 6.7 (AMPD1 and AMPD2) or 7.0 (AMPD3), containing 200 mM potassium chloride and 0.1% -mercaptoethanol. Protein was then eluted with a 100 ml linear gradient of 0.02– 0.45 M potassium phosphate. Fractions were collected and assayed for AMPD activity and protein. Peak fractions were concentrated by ammonium sulfate precipitation [26 –30% (w/v) cut] and resuspended in 50 mM imidazole hydrochloride, pH 6.5 (AMPD1 and AMPD2) or 7.0 (AMPD3), containing 500 mM potassium chloride. Resuspended enzyme was diluted 1:1 with 95% glycerol and frozen at ⫺80°C until further use. Polyacrylamide gel electrophoresis. Approximately 0.9 –1.8 U of each human AMPD recombinant enzyme was fractionated by SDS– polyacrylamide gel electrophoresis (SDS–PAGE) as previously described (16). Nine percent SDS–polyacrylamide gels were fixed and stained with either Coomassie blue or silver nitrate. Western blot analysis. Selected N-truncated human AMPD recombinant enzymes were fractionated by SDS–PAGE, electroblotted onto nitrocellulose membranes, and incubated with rabbit polyclonal antisera raised against immobilized human AMPD1 and AMPD3 recombinant enzymes as previously described (16). Rabbit polyclonal antiserum was raised against the ⌬E129AMPD2 recombinant enzyme for use in this study. Actomyosin binding assay. Actomyosin binding capacities of human AMPD recombinant enzymes were analyzed as previously described (16). Briefly, 0.3 U of each AMPD recombinant enzyme was mixed at room temperature with 60 g of purified adult rat skeletal muscle actomyosin prepared free of endogenous AMPD activity. The mixtures were diluted to ⬍100 mM salt concentration and actomyosin was precipitated by centrifugation. Pellets were washed twice and resuspended in high salt buffer and assayed for AMPD activity. Parallel mixtures containing ␥-globulin in place of actomyosin served as nonspecific protein controls and routinely generated ⬍4% of added AMPD activity in resuspended pellets. Computer-assisted statistical analysis. Statistical analysis was accomplished using the Instat program software.
RESULTS Expression and Purification of Wild-Type and N-Truncated Human AMPD Recombinant Enzymes All enzymes were expressed from modified baculoviral genomes in infected Sf 9 cells and isolated by phosphocellulose chromatography. With the exception of the largest N-terminal deletion mutants (⌬M151AMPD1, ⌬M255AMPD2, and ⌬M160AMPD3), relatively high enzyme activities were achieved with all constructs. However, lower levels of expression were consistently observed with other N-truncated AMPD3 variants compared to the wild type enzyme. A single elution from phosphocellulose resin with a potassium chloride gradient accomplished 21- to 64-
490
Vol. 285, No. 2, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 1. SDS–PAGE analysis of purified human AMP deaminase recombinant enzymes. Approximately 0.9 –1.8 units of each enzyme were fractionated on 9% polyacrylamide gels, fixed, and stained with Coomassie blue. Predicted subunit molecular masses (in kDa) are AMPD1: exon 2⫹ (86.5), exon 2⫺ (86.0), ⌬M54 (80.4), ⌬L84M (76.9), ⌬L96 (75.6), ⌬V112M (73.8); AMPD2: ⌬M55 (94.9), ⌬E129 (88.2), ⌬D184 (80.6), ⌬L215M (77.0); AMPD3: 1b (88.8), ⌬L20M (86.5) ⌬I49M (83.3), ⌬E65M (81.4), ⌬M90 (78.5), ⌬M127 (74.8). Migration of molecular weight standards denoted in kilodaltons (kDa). All enzymes were purified by two separate elutions from a phosphocellulose column followed by ammonium sulfate precipitation. AMPD1 spliceoforms were only partially purified owing to poor recoveries from phosphocellulose resin [subunits identified with an asterisk (*) in these two preparations]. Lower bands in the ⌬M55 (AMPD2), ⌬E129 (AMPD2), and 1b (AMPD3) lanes represent N-terminal proteolytic products of these subunits. The lower band in the ⌬I49M (AMPD3) lane represents a contaminating protein.
fold purifications with recoveries ranging from 69 to 98%. Notable exceptions included both wild type AMPD1 spliceoforms, for which only 9- to 12-fold purifications were obtained in association with low recoveries (3–20%). Poor recoveries from phosphocellulose resin precluded any further purification of these two enzymes. Pooled peak fractions from three to four preparations of all other enzymes were passed over phosphocellulose resin a second time with potassium phosphate gradient elution, followed by ammonium sulfate precipitation. Estimated specific activities of the more purified preparations of N-truncated AMPD1 enzymes ranged from 5000 to 7800 U/mg of protein, N-truncated AMPD2 enzymes from 560 to 760 U/mg of protein, and wild-type and N-truncated AMPD3 enzymes from 1300 to 2039 U/mg of protein. An exception was the ⌬I49M derivative of AMPD3 with a specific activity of 550 U/mg of protein. The relatively low specific activity of the ⌬I49M enzyme was attributed to the persistence of a lower molecular mass contaminant in the purified preparation. SDS–PAGE analysis of all active recombinant enzymes revealed observed subunit molecular masses that were consistent with predicted sizes in each case (Fig. 1). Each inactive human AMPD recombinant protein and the corresponding active variant with the next largest N-terminal deletion were evaluated by Western blot analysis. Results presented in Fig. 2 revealed immunoreactive bands of the appropriate size for the ⌬M151AMPD1, ⌬M255AMPD2, and ⌬M160AMPD3 polypeptides in whole cell extracts. However, unlike their active N-truncated counterparts (⌬V112MAMPD1, ⌬L215MAMPD2, and
⌬M127AMPD3, respectively), these inactive enzymes partitioned predominantly into the insoluble fraction of the extract.
FIG. 2. Western blot analysis of human AMP deaminase recombinant enzymes. Whole cell extracts (W) were prepared by sonication of insect cells infected with recombinant baculovirus containing human AMPD constructs. Each extract was partitioned into soluble (S) and insoluble (I) fractions by centrifugation at 10,000g for 10 min. The insoluble pellet was resuspended in nonreducing denaturing buffer. Samples (W, 5 g of total protein; S, volume equivalent to W; I, load based on that amount lost from each whole cell extract after centrifugation) were fractionated by 9% SDS–PAGE, electroblotted onto nitrocellulose membranes, and incubated with polyclonal antisera raised against recombinant AMPD1 (top), AMPD2 (middle), or AMPD3 (bottom) enzymes. Predicted subunit molecular masses (in kDa) of immunoreactive bands are AMPD1: ⌬V112M (73.8), ⌬M151 (69.3); AMPD2: ⌬L215M (77.0), ⌬M255 (72.4); AMPD3: ⌬M127 (74.8), ⌬M160 (71.1). Enzyme activities (in mU) loaded in W and S are listed.
491
Vol. 285, No. 2, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS TABLE 1
Catalytic Properties of Human AMPD Recombinant Enzymes K m (app), mM ⫺ATP
Enzyme
⫹ATP
AMPD1 2⫹ WT 2⫺ WT ⌬M54 ⌬L84M ⌬L96 ⌬V112M
(4)* (4) (3) (3) (3) (3)
4.1 ⫾ 1.1 3.4 ⫾ 1.0 2.2 ⫾ 0.1# 1.7 ⫾ 0.8# 1.3 ⫾ 0.4# ,@ ,ˆ 1.2 ⫾ 0.63 ,@ ,ˆ
1.3 ⫾ 0.3 2.1 ⫾ 0.4# 1.8 ⫾ 1.0 1.5 ⫾ 0.5 1.6 ⫾ 0.8 1.1 ⫾ 0.2@ n H (⫺ATP)
AMPD2 (1B/2) ⌬M55 ⌬E129 ⌬D184 ⌬L215M
(4)* (3)* (3)* (3)*
18.9 ⫾ 12.8 15.2 ⫾ 2.3 14.1 ⫾ 6.2 10.1 ⫾ 2.6
1.6 ⫾ 0.4 2.0 ⫾ 0.8 1.3 ⫾ 0.2 1.2 ⫾ 0.3
2.0 ⫾ 0.9 2.2 ⫾ 0.2 1.6 ⫾ 0.6 2.0 ⫾ 1.3
AMPD3 1b WT ⌬L20M ⌬I49M ⌬E65M ⌬M90 ⌬M127
(5) (3) (3)* (3)* (3) (5)
1.5 ⫾ 1.1 0.8 ⫾ 0.3 1.1 ⫾ 0.1 0.9 ⫾ 0.2 1.0 ⫾ 0.1 1.3 ⫾ 0.6
0.8 ⫾ 0.2 0.8 ⫾ 0.2 0.7 ⫾ 0.1 0.6 ⫾ 0.1 0.8 ⫾ 0.3 1.3 ⫾ 1.0
in Fig. 3 show that both wild-type AMPD1 exon 2 spliceoforms and the ⌬M54 enzyme exhibited high actomyosin-binding capacities (70 –90% bound, Fig. 3A). However, the exon 2⫺ spliceoform had a significantly higher capacity than the other two enzymes. Further N-terminal truncations of the AMPD1 polypeptide produced enzymes (⌬L84M, ⌬L96, and ⌬V112M) with intermediate actomyosin binding capacities (40 –50% bound) that were each significantly lower than all three larger constructs. N-truncated human AMPD2 recombinant enzymes had intermediate (35– 40% bound) to high (75– 80% bound) actomyosin binding capacities (Fig. 3B), with the ⌬M55 and ⌬E129 constructs both exhibiting significantly lower capacities than either the ⌬D184 or ⌬L215M activities. Finally, N-truncations of the AMPD3 polypeptide produced enzymes with intermediate actomyosin binding capacities (40 –50% bound) that were all significantly higher than the low capacities of the wild-type (9 ⫾ 3% bound) and ⌬L20M (16 ⫾ 5% bound) enzymes (Fig. 3C). However, the ⌬M90 variant (27 ⫾ 4% bound) was significantly lower than that of the other N-truncated AMPD3 enzymes with intermediate actomyosin binding capacities.
* Significant differences (P ⬍ 0.05) between ⫺ATP and ⫹ATP values in a two-tailed t test. # Significant difference (P ⬍ 0.05) compared to the exon 2⫹ wild-type enzyme in a two-tailed t test. @ Significant difference (P ⬍ 0.05) compared to the exon 2⫺ wild-type enzyme in a two-tailed t test. ˆ Significant difference (P ⬍ 0.05) compared to the ⌬M54 enzyme in a two-tailed t test.
Catalytic Properties of Human AMPD Recombinant Enzymes In contrast to the dramatic effects produced by the largest N-terminal deletions, Table 1 shows that lesser truncations had little or no effect on the catalytic behavior of each isoform. For example, a gradual increase in affinity for substrate in the absence of ATP (up to 3.4-fold) was observed with sequential N-terminal truncations of the AMPD1 polypeptide. Consistent with the reported behavior of isoform L isolated from human tissues (13), all N-truncated AMPD2 enzymes exhibited a higher K m in the absence of ATP and were allosterically activated by this nucleotide effector, with Hill coefficients of approximately 2. Actomyosin Binding Capacities of Human AMPD Recombinant Enzymes Actomyosin binding capacities of each human AMPD recombinant enzyme were evaluated. Data presented
FIG. 3. In vitro actomyosin binding capacities of human AMPD recombinant enzymes. Graphical presentation of the percentage of AMPD activity recovered in resuspended pellets following coprecipitation with rat skeletal muscle actomyosin (⫾SD). Number of trials using independent preparations of each enzyme are shown in parentheses. (A) AMPD1; (B) AMPD2; (C) AMPD3. Using two-tailed t tests for comparisons within each gene-specific group of enzymes: *P ⬍ 0.05 compared to all other AMPD1 enzymes, except ⌬M54; §P ⬍ 0.05 compared to all other AMPD1 enzymes; †P ⬍ 0.05 compared to ⌬L84M, ⌬L96, and ⌬V112M; £P ⬍ 0.05 compared to ⌬D184 and ⌬L215M; ¥P ⬍ 0.05 compared to ⌬I49M, ⌬E65M, ⌬M90, and ⌬M127; ¶P ⬍ 0.05 compared to ⌬M90 and ⌬M127; ⌺P ⬍ 0.05 compared to ⌬M90; ⍀P ⬍ 0.05 compared to ⌬M127. Substitution of gammaglobulin for actomyosin in each binding assay typically yields 1– 4% of recovered activities in resuspended pellets.
492
Vol. 285, No. 2, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS TABLE 2
Alignment of the Proposed Myosin Binding Region in the Rat (r) AMPD1 Polypeptide with All Corresponding Human (h) AMP Deaminase Sequences a rAMPD1 hAMPD1 hAMPD2 hAMPD3 rAMPD1 hAMPD1 hAMPD2 hAMPD3 rAMPD1 hAMPD1 hAMPD2 hAMPD3
178 178 293 192 230 230 345 244 282 282 397 296
ESFYPVFTPPPKKGEDPFRREDLPANLGYHLKMKGGVIYIYPDEAAASRDEP ----------V--------TDN--E---------D--V-V--N---V-K--ADAPVHPPALEQHPYEHCEPSTM-GD--LG-R-VR--VHV-TRREPDEHCSE -EGL-D-H--PLPQ---YCLD-A-P--D-LVH-Q--ILFV-DNKKMLEHQ-KPYPYPNLDDFLDDMNFLLALIAQGPVKTYTHRRLKFLSSKFQVHQMLNEMD --L------T-----------------------------------------VEL---D-QE-VA-V-V-M---IN--I-SFCY---QY------M-VL----K HSL---D-ETYTV--SHI----TD--T---C----N--E---SL-E-----S ELKELKNNPHRDFYNCRKVDTHIHAAACMNQKHLLRFIKKSYHIDADRVVYS ------------------------------------------Q----------AAQ-KV-------I---------SS------------RAMKRHLEEI-HV -F----S--------V-----------------------HT-QTEP--T-AE
a Stretches of rat AMPD1 sequence exhibiting homology with extended sequences in titin (31) are overlined. Dashes (-) represent identity to the rat AMPD1 sequence.
DISCUSSION AMPD1 (1, 6) and AMPD3 (7, 10) expression is highest in adult skeletal muscle, where it is well established that enzyme activity can associate with elements of the contractile apparatus (24 –29). Moreover, these interactions appear to be important physiological regulators of catalytic activity (30). However, there are conflicting reports with respect to AMPD isoform specificity of contractile protein binding (25, 26). Data presented in this study further our understanding of the structural basis for these interactions and also address the issue of isoform-specificity. Although extreme N-terminal sequences in each human AMPD polypeptide have only subtle effects on the catalytic and regulatory behaviors of the enzyme, discrete stretches of these divergent residues appear to play more dramatic structural roles in contractile protein binding. For example, two separate stretches of N-terminal sequence in the AMPD1 polypeptide have a significant effect on contractile protein binding of isoform M. (i) Exon 2-encoded residues (8-AEEKQ-12) modestly suppress actomyosin binding capacity. The higher actomyosin binding capacity of the human exon 2⫺ spliceoform is consistent with a previous report showing that the rat enzyme was relatively refractive to a variety of experimental conditions that resulted in lower colocalization with myosin heavy chain (MHC) compared to the exon 2⫹ spliceoform (31). (ii) Residues 54 – 83 significantly enhance actomyosin binding capacity. Relatedly, a rat ⌬N65AMPD1 construct exhibited significantly lower colocalization with MHC compared to either wild-type spliceoform (31). These combined data suggest that a secondary contractile protein binding region is located within residues 54 – 65 of the AMPD1 polypeptide. Acting in concert with the previously identified primary contractile protein binding domain contained within residues 178 –333 (31), these two regions
provide a greater ability to associate with elements of the contractile apparatus during exercise. Conversely, stretches of N-terminal sequence in the AMPD2 and AMPD3 polypeptides suppress actomyosin binding of isoforms L and E, respectively. Residues 129 –183 in the AMPD2 polypeptide are associated with an approximate 50% reduction in actomyosin binding capacity, whereas the N-terminal 48 residues of the AMPD3 polypeptide nearly eliminate contractile protein binding. Further deletions up to residue 215 in the AMPD2 polypeptide and residue 127 in the AMPD3 polypeptide result in activities with high and intermediate actomyosin binding capacities, respectively. Whereas extreme N-terminal sequence in each human AMPD isoform exerts unique modulatory effects on isoform-specific actomyosin binding capacities, the combined data also provide compelling evidence for a conserved contractile protein binding region located within the C-terminal catalytic domain of all AMPD polypeptides. Although dramatic reductions in catalytic activity and protein solubility precluded further N-terminal truncations to map this element, a previous study identified a myosin binding domain within residues 178 –333 in the rat AMPD1 polypeptide (31). Table 1 presents an alignment between residues 178 and 333 in the rat sequence with the corresponding regions in all three human polypeptides. Rat and human AMPD1 polypeptides are 91% identical in this region, whereas less conservation is apparent in the AMPD2 (44% identity, residues 293– 448) and AMPD3 (57% identity, residues 192–347) sequences. These results are not surprising since this stretch of sequence spans the boundary between divergent N-terminal and conserved C-terminal amino acids across AMPD isoforms. The divergent part of this domain is proposed to contain important contractile protein binding elements for the AMPD1 isoform based on homology between small stretches of the rat sequence and several extended
493
Vol. 285, No. 2, 2001
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
regions of titin [overlined residues in Table 2 (31)]. However, the homology to titin is apparent only in the AMPD1 polypeptide. Therefore, the hypothesis for a conserved AMPD contractile protein binding domain also predicts the involvement of sequences located in the C-terminal portion of this region where identities are higher. The near complete masking of this putative conserved contractile protein binding site by extreme N-terminal sequence in the AMPD3 polypeptide likely facilitates other intracellular interactions involving isoform E, such as its reported association with membranes (32, 33). Finally, data presented in this study may also be used to explain conflicting reports on AMPD isoform specificity of contractile protein binding. Studies also performed in vitro using contractile proteins and AMPD activities isolated from adult rat tissues have reported AMPD1 isoform specificity in one case (26), and binding of all three isoforms in another (25). Consistent with the first report, human AMPD1 exon 2 spliceoforms and the AMPD3 recombinant enzyme exhibit relative high and low actomyosin binding capacities, respectively. However, contractile protein binding is profoundly influenced by divergent N-terminal sequences in these two AMPD isoforms. Moreover, these regions of the AMPD1 and AMPD3 polypeptides are also sensitive to proteolysis during and after purification (16). Genetically engineered enzymes modeled after N-truncated proteolytic products of these human AMPD recombinant enzymes (⌬L84MAMPD1 and ⌬M90AMPD3) both exhibit intermediate actomyosin binding capacities. Therefore, data generated using N-truncated enzymes are consistent with the second report (25). Consequently, it appears reasonable to conclude that contractile protein binding behaviors exhibited by different AMPD activities purified from mammalian tissues would depend on the relative degree of N-terminal proteolysis in the preparation.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
ACKNOWLEDGMENT This work was supported by Public Health Service Grant DK50902 from the National Institutes of Health.
16.
REFERENCES 17. 1. Sabina, R. L., Marquetant, R., Desai, N. M., Kaletha, K., and Holmes, E. W. (1987) Cloning and sequence of rat myoadenylate deaminase cDNA: Evidence for tissue-specific and developmental regulation. J. Biol. Chem. 262, 12397–12400. 2. Sabina, R. L., Morisaki, T., Clarke P., Eddy, R., Shows, T. B., Morton, C. C., and Holmes, E. W. (1990) Characterization of the human and rat myoadenylate deaminase genes. J. Biol. Chem. 265, 9423–9433. 3. Morisaki, T., Sabina, R. L., and Holmes, E. W. (1990) Adenylate deaminase: A multigene family in humans and rats. J. Biol. Chem. 265, 11482–11486. 4. Sabina, R. L., Fishbein, W. N., Pezeshkpour, G., Clarke, P. R. H.,
18.
19.
20.
494
and Holmes, E. W. (1992) Molecular analysis of the myoadenylate deaminase deficiencies. Neurology 42, 170 –177. Mahnke-Zizelman, D. K., and Sabina, R. L. (1992) Cloning of human AMP deaminase isoform E cDNAs: Evidence for a third AMPD gene exhibiting alternatively spliced 5⬘-exons. J. Biol. Chem. 267, 20866 –20877. Bausch-Jurken, M. T., Mahnke-Zizelman, D. K., Morisaki, T., and Sabina, R. L. (1992) Molecular cloning of AMP deaminase isoform L: Sequence and bacterial expression of human AMPD2 cDNA. J. Biol. Chem. 267, 22407–22413. Mahnke-Zizelman, D. K., Eddy, R., Shows, T. B., and Sabina, R. L. (1996) Characterization of the human AMPD3 gene reveals that 5⬘ exon usage is subject to transcriptional control by three tandem promoters and alternative splicing. Biochim. Biophys. Acta 1306, 75–92. Mahnke-Zizelman, D. K., Van den Bergh, F., Bausch-Jurken, M. T., Eddy, R., Sait, S., Shows, T. B., and Sabina, R. L. (1996) Cloning, sequence and characterization if the human AMPD2 gene: Evidence for transcriptional regulation by two closely spaced promoters. Biochim. Biophys. Acta 1308, 122–132. Wang, X., Morisaki, H., Sermsuvitayawong, K., Mineo, I., Toyama K., Ogasawara, N., Mukai, T., and Morisaki T. (1997) Cloning and expression of cDNA encoding heart-type isoform of AMP deaminase. Gene 188, 285–290. Mahnke-Zizelman, D. K., D’Cunha, J., Wojnar, J. M., Brogley, M. A., and Sabina, R. L. (1997) Regulation of rat AMP deaminase 3 (isoform C) by development and skeletal muscle fibre type. Biochem. J. 326, 521–529. Mineo, I., Clarke, P. R. H., Sabina, R. L., and Holmes, E. W. (1990) A novel pathway for alternative splicing: Identification of an RNA intermediate that generates an alternative 5⬘ splice donor site not present in the primary transcript of AMPD1. Mol. Cell. Biol. 10, 5271–5278. Van den Bergh, F., and Sabina, R. L. (1995) Characterization of human AMPD2 gene expression reveals alternative transcripts encoding variable N-terminal extensions of isoform L. Biochem. J. 312, 401– 410. Ogasawara, N., Goto, H., Yamada, Y., and Watanabe, T. (1978) Distribution of AMP-deaminase isozymes in rat tissues. Eur. J. Biochem. 87, 297–304. Ogasawara, N., Goto, H., Yamada, Y., Watanabe, T., and Asano, T. (1982) AMP deaminase isozymes in human tissues. Biochim. Biophys. Acta 714, 298 –306. Ogasawara, N., Goto, H., and Watanabe, T. (1975) Isozymes of rat brain AMP deaminase: Developmental changes and characterization of five forms. FEBS Lett. 58, 245–248. Mahnke-Zizelman, D. K., Tullson, P. C., and Sabina, R. L. (1998) Novel aspects of tetramer assembly and N-terminal domain structure and function are revealed by recombinant expression of human AMP deaminase isoforms. J. Biol. Chem. 273, 35118 – 35125. Chang, Z., Nygaard, P., Chinault, A. C., and Kellems, R. E. (1991) Deduced amino acid sequence of Escherichia coli adenosine deaminase reveals evolutionarily conserved amino acid residues: Implications for catalytic function. Biochemistry 30, 2273– 2280. Wilson, D. K., Rudolph, F. B., and Quiocho, F. A. (1991) Atomic structure of adenosine deaminase complexed with a transitionstate analog: Understanding catalysis and immunodeficiency mutations. Science 252, 1278 –1284. Gross, M., Morisaki, H., Morisaki, H., and Holmes, E. W. (1994) Identification of functional domains in AMPD1 by mutational analysis. Biochem. Biophys. Res. Commun. 205, 1010 –1017. Ranieri-Raggi, M., and Raggi, A. (1980) Effects of storage on
Vol. 285, No. 2, 2001
21.
22.
23.
24.
25.
26.
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
activity and subunit structure of rabbit skeletal-muscle AMP deaminase. Biochem. J. 189, 367–368. Chilson, O. P., Kelly-Chilson, A. E., and Siegel, N. R. (1997) AMP-deaminases from chicken and rabbit muscle: Partial primary sequences of homologous 17-kDa CNBr fragments: Autorecognition by rabbit anti-[chicken AMPD]. Comp. Biochem. Physiol. 116B, 371–377. Ronca, F., Ranieri-Raggi, M., Brown, P. E., Moir, A. J. G., and Raggi, A. (1994) Evidence for species-differentiated regulatory domain within the N-terminal region of skeletal muscle AMP deaminase. Biochim. Biophys. Acta 1209, 123–129. Sims, B., Mahnke-Zizelman, D. K., Profit, A. A., Prestwich, G. D., Sabina, R. L., and Theibert, A. B. (1999) Regulation of AMP deaminase by phosphoinositides. J. Biol. Chem. 274, 25701– 25707. Ashby, B., and Frieden, C. (1977) Interaction of AMPaminohydrolase with myosin and its subfragments. J. Biol. Chem. 252, 1869 –1872. Ogasawara, N., Goto, H., and Yamada, Y. (1978) Effects of various ligands on interaction of AMP deaminase with myosin. Biochim. Biophys. Acta 524, 442– 446. Shiraki, H., Ogawa, H., Matsuda, Y., and Nakagawa, H. (1979) Interaction of rat muscle AMP deaminase with myosin. I. Biochemical study of the interaction of AMP deaminase and myosin in rat muscle. Biochim. Biophys. Acta 566, 335–344.
27. Rundell, K. W., Tullson, P. C., and Terjung, R. L. (1992) AMP deaminase binding in contracting rat skeletal muscle. Am. J. Physiol. 263, C287–C293. 28. Koretz, J. F., Irving, T. C., and Wang, K. (1993) Filamentous aggregates of native titin and binding of C-protein and AMPdeaminase. Arch. Biochem. Biophys. 304, 305–309. 29. Soteriou, A., Gamage, M., and Trinick, J. (1993) A survey of interactions made by the giant protein titin. J. Cell Sci. 104, 119 –123. 30. Rundell, K. W., Tullson, P. C., and Terjung, R. L. (1992) Altered kinetics of AMP deaminase by myosin binding. Am. J. Physiol. 263, C294 –C299. 31. Hisatome, I., Morisaki, T., Kamma, H., Sugama, T., Morisaki, H., Ohtahara, A., and Holmes, E. W. (1998) Control of AMP deaminase I binding to myosin heavy chain. Am. J. Physiol. 275, C870 –C881. 32. Rao, S. N., Hana, L., and Askari, A. (1968) Alkali cation activated AMP deaminase of erythrocytes: Some properties of the membrane bound enzyme. Biochim. Biophys. Acta 151, 651– 654. 33. Pipoly, G. M., Nathans, G. R., Chang, D., and Deuel, T. F. (1979) Regulation of the interaction of purified human erythrocyte AMP deaminase and the human erythrocyte membrane. J. Clin. Invest. 63, 1066 –1076.
495