Search tips
Search criteria 


Logo of pursigspringer.comThis journalToc AlertsSubmit OnlineOpen Choice
Purinergic Signal. 2010 September; 6(3): 327–337.
Published online 2010 March 30. doi:  10.1007/s11302-010-9180-0
PMCID: PMC2947659

Proline residues link the active site to transmembrane domain movements in human nucleoside triphosphate diphosphohydrolase 3 (NTPDase3)


The active sites of the membrane-bound nucleoside triphosphate diphosphohydrolases (NTPDases) regulate and are regulated by coordinated and spatially distant movements of their transmembrane helices, modulating enzyme activity, and substrate specificity. Using site-directed mutagenesis, the roles of the conserved proline residues (N-terminal: P52 and P53; C-terminal: P472, P476, P481, P484, and P485) of human NTPDase3, located in the “linker regions” that connect the N- and C-terminal transmembrane helices with the extracellular active site, were examined. Single cysteine substitutions were strategically placed in the transmembrane domain (N-terminal helix: V42C; C-terminal helix: G489C) to serve as cross-linking “sensors” of helical interactions. These “sensor” background mutant proteins (V42C and G489C NTPDase3) are enzymatically active and are cross-linked by copper phenanthroline less efficiently in the presence of adenosine triphosphate (ATP). Proline to alanine substitutions at P53, P481, P484, and P485 in the V42C background, as well as P53, P481, and P484 in the G489C background, exhibited decreased nucleotidase activities. More importantly, alanine substitutions at P53 and P481 in the V42C background and P481 in the G489C background no longer exhibited the ATP-induced decrease in transmembrane cross-linking efficiency. Interestingly, the P485A mutation abolished oxidative cross-linking at G489C both in the presence and absence of ATP. Taken together, these results suggest a role for proline residues 53 and 481 in the linker regions of human NTPDase3 for coupling nucleotide binding at the enzyme active site to movements and/or rearrangements of the transmembrane helices necessary for optimal nucleotide hydrolysis.

Keywords: Ecto-nucleotidase, NTPDase3, Linker region, Conserved proline residues, Site-directed mutagenesis, Transmembrane cross-linking


The nucleoside triphosphate diphosphohydrolases (NTPDases) are a family of nucleotidases, some of which regulate purinergic signaling by divalent cation-dependent hydrolysis of nucleotides acting as agonists at purinergic receptors. In humans, there are six membrane-bound enzymes (NTPDases 1–4, 7, and 8), whose carboxy- and amino-terminal ends are each anchored to the membrane [1, 2], and two enzymes that lack a carboxy-terminal transmembrane (TM) helix (NTPDase5 and NTPDase6), which can be secreted as soluble enzymes after cleavage of their respective N-terminal signal sequences [3, 4]. Due to the scarcity of specific inhibitors and genetically modified animals, the functions of most individual NTPDases are poorly understood and still under investigation. However, the NTPDases are implicated to play a role in many biological and physiological processes, including secretion [5], cell adhesion [6], cancer and malignant transformation [7, 8], adenosine recycling, and platelet aggregation [9].

Membrane-bound NTPDase3 is expressed on the cell surface and has a large extracellular domain, N- and C-terminal “linker regions,” and a transmembrane domain (TMD) consisting of an N- and a C-terminal TM helix [10]. The extracellular domain contains five disulfide bonds that are conserved among all cell surface NTPDases, which when disrupted in NTPDase3 yield five distinct biochemical phenotypes [2]. There are also five extracellular apyrase conserved regions (ACRs) shared among all members of the NTPDase family [11]. Based on sequence homology of the NTPDases with the actin/heat shock protein/sugar kinase superfamily of proteins [12], ACR1 and ACR4 are recognized as phosphate-binding domains of the extracellular active site. The roles and importance of the ACRs and specific amino acids contained in those regions were confirmed and refined by the recent publication of crystal structures of the extracellular portion of rat NTPDase2, both with and without a co-crystallized non-hydrolyzable nucleotide analogue [13].

The NTPDase3 “linker regions” (N-terminal: 46HKQEVLPPGLK56; C-terminal: 472PAESPLIRLPIEPP485) are stretches of residues in the amino acid sequence between the extracellular ends of the N- and C-terminal TM helices and ACR 1 and ACR5 in the extracellular lobes (location of the enzyme active site). These linker regions contain several conserved proline residues (N-terminal: P52 and P53; C-terminal: P472, P476, P481, P484, and P485—see Fig. 1a). Proline is structurally unique among the 20 protein-forming amino acids since its side-chain binds to the backbone amide position resulting in a distinctive cyclic structure. This replaces the amide proton with a –CH2 group, limiting proline to act only as a hydrogen bond acceptor. The bulkiness of the N–CH2 group places restrictions on the conformation of the residue preceding proline [14], thereby destabilizing α-helix secondary structures [15, 16]. The elimination of the hydrogen bond donor, together with the bulkiness of the side-chain, produces the well-known “helix-breaker” (and β-sheet breaker) secondary structure effect of proline residues. In addition, the ring structure of proline also locks the available backbone ø dihedral angle at approximately 65°, resulting in a conformational rigidity unique to proline [17, 18]. Thus, proline residues confer unique structural features and rigidity needed for some conformational changes in proteins.

Fig. 1
Rationale and experimental approach. a Multiple sequence alignments for portions of NTPDase sequences between the N- and C-terminal transmembrane (TM) helices and the extracellular domain containing the active site. Conserved proline residues found in ...

Previous studies have demonstrated that interactions between the extracellular and TM domains exist in NTPDases. One of the earliest hints of this phenomenon was the loss of activity caused by detergents during enzyme purification. Almost every detergent traditionally used to solubilize membrane-bound enzymes inhibits the enzymatic activity of many NTPDases [9]. Later, it was shown that both TM helices are required for full enzymatic activity, and the nucleotide specificity of the NTPDases is altered by the lack of, or manipulation of, either TM helix, further suggesting interplay between the active site and the TMD [1922]. Grinthal and Guidotti expanded this work by demonstrating that the TM helices of rat CD39/NTPDase1 undergo dynamic motions that reflect their functional relationship with the enzyme active site upon substrate binding [23]. Specifically, in the absence of substrate (adenosine triphosphate (ATP) or adenosine diphosphate (ADP)), oxidative cross-linking of single cysteine residues substituted into the TM helices of rat CD39/NTPDase1 was found to be efficient, if the residues were located near the extracellular surface of either the N- or C-terminal TM helices. In addition, Grinthal and Guidotti [23] observed that ATP greatly attenuated dimer formation via oxidation of cysteine mutations of alanine 34 (corresponding to valine 42 in human NTPDase3) on the N-terminal TM helix and serine 481 (corresponding to glycine 489 in human NTPDase3) on the C-terminal TM helix (see Fig. 1a). Therefore, V42 and G489 in human NTPDase3 were chosen for mutation to cysteine to serve as TMD “sensor” residues to monitor the effect of ATP binding in the extracellular domain on changes in the cross-linking efficiency of the TM helices.

In the present study, we investigated the hypothesis that the conserved, linker region proline residues provide the rigidity necessary to facilitate the structural linkage and/or “coupling” of the presumed movements of the active site lobes (active site lobe movements presumed by analogy to other members of the actin superfamily of nucleotide-handling enzymes [12]) to movements of the TM helices. Site-directed mutagenesis was utilized to introduce cysteine residues at strategic locations (V42C and G498C mutations) near the extracellular ends of the N- and C-terminal TM helices that were used as “sensors” to determine differences in TM helix interactions assessed by changes in oxidative cross-linking efficiency in the presence and absence of substrate. Then, the conserved proline residues in the linker regions were singly mutated to alanine, and the resultant mutants were assessed for enzymatic activity and TM cross-linking efficiency in the presence and absence of substrate. Our data demonstrates that alanine substitution of proline 53 and 481 results in a decrease in enzymatic activity and abolishment of the ATP-induced decrease in TM cross-linking efficiency. This result suggests that P53 and P481 are involved in the structural framework used to link events and movements that occur in the enzyme active site with movements of the TM helices.

Materials and methods

Site-directed mutagenesis of NTPDase3

Mutagenesis of human NTPDase3 in the pcDNA3 vector was performed using a QuikChange site-directed mutagenesis kit (Stratagene) as previously described [24]. The University of Cincinnati DNA Core Facility produced the synthetic oligonucleotide primers needed to engineer each mutant. The sense primers used for mutagenesis are as follows:


The altered codons are bolded and underlined (the complimentary anti-sense oligonucleotides, also needed for mutagenesis, are not shown). The mutants were made in the wild-type-like “free sulfhydryl-less” C10S/C501S/C509S NTPDase3 background to eliminate all other free sulfhydryls [25]. The resulting cDNA constructs were sequenced by the University of Cincinnati DNA Core Facility to verify the presence of the desired mutation and the absence of any unwanted changes.

Transient transfection and preparation of COS-1 cell membranes

COS-1 cells were grown in Dulbecco's modified Eagle's medium (DMEM) with 10% bovine serum and a 2% mixture of antibiotics and antimycotics (Invitrogen). The cells were transfected with 4 μg plasmid DNA per 100 mm cell culture plate using Lipofectamine and PLUS reagents (Invitrogen) as previously described [26]. Cells were transfected with 4 μg empty pcDNA3 vector as a control. The COS-1 cells were harvested ~48 h post-transfection. The crude cell membrane preparations were obtained as previously described [26].

Protein assay

Protein concentrations were determined using the Bio-Rad protein assay reagent with the modifications of Stoscheck [27]. Bovine serum albumin was used as a standard.

Nucleotidase assay

Nucleotidase activities were determined by measuring the amount of inorganic phosphate (Pi) released from nucleotide substrates (Sigma) at 37°C using modifications of the technique Fiske and Subbarow [28] as previously described [24]. Either 5 mM CaCl2 or 5 mM MgCl2 buffer (both in 20 mM 3-(N-morpholino)-propanesulfonic acid (MOPS), pH 7.1 buffer) was used for cation-dependent nucleotidase activity. The reactions were initiated by the addition of nucleotide to a final concentration of 2.5 mM in the 0.3-ml assay solutions. Hydrolysis was allowed to proceed for 30 min or 1 h, depending on the substrate used. The activities were corrected for pcDNA3 vector/COS-1 cell background (membranes from COS-1 cells transfected with an empty vector), as well as differences in expression levels as determined by quantitative Western blotting of each sample (i.e., the activities determined in micromoles per milligram COS protein/hour were divided by the relative NTPDase3 expression level of each mutant (see Table 1), and this corrected activity is used for comparison throughout this study).

Table 1
Expression levels and nucleotidase activities of human nucleoside triphosphate diphosphohydrolase 3 (NTPDase3) mutants used in this study

SDS-PAGE and Western blotting

Pre-cast 10-well or 15-well 4–15% gradient mini-gels (Bio-Rad) were used to resolve aliquots of crude membrane proteins (0.5–2 μg, depending on the sample and the experimental purpose), usually after incubating for 10 min at 60°C in non-reducing sodium dodecyl sulfate (SDS) loading buffer. Following SDS-polyacrylamide gel electrophoresis (PAGE), the proteins were electrotransferred to Immun-Blot polyvinylidene fluoride (PVDF) membrane (Bio-Rad) for 3 h at 33 V in cold 10 mM CAPS/NaOH, pH 11. After transfer, the PVDF membrane was incubated for 1 h in blocking solution (5% nonfat dry milk in Tris-buffered saline (TBS)) at room temperature (22°C). Next, the PVDF membrane was incubated overnight at room temperature in blocking solution containing 0.02% sodium azide and a 1:5,000 dilution of rabbit polyclonal primary antisera of KLH1, generated against the cytoplasmic C-terminal peptide (amino acid residues 515–529) of human NTPDase3 [29]. After washing the blot in a TBS 0.05% Tween 20 solution, a goat anti-rabbit horseradish peroxidase-conjugated secondary antibody (Pierce) was applied at a 1:4,000 dilution for 1 h at room temperature followed by washing and application of Pierce SuperSignal West Dura Extended Duration Substrate to the PVDF membrane for 5 min to detect immunoreactivity by chemiluminescence. Chemiluminescence was recorded and quantified using a FluorChem IS-8800 system (Alpha Innotech). Quantification of cross-linking efficiency was calculated by dividing the NTPDase3 monomer band densities in copper phenanthroline (CuPhen)-treated samples by the monomer band densities in non-cross-linked control samples from the same mutant processed in parallel.

Oxidative cross-linking in the presence or absence of substrate

CuPhen was the oxidative cross-linker used and was prepared fresh each day, just prior to use, by combining cupric sulfate and 1,10-phenanthroline (Sigma) at a 1:3 ratio in 20% ethanol [30]. Cysteine-substituted protein from COS-1 cell crude membranes was diluted in 50 mM Tris–HCl, pH 7.5 to a final concentration of 0.1 mg/ml and oxidized with a final concentration of 5 mM CuPhen in the absence or presence of a final concentration of 5 mM ATP (Sigma), added just before the CuPhen, at 37°C for 5 min in a total volume of 25 μl. The reaction was stopped by adding an equal volume of non-reducing SDS loading buffer containing 20 mM N-ethylmaleimide (NEM) and 20 mM EDTA. The samples were then loaded on a gel for SDS-PAGE and Western blotting as described above.

Measurement of enzyme activity after cross-linking

CuPhen reactions were stopped by adding an equal volume of 50 mM Tris–HCl, pH 7.5 containing 20 mM NEM and 20 mM EDTA. The samples were incubated at room temperature for 30 min and then centrifuged for 1 min in an Eppendorf tabletop centrifuge at 13,500g. The supernatant was removed, and the pellet was re-suspended in 100 μl of tissue homogenization buffer (250 mM sucrose, 2 mM EDTA, 30 mM MOPS-NaOH, pH 7.4). The entire 100-μl sample was used for the Ca-ATPase assay as described above.


Rationale and experimental approach

Multiple sequence alignments of the N- and C-terminal “linker regions” of several known NTPDase3 sequences with other membrane-bound NTPDases (rat NTPDase1 and human NTPDase 1, 2, 4/7, and 8) revealed complete conservation of P472, a high degree of conservation of P53, P476, and P481 (completely conserved in the NTPDase3 enzymes) and a lesser degree of conservation of P52, P484, and P485 (Fig. 1a).

Cysteine residues (V42C and G489C) were introduced in the “free sulfhydryl-less” C10S/C501S/C509S NTPDase3 background (C10S/C501S/C509S, which has very similar enzymatic properties to WT NTPDase3 [25]) for subsequent site-specific, oxidative, intermolecular cross-linking of the TM helices of human NTPDase3. These amino acids correspond to those of rat CD39/NTPDase1, which were shown to be excellent “sensors” of ATP effects on TM helix cross-linking [23]. Thus, V42 (corresponding to A34 in rat NTPDase1) and G489 (corresponding to S481 in rat NTPDase1) in human NTPDase3 were singly mutated to cysteine (see Fig. 1a). These mutants were used as background/control enzymes and TMD-specific cross-linking efficiency “sensors” for subsequent proline to alanine mutations used to test our hypothesis (see Fig. 1b).

Characterization of the TM “sensor” mutations (V42C and G489C NTPDase3)

The V42C and G489C mutants were constructed in the C10S/C501S/C509S NTPDase3 background to ensure the introduced cysteine is the only free sulfhydryl available for specific cross-linking at either the N- or C-terminal TM helix. The relative expression level and nucleotidase activities of the V42C and G489C mutants were similar to the C10S/C501S/C509S NTPDase3 control, with only slight but statistically significant decreases in nucleotidase activities for G489C (Table 1, NTPDase3 cysteine sensor V42 and G489 mutants in C10S/C501S/C509S background). CuPhen cross-linking of the TM helices via V42C or G489C caused a significant decrease in Ca-ATPase enzymatic activity (Fig. 2). However, the decrease in Ca-ATPase activity was attenuated in the presence of 5 mM ATP (Fig. 2). The cross-linking efficiency of V42C and G489C in the presence and absence of various nucleotides (ATP, ADP, adenosine monophosphate (AMP), and adenosine 5’-(β,γ-imido)triphosphate (AMP-PNP) all at 5 mM) and 5 mM inorganic phosphate (Pi) was then assessed, to determine the nucleotide specificity of the modulation of CuPhen TM cross-linking at these introduced cysteine “sensors,” as well as the utility of these background sensor mutations for the subsequent proline residue mutations. In the presence of 5 mM CuPhen, both V42C (Fig. 3a) and G489C (Fig. 3b) mutants are efficiently cross-linked, forming dimers (it should be noted that the antibody reacts much stronger with dimeric than monomeric NTPDase3, and thus only monomer bands were used in the calculation of the cross-linking efficiency [31]). When 5 mM ATP was present with CuPhen, the cross-linking efficiency of both V42C and G489C mutants decreased (the percent monomer remaining increased), mimicking the results obtained using cysteine substitutions at the same positions in rat CD39/NTPDase1 [23]. Also consistent with those previous NTPDase1 results, the cross-linking efficiency of V42C and G489C was attenuated by the addition of 5 mM ADP and AMP-PNP (Fig. 3a, b), but not by 5 mM AMP or Pi. Lastly, in the presence of Mg+2 or Ca+2, the decrease in CuPhen cross-linking efficiency caused by 5 mM ATP is attenuated in a divalent cation concentration-dependent manner for human NTPDase3 (data not shown), similar to what was observed earlier for NTPDase1 [23].

Fig. 2
Adenosine triphosphate (ATP) attenuates the copper phenanthroline (CuPhen)-induced decrease in nucleoside triphosphate diphosphohydrolase 3 (NTPDase3) Ca-ATPase activity. Nucleotidase activities were assayed as described in “Materials and methods ...
Fig. 3
Adenosine triphosphate (ATP), adenosine diphosphate (ADP), and adenosine 5’-(β,γ-imido)triphosphate (AMP-PNP) binding at the extracellular domain decreases the copper phenanthroline (CuPhen) cross-linking efficiency of V42C and ...

This data validates the use of V42C and G489C NTPDase3 mutants as “sensors” for active site-dependent changes of the TM helices in NTPDase3 and allows testing of the hypothesis that the linkage between the active site and the TM helices is dependent on the presence of one or more of the conserved proline residues located in the “linker regions” of NTPDase3, denoted in Fig. 1a.

Characterization of proline to alanine mutants in the V42C or G489C NTPDase3 background

To determine if the proline residues play a role in “coupling” ATP binding at the enzyme active site to movements of the TM helices, site-directed mutagenesis was used to singly substitute the conserved proline residues in the “linker regions” to alanine in both the V42C and G489C NTPDase3 backgrounds. In the V42C background, proline to alanine substitution at all the positions except 52 resulted in a significant decrease in relative expression level as compared to the V42C background control (Table 1, NTPDase3 proline residue to alanine mutants in V42C background). Proline to alanine substitutions at positions 53 and 481 resulted in substantial decreases in Ca-ATPase activity as compared to the V42C background control (Table 1, NTPDase3 proline residue to alanine mutants in V42C background). In addition, proline to alanine substitutions at positions 484 and 485 resulted in smaller but significant decreases in Ca-ATPase activity as compared to the V42C background control. In the G489C background, proline to alanine substitution at positions 53 and 481 resulted in significant decreases in relative expression levels as compared to the G489C background control (Table 1, NTPDase3 proline residue to alanine mutants in G489C background). Proline to alanine substitutions at positions 53, 481 and, to a lesser extent, 484, resulted in substantial decreases in Ca-ATPase activity, compared to the G489C background control (Table 1, NTPDase3 proline residue to alanine mutants in G489C background). Other nucleotidase activities were also measured for these mutants, and the changes in Ca-ADPase, Mg-ATPase, and Mg-ADPase activities were similar to the results for Ca-ATPase activities (see Table 1).

Each mutant was also oxidized with 5 mM CuPhen in the presence and absence of 5 mM ATP to assess CuPhen cross-linking efficiency (measured by quantifying the monomer band remaining relative to the untreated control). In the V42C background, the ATP-induced decrease in CuPhen cross-linking efficiency was abolished by proline to alanine substitution at positions 53 and 481 (Figs. 4a and and5a).5a). Proline to alanine substitution at position 481 abolished the ATP-induced decrease in CuPhen cross-linking efficiency in the G489C background (Figs. 4b and and5b).5b). Interestingly, alanine substitution at position 485 in the G489C background resulted in an inability of the mutant to be cross-linked at G489C by CuPhen, both in the presence or absence of ATP (Figs. 4b and and5b5b).

Fig. 4
Copper phenanthroline (CuPhen) cross-linking efficiency of proline to alanine mutants in the presence and absence of 5 mM adenosine triphosphate (ATP). Cross-linking was performed, and ATP was added before 5 mM CuPhen, as described in ...
Fig. 5
Specific proline to alanine substitutions abolish the adenosine triphosphate (ATP)-induced decrease in copper phenanthroline (CuPhen) cross-linking efficiency. Cross-linking was performed, and the efficiency of cross-linking was quantified by comparing ...

All the results are visually summarized in the cartoon representation of NTPDase3 monomers shown in Fig. 6. Mutations which abolish the ability of substrate (ATP) to diminish the cross-linking efficiency of the introduced TM helix cysteine residues (“uncoupling mutations”) are indicated by boxes around the mutations. Mutations exhibiting statistically significant decreases in nucleotidase activities are shown in bold type and include the P53A and P481A mutants in both sensor backgrounds. Taken together, this data suggests that a few specific proline residues in the linker region (i.e., P53 and P481) facilitate the linkage between binding of substrate at the enzyme active site to movements or reorientations of the TM helices that are necessary to achieve full enzymatic activity of NTPDase3.

Fig. 6
Summary of the proline to alanine nucleoside triphosphate diphosphohydrolase 3 (NTPDase3) mutant data. Shown are cartoon representations of two monomers of human NTPDase3. On the left, the results for the proline substitutions in the V42C N-terminal transmembrane ...


Proteins often undergo domain rearrangements that involve substrate or ligand binding in one area of a protein resulting in the movement of another part(s) of the protein [32]. These dynamic motions of domain rearrangement are necessary for optimal function of these proteins. Specifically, interplay between the extracellular domain and TMD of the cell surface NTPDases govern not only enzymatic activity but also substrate specificities [33]. Thus, NTPDase activities and substrate specificities are modulated not only by mutation of amino acids in the active site [10, 34], post-translational modifications such as glycosylation [35, 36] and N-acetylation [37] in the extracellular domain, but also by removal of the TM helices and perturbations of TM helical interactions [19, 3841] and modification of TMD cysteine by p-chloromercuriphenylsulfonate ([25, 42]).

The hypothesis tested in the present study is that one or more of the conserved proline residues found in the “linker regions” of human NTPDase3 mediate the phenomenon of substrate binding at the active site (located between the two lobes of the extracellular domain) causing movements and/or rearrangement of the TM helices to facilitate optimum nucleotide hydrolysis. When proline to alanine mutations were added to the “sensor” background/control NTPDase3 enzymes (V42C and G489C), alanine substitution at positions P53 and P481 in V42C caused a substantial decrease in activity, while a significant but smaller decrease in activity was noted for P484 and P485 (Table 1, NTPDase3 proline residue to alanine mutants in V42C background). Alanine substitution at positions P53 and P481 in G489C resulted in a substantial decrease in activity, while the P484A mutant exhibited a smaller decrease in activity (Table 1, NTPDase3 proline residue to alanine mutants in G489C background). The reduction in activity of the proline to alanine mutants (30–50%) is a little less than that seen upon deglycosylation (where a 50–60% decrease in Mg-ATPase activity was previously observed for NTPDase3 [35]). The decrease in enzyme activity of the proline to alanine mutants is slightly larger than that caused by the detergent Triton X-100 when the activity of NTPDase3 is assayed in the presence of Ca+2 (30–50% versus 20–40%), but this 30–50% decrease in enzyme activity of the proline to alanine mutants is smaller than the decrease observed in Mg-nucleotidase activities after Triton X-100 treatment, (70–90% [31]). Interestingly, alanine substitution at P472 increased nucleotidase activities, but these increases were smaller and not statistically significant in the V42C sensor background (Table 1, NTPDase3 proline residue to alanine mutants in V42C background and NTPDase3 proline residue to alanine mutants in G489C background). Taken together, this data suggests P53, P481, and P484, which are completely conserved in NTPDase3, but not in other NTPDases (see Fig. 1a), are needed for optimal enzymatic activity of NTPDase3.

Assessment of the proline to alanine mutations for possible “uncoupling” effects (abolishment of the ATP-dependent decrease in TM cross-linking efficiency) revealed that mutations at positions P53 and P481 in the V42C background were indeed capable of uncoupling ATP binding at the active site to reorientations or movements of the TM helices as assessed by CuPhen cross-linking (Figs. 4a and and5a).5a). In the G489C background, proline to alanine substitution at position P481 also caused this effect (Figs. 4b and and5b).5b). Only alanine substitutions at positions P53 and P481 resulted in both decreases in nucleotidase activities and abolishment of the decreased cross-linking efficiency of the TM helices in the presence of ATP. Surprisingly, alanine substitution at position P485 abolished CuPhen-induced cross-linking of G489C both in the presence and absence of ATP, although this was not true for cross-linking of V42C after alanine substitution at P485 (Figs. 4 and and5).5). This suggests that the placement of the cysteine residue used as the cross-linking TM sensor is important for the ability to measure the effects of the proline mutations, and that such proline to alanine mutations may affect the TM helix closer to the mutation more than the TM helix on the other end of the protein. Thus, alanine substitution at P485, which is only four amino acids away from the G489C mutation used as a sensor in the C-terminal helix, changes the orientation or freedom of movement of the C-terminal TM helix of NTPDase3 in a way that inhibits intermolecular, CuPhen-induced cross-linking at G489C, but not cross-linking at V42C, which is located in the N-terminal TM helix.

The current study is the first to demonstrate the importance of specific, conserved proline residues in the “linker regions” of NTPDase3 that “couple” the extracellular domain (site of nucleotide binding) to the dynamic motions and interactions of the TM helices, a linkage which is functionally important for optimal nucleotide hydrolysis. Analogous to results obtained using rat CD39/NTPDase1 [23], cysteine mutations in both TM helices near the extracellular surface (V42C and G489C in NTPDase3) proved to be useful TM cross-linking “sensors.” Both V42C and G489C NTPDase3 mutants were enzymatically active (Table 1) and efficiently cross-linked in the presence of CuPhen (Fig. 3). In the presence of ATP (as well as ADP and AMP-PNP), the CuPhen-induced cross-linking efficiency of the V42C and G489C NTPDase3 mutants was attenuated, resulting in an increase in the amount of monomer remaining (Fig. 3). Disulfide cross-linking of V42C and G489C via CuPhen also inhibited the nucleotidase activity of the mutants (Fig. 2), as was also observed with CD39/NTPDase1, suggesting the need for TM helix rearrangements and/or movements for full enzymatic activity. Not only does 5 mM ATP attenuate TM cross-linking (Fig. 3) but it also attenuates the CuPhen-induced decrease in enzyme activity (see Fig. 2). Thus, TM helix rearrangements and/or movements may be important for full activity of all cell surface NTPDases.

As defined by the polar TM interactions described in a recent study detailing the spatial organization and interactions of the TM helices of human NTPDase3 [31], efficient intermolecular cross-linking between monomers within an NTPDase3 dimer is not expected for either V42C or G489C, since that TM helix interaction model does not place either of these residues in close proximity to the same residue in the other monomer comprising the dimer [31]. This suggests that similar to rat NTPDase1 [23], NTPDase3 TM helices must have considerable rotational freedom near the extracellular surface of the membrane, and that this rotational freedom is substantially decreased in the presence of ATP, resulting in the nucleotide-induced decrease in CuPhen cross-linking efficiency.

In summary, consistent with our hypothesis, some proline to alanine mutants exhibit decreases in enzymatic activities, uncoupling of the interactions between the active site and TMD, or both. Importantly, other proline to alanine mutants also located in the linker region had little, if any, effect, demonstrating that the results obtained are not simply nonspecific artifacts, in the sense that any proline to alanine substitution in this area of the protein would result in decreases in activity and changes in ATP-induced cross-linking efficiencies as a result of adverse effects on protein folding. These conserved proline residues in the linker regions of NTPDase3 constitute proline-rich regions that mediate interactions between the extracellular and TM domains, suggesting the possible existence of polyproline II (PPII) helices. Typically, PPII helices are shorter than five amino acids, contain multiple proline residues, and often also contain positively charged amino acids [4345]. These characteristics are consistent with the amino acid sequence of the linker regions of NTPDase3, as well as with the importance of the P53 and P481 residues identified in this study (N-terminal: QEVLPP53GLK; C-terminal: PAESPLIRLP481IEPP). Interestingly, the X-ray structure of the soluble portion of rat NTPDase2 does contain parts of these linker regions, including residues corresponding to some of these proline residues (N-terminal: 53P; C-terminal: 472P and 476P, [13]). In the crystal structure, there is a short left-handed 310 helix in the C-terminal linker region, suggesting the possibility of polyproline II helix in this linker region of NTPDase3.

Taken together, the data presented in this study suggest that the interplay between the extracellular active site and the TMD, as determined by reduction of oxidative cross-linking in the presence of substrates and previously observed for NTPDase1, may be generalized to all cell surface NTPDases. Also, conserved prolines in the linker regions connecting the N- and C-terminal helices may couple these domain interactions for other NTPDases. However, it should be noted that the most conserved proline residue in these linker regions (P472 in NTPDase3), as well as a residue conserved in the cell surface NTPDases (P52 in NTPDase3), are not essential for NTPDase3 optimal activity and, when mutated, do not abolish the ability of ATP to modulate NTPDase3 TM cross-linking. For NTPDase3, proline residues at positions 53 and 481 help mediate this linkage, needed for optimal activity, which couples binding of nucleotide binding at the enzymatic active site to dynamic movements and/or rearrangement of the TM helices.


This work was supported by the National Institutes of Health R01 grant HL72382/HL72382-S1 to T.L.K.


Nucleoside triphosphate diphosphohydrolase 3
Transmembrane domain
Copper phenanthroline
3-(N-Morpholino)-propanesulfonic acid


1. Zimmermann H. Two novel families of ectonucleotidases: molecular structures, catalytic properties and a search for function. Trends Pharmacol Sci. 1999;20:231–236. doi: 10.1016/S0165-6147(99)01293-6. [PubMed] [Cross Ref]
2. Ivanenkov VV, Meller J, Kirley TL. Characterization of disulfide bonds in human nucleoside triphosphate diphosphohydrolase 3 (NTPDase3): implications for NTPDase structural modeling. Biochemistry. 2005;44:8998–9012. doi: 10.1021/bi047487z. [PubMed] [Cross Ref]
3. Murphy-Piedmonte DM, Crawford PA, Kirley TL. Bacterial expression, folding, purification and characterization of soluble NTPDase5 (CD39L4) ecto-nucleotidase. Biochim Biophys Acta. 2005;1747:251–259. [PubMed]
4. Ivanenkov VV, Murphy-Piedmonte DM, Kirley TL. Bacterial expression, characterization, and disulfide bond determination of soluble human NTPDase6 (CD39L2) nucleotidase: implications for structure and function. Biochemistry. 2003;42:11726–11735. doi: 10.1021/bi035137r. [PubMed] [Cross Ref]
5. Stout JG, Strobel RS, Kirley TL. Identification and immunolocalization of ecto-ATPDase in chicken stomach. Biochem Mol Biol Int. 1995;36:529–535. [PubMed]
6. Kansas GS, Wood GS, Tedder TF. Expression, distribution, and biochemistry of human CD39: role in activation-associated homotypic adhesion of lymphocytes. J Immunol. 1991;146:2235–2244. [PubMed]
7. Knowles AF. Differential expression of ectoMg2 + -ATPase and ectoCa2 + -ATPase activities in human hepatoma cells. Arch Biochem Biophys. 1988;263:264–271. doi: 10.1016/0003-9861(88)90635-2. [PubMed] [Cross Ref]
8. Karasaki S, Simard A, Lamirande G. Surface morphology and nucleoside phosphatase activity of rat liver epithelial cells during oncogenic transformation in vitro. Cancer Res. 1977;37:3516–3525. [PubMed]
9. Plesner L. Ecto-ATPases: identities and functions. Int Rev Cytol. 1995;158:141–214. doi: 10.1016/S0074-7696(08)62487-0. [PubMed] [Cross Ref]
10. Kirley TL, Crawford PA, Smith TM. The structure of the nucleoside triphosphate diphosphohydrolases (NTPDases) as revealed by mutagenic and computational modeling analysis. Purinergic Signal. 2006;2:379–389. doi: 10.1007/s11302-005-5301-6. [PMC free article] [PubMed] [Cross Ref]
11. Handa M, Guidotti G. Purification and cloning of a soluble ATP-diphosphohydrolase (apyrase) from potato tubers (Solanum tuberosum) Biochem Biophys Res Commun. 1996;218:916–923. doi: 10.1006/bbrc.1996.0162. [PubMed] [Cross Ref]
12. Bork P, Sander C, Valencia A. An ATPase domain common to prokaryotic cell cycle proteins, sugar kinases, actin, and hsp70 heat shock proteins. Proc Natl Acad Sci USA. 1992;89:7290–7294. doi: 10.1073/pnas.89.16.7290. [PubMed] [Cross Ref]
13. Zebisch M, Sträter N. Structural insight into signal conversion and inactivation by NTPDase2 in purinergic signaling. Proc Natl Acad Sci USA. 2008;105:6882–6887. doi: 10.1073/pnas.0802535105. [PubMed] [Cross Ref]
14. Hurley JH, Mason DA, Matthews BW. Flexible-geometry conformational energy maps for the amino acid residue preceding a proline. Biopolymers. 1992;32:1443–1446. doi: 10.1002/bip.360321104. [PubMed] [Cross Ref]
15. MacArthur MW, Thornton JM. Influence of proline residues on protein conformation. J Mol Biol. 1991;218:397–412. doi: 10.1016/0022-2836(91)90721-H. [PubMed] [Cross Ref]
16. Nicholson H, Tronrud DE, Becktel WJ, Matthews BW. Analysis of the effectiveness of proline substitutions and glycine replacements in increasing the stability of phage T4 lysozyme. Biopolymers. 1992;32:1431–1441. doi: 10.1002/bip.360321103. [PubMed] [Cross Ref]
17. Balasubramanian R, Lakshminarayanan AV, Sabesan MN, Tegoni G, Venkatesan K, Ramachandran GN. Studies on the conformation of amino acids. VI. Conformation of the proline ring as observed in crystal structures of amino acids and peptides. Int J Protein Res. 1971;3:25–33. doi: 10.1111/j.1399-3011.1971.tb01689.x. [PubMed] [Cross Ref]
18. Morris AL, MacArthur MW, Hutchinson EG, Thornton JM. Stereochemical quality of protein structure coordinates. Proteins. 1992;12:345–364. doi: 10.1002/prot.340120407. [PubMed] [Cross Ref]
19. Wang TF, Ou Y, Guidotti G. The transmembrane domains of ectoapyrase (CD39) affect its enzymatic activity and quaternary structure. J Biol Chem. 1998;273:24814–24821. doi: 10.1074/jbc.273.38.24814. [PubMed] [Cross Ref]
20. Wang CJ, Vlajkovic SM, Housley GD, Braun N, Zimmermann H, Robson SC, Sevigny J, Soeller C, Thorne PR. C-terminal splicing of NTPDase2 provides distinctive catalytic properties, cellular distribution and enzyme regulation. Biochem J. 2005;385:729–736. doi: 10.1042/BJ20040852. [PubMed] [Cross Ref]
21. Chiang WC, Knowles AF. Transmembrane domain interactions affect the stability of the extracellular domain of the human NTPDase 2. Arch Biochem Biophys. 2008;472:89–99. doi: 10.1016/ [PubMed] [Cross Ref]
22. Knowles AF, Li C. Molecular cloning and characterization of expressed human ecto-nucleoside triphosphate diphosphohydrolase 8 (E-NTPDase 8) and its soluble extracellular domain. Biochemistry. 2006;45:7323–7333. doi: 10.1021/bi052268e. [PubMed] [Cross Ref]
23. Grinthal A, Guidotti G. Dynamic motions of CD39 transmembrane domains regulate and are regulated by the enzymatic active site. Biochemistry. 2004;43:13849–13858. doi: 10.1021/bi048644x. [PubMed] [Cross Ref]
24. Smith TM, Kirley TL. Site-directed mutagenesis of a human brain ecto-apyrase: evidence that the E-type ATPases are related to the actin/heat shock 70/sugar kinase superfamily. Biochemistry. 1999;38:321–328. doi: 10.1021/bi9820457. [PubMed] [Cross Ref]
25. Murphy DM, Ivanenkov VV, Kirley TL. Identification of cysteine residues responsible for oxidative cross-linking and chemical inhibition of human nucleoside triphosphate diphosphohydrolase 3. J Biol Chem. 2002;277:6162–6169. doi: 10.1074/jbc.M110105200. [PubMed] [Cross Ref]
26. Smith TM, Kirley TL. Cloning, sequencing, and expression of a human brain ecto-apyrase related to both the ecto-ATPases and CD39 ecto-apyrases. Biochim Biophys Acta. 1998;1386:65–78. [PubMed]
27. Stoscheck CM. Protein assay sensitive at nanogram levels. Anal Biochem. 1987;160:301–305. doi: 10.1016/0003-2697(87)90051-0. [PubMed] [Cross Ref]
28. Fiske CH, Subbarow Y. The colorometric determination of phosphorous. J Biol Chem. 1925;66:375–400.
29. Smith TM, Kirley TL. Glycosylation is essential for functional expression of a human brain ecto-apyrase. Biochemistry. 1999;38:1509–1516. doi: 10.1021/bi9821768. [PubMed] [Cross Ref]
30. Careaga CL, Falke JJ. Thermal motions of surface alpha-helices in the D-galactose chemosensory receptor. Detection by disulfide trapping. J Mol Biol. 1992;226:1219–1235. doi: 10.1016/0022-2836(92)91063-U. [PMC free article] [PubMed] [Cross Ref]
31. Gaddie KJ, Kirley TL. Conserved polar residues stabilize transmembrane domains and promote oligomerization in human nucleoside triphosphate diphosphohydrolase 3. Biochemistry. 2009;48:9437–9447. doi: 10.1021/bi900909g. [PMC free article] [PubMed] [Cross Ref]
32. Gerstein M, Lesk AM, Chothia C. Structural mechanisms for domain movements in proteins. Biochemistry. 1994;33:6739–6749. doi: 10.1021/bi00188a001. [PubMed] [Cross Ref]
33. Grinthal A, Guidotti G. Transmembrane domains confer different substrate specificities and adenosine diphosphate hydrolysis mechanisms on CD39, CD39L1, and chimeras. Biochemistry. 2002;41:1947–1956. doi: 10.1021/bi015563h. [PubMed] [Cross Ref]
34. Drosopoulos JH, Broekman MJ, Islam N, Maliszewski CR, Gayle RB, 3rd, Marcus AJ. Site-directed mutagenesis of human endothelial cell ecto-ADPase/soluble CD39: requirement of glutamate 174 and serine 218 for enzyme activity and inhibition of platelet recruitment. Biochemistry. 2000;39:6936–6943. doi: 10.1021/bi992581e. [PubMed] [Cross Ref]
35. Murphy DM, Kirley TL. Asparagine 81, an invariant glycosylation site near apyrase conserved region 1, is essential for full enzymatic activity of ecto nucleoside triphosphate diphosphohydrolase 3. Arch Biochem Biophys. 2003;413:107–115. doi: 10.1016/S0003-9861(03)00084-5. [PubMed] [Cross Ref]
36. Mateo J, Kreda S, Henry CE, Harden TK, Boyer JL. Requirement of Cys399 for processing of the human ecto-ATPase (NTPDase2) and its implications for determination of the activities of splice variants of the enzyme. J Biol Chem. 2003;278:39960–39968. doi: 10.1074/jbc.M307854200. [PubMed] [Cross Ref]
37. Basu S, Kirley TL. Identification of a tyrosine residue responsible for N-acetylimidazole-induced increase of activity of ecto-nucleoside triphosphate diphosphohydrolase 3. Purinergic Signalling. 2005;1:271–280. doi: 10.1007/s11302-005-0650-8. [PMC free article] [PubMed] [Cross Ref]
38. Papanikolaou A, Papafotika A, Murphy C, Papamarcaki T, Tsolas O, Drab M, Kurzchalia TV, Kasper M, Christoforidis S. Cholesterol-dependent lipid assemblies regulate the activity of the ecto-nucleotidase CD39. J Biol Chem. 2005;280:26406–26414. doi: 10.1074/jbc.M413927200. [PubMed] [Cross Ref]
39. Mukasa T, Lee Y, Knowles AF. Either the carboxyl- or the amino-terminal region of the human ecto-ATPase (E-NTPDase 2) confers detergent and temperature sensitivity to the chicken ecto-ATP-diphosphohydrolase (E-NTPDase 8) Biochemistry. 2005;44:11160–11170. doi: 10.1021/bi050019k. [PubMed] [Cross Ref]
40. Chen W, Guidotti G. Soluble apyrases release ADP during ATP hydrolysis. Biochem Biophys Res Commun. 2001;282:90–95. doi: 10.1006/bbrc.2001.4555. [PubMed] [Cross Ref]
41. Zebisch M, Sträter N. Characterization of rat NTPDase1, -2, and -3 ectodomains refolded from bacterial inclusion bodies. Biochemistry. 2007;46:11945–11956. doi: 10.1021/bi701103y. [PubMed] [Cross Ref]
42. Chiang WC, Knowles AF. Inhibition of human NTPDase 2 by modification of an intramembrane cysteine by p-chloromercuriphenylsulfonate and oxidative cross-linking of the transmembrane domains. Biochemistry. 2008;47:8775–8785. doi: 10.1021/bi800633d. [PubMed] [Cross Ref]
43. Stapley BJ, Creamer TP. A survey of left-handed polyproline II helices. Protein Sci. 1999;8:587–595. [PubMed]
44. Creamer TP. Left-handed polyproline II helix formation is (very) locally driven. Proteins. 1998;33:218–226. doi: 10.1002/(SICI)1097-0134(19981101)33:2<218::AID-PROT6>3.0.CO;2-E. [PubMed] [Cross Ref]
45. Ma K, Kan L, Wang K. Polyproline II helix is a key structural motif of the elastic PEVK segment of titin. Biochemistry. 2001;40:3427–3438. doi: 10.1021/bi0022792. [PubMed] [Cross Ref]

Articles from Purinergic Signalling are provided here courtesy of Springer