Expression and Biochemical Characterization of the Human Enzyme N-Terminal Asparagine Amidohydrolase (hNTAN1)

Abstract

The enzymatic deamidation of N-terminal L-Asn by N-terminal asparagine amidohydrolase (NTAN1) is a feature of the ubiquitin-dependent N-end rule pathway of protein degradation, which relates the in vivo half-life of a protein to the identity of its N-terminal residue. Herein we report the bacterial expression, purification, and biochemical characterization of the human NTAN1 (hNTAN1). We show here that hNTAN1 is highly selective for the hydrolysis of N-terminal peptidyl L-Asn, but fails to deamidate free L-Asn or L-Gln, N-terminal peptidyl L-Gln, or acetylated N-terminal peptidyl L-Asn. Similar to other N-terminal deamidases, hNTAN1 is shown to possess a critical Cys residue that is absolutely required for catalysis, corroborated in part by abolishment of activity through the point mutation Cys75Ala. We also present evidence that the exposure of a conserved L-Pro at the N-terminus of hNTAN1 following removal of the initiating L-Met is important for function of the enzyme. The results presented here should assist in the elucidation of molecular mechanisms underlying the neurological defects of NTAN1-deficient mice observed in other studies, and in the discovery of potential physiological substrates targeted by the enzyme in the modulation of protein turnover via the N-end rule pathway.

Targeted proteolytic degradation is essential for the regulation of a variety of cellular processes by achieving the correct balance between protein folding and the degradation of misfolded or damaged proteins (1). In eukaryotes, the ubiquitin (Ub)-proteasome system (UPS) is the primary pathway for the time-specific elimination of targeted proteins and the N-end rule pathway is a subset of this system that relates the in vivo half-life of a protein to the identity of its N-terminal residue (2, 3). Degradation signals (degrons) that can be targeted by the N-end rule pathway are of two discrete subsets: N-terminal degrons (N-degrons) and internal degrons (2, 4). The pathway has been observed in all organisms examined, from prokaryotes to eukaryotes, though the former exists in a Ub-independent manner (5). In eukaryotes, an N-degron consists of three determinants: a destabilizing N-terminal residue (Figure 1), one (or more) internal Lys residues (the site of poly-Ub chain formation), and a conformationally flexible region in the vicinity of this internal Lys (6, 7). An N-degron can be generated from a precursor, called a pre-N-degron, through specific protease-mediated post-translational modifications (8).

Figure 1. The mammalian N-end rule pathway.

N-terminal residues are indicated by single-letter abbreviations. Green boxes denote the remainder of the protein substrate. C* denotes oxidized N-terminal Cys (either Cys-sulfinate or Cys-sulfonate). The Cys oxidation requires nitric oxide (NO) and oxygen or its derivatives. “Tertiary”, “secondary”, and “primary” denote mechanistically distinct subsets of destabilizing N-terminal residues (See “Introduction”). The enlarged and blue-colored “NTAN1” denotes the N-terminal asparagine amidohydrolase characterized in this study.

The N-end rule has a hierarchical structure (Figure 1). In eukaryotes, N-terminal L-Asn and L-Gln function as tertiary destabilizing residues through their deamidation by N-terminal amidohydrolases (Nt-amidases) into the secondary destabilizing N-terminal residues L-Asp and L-Glu, respectively (9–11). The secondary destabilizing activity of N-terminal L-Asp and L-Glu requires their conjugation to L-Arg by the ATE1-encoded Arg-tRNA protein transferase (R-transferase) (12–14). In mammals and other eukaryotes that produce nitric oxide (NO), N-terminal L-Cys can also function as a tertiary destabilizing residue through its oxidation to Cys-sulfinate or Cys-sulfonate in an O₂- or NO-dependent manner, whereby the oxidized L-Cys can be arginylated by ATE1 (8, 13). Together with other primary destabilizing residues, the N-terminal L-Arg allows for protein substrate recognition by specific E3 Ub ligases and subsequent degradation (15).

A number of genetic studies with knock-out mice have implicated involvement of the N-end rule pathway in cardiac development and signaling, angiogenesis (12, 16), meiosis (17), DNA repair (18), neurogenesis (16), pancreatic functions (19), learning and memory (10, 20), female development (17), muscle atrophy (21), and olfaction (22). Other functions of the pathway include: (i) selective degradation of misfolded proteins (23), (ii) sensing of nitric oxide, oxygen, and heme (12–14, 24), (iii) regulation of short peptide import (25, 26), (iv) fidelity of chromosome segregation (27), (v) regulation of apoptosis through NTAN-1 mediated degradation of the caspase-mediated C-terminal fragment of Drosophila inhibitor of apoptosis 1 (DIAP1) (28), and (vi) regulation of leaf senescence in plants (29).

In contrast to S. cerevisiae which possesses a single enzyme (Nta1) that can deamidate either L-Asn or L-Gln at the N-terminus of a protein substrate (9), vertebrates possess one distinct enzyme for the specific deamidation of N-terminal L-Asn (NTAN1) and one for the specific deamidation of N-terminal L-Gln (NTAQ1) (11). NTAN1-deficient mice were found to display neurological defects such as impairment of spontaneous activity, spatial memory, and in socially conditioned exploratory behavior (10).

In an earlier study, an enzyme with N-terminal L-Asn hydrolase activity from porcine liver was purified and partially characterized (30). A gene encoding a putative NTAN1 enzyme, Ntan1, was subsequently isolated from a mouse cDNA library and shown to complement a S. cerevisiae Δnta1 mutant strain (31). Herein, we describe the expression and biochemical characterization of the human NTAN1 (hNTAN1). Among our results, we have identified an essential L-Cys residue that is critical for the hydrolysis of N-terminal L-Asn by hNTAN1, and thus may act as a nucleophile in catalysis, in analogy with both the S. cerevisiae Nta1 (9) and the dedicated murine N-terminal L-Gln deamidase NTAQ1 (11).

EXPERIMENTAL PROCEDURES

Materials

Synthetic oligonucleotides were purchased from Integrated DNA Technologies (Coralvillle, IA). Restriction enodnucleases, Vent DNA polymerase, T4 DNA ligase, and dNTPs were from New England Biolabs (Ipswich, MA). Nickel-nitrilotriacetic acid (Ni²⁺-NTA) resin was from QIAGEN (Valencia, CA). Strep-Tactin Superflow high capacity suspension and D-Desthiobiotin were obtained from IBA GmbH (St. Louis, MO). Anti-FLAG M2 magnetic beads, FLAG peptide, Val⁵-angiotensin II (AII), and Asn¹Val⁵-angiotensin II (N¹-AII) were purchased from Sigma-Aldrich (St. Louis, MO). Gln¹Val⁵-angiotensin II (Q¹-AII), Ac-Asn¹Val⁵-angiontensinII (Ac-N¹-AII), and a 9mer peptide (N¹-DIAP) corresponding to the N-terminus of the caspase-mediated C-terminal fragment of DIAP1 (28) were purchased from Abgent Inc. (San Diego, CA). Difco 2xYT growth medium was from Becton Dickinson (Franklin Lakes, NJ). All other reagents were obtained from Sigma-Aldrich (St. Louis, MO) unless otherwise noted.

Construction of hNTAN Expression Plasmids

A gene encoding the human protein N-terminal asparagine amidohydrolase (hNTAN1) (933bp) with a fused N-terminal His₆ affinity tag was synthetically assembled using a set of codon-optimized overlapping oligonucleotides designed by the program DNAWorks as described previously (32). NcoI and EcoRI restriction sites were incorporated into the outermost sense and antisense oligonucleotides respectively. Sense primers (Table S1 of the Supporting Information) were designed to: (i) introduce the hNTAN1 initiation codon within an NcoI site thereby eliminating the N-terminal His₆ affinity tag (NTAN-N-NoHis) and (ii) modify the ribosome binding site (RBS) and incorporate an upstream XbaI site in order to improve the translation initiation rate of hNTAN1 in E. coli (NTAN-RBS2). Antisense primers were designed to append a C-terminal His₆ affinity tag (NTAN-C-H), a C-terminal strepII affinity tag (NTAN-C-SII), and a C-terminal FLAG-strepII tandem affinity tag (NTAN-C-FSII), all followed by a NotI site. Lastly, a sense primer (NTAN-MP-RBS2) was designed to eliminate the GGA codon immediately following the initiation ATG (within an NcoI site) that was incorporated to maintain a correct open reading frame in the cassettes described above. Individual PCR products for each cassette were digested accordingly and ligated into pET28-a (Novagen). The RBS and the N- and C-terminal amino acid sequences of each hNTAN1 construct created are listed in Table 1. E. coli BL21 (DE3) was used for protein expression and purification. Genes encoding hNTAN1 mutants C75A, C75S, C75T, or C160A, were constructed by overlap-extension PCR using the pNTAN2FLAGSII plasmid as template and the primer pairs shown in Table S1 of the Supporting Information.

Table 1. hNTAN1 constructs and their relevant expression characteristics in BL21 (DE3).

The sequences of the sense and antisense primers used for the modified termini and RBS are found in Supplementary Table 1. The resultant amino acid sequences are shown where residues following the ellipsis are indicative of the C-terminus of the sequence. The purity of hNTAN1 was evaluated by SDS-PAGE, where band intensities were estimated by densitometry imaging (Quantity 1, BioRad).

Plasmid	RBS	Amino Acid Sequence	hNTAN1 Purity
pHisNTAN*	AAGGAGA	MGGSSHHHHHHSSGPLLV…SGPS	30 %
pNTANHis	AAGGAGA	MGPLLV…SGPSGGSSHHHHHH	30 %
pR1NTANSII	AAGGAGA	MGPLLV…SPGSSAWSHPQFEK	30 %
pR2NTANSII	AACAATAATAAGGAGATAAGAA	MGPLLV…SPGSSAWSHPQFEK	70 %
pNTANFLAGSII	AACAATAATAAGGAGATAAGAA	MGPLLV…SPGSGGGSDYKDDDDKSAWSHPQFEK	> 98%
pNTAN2FLAGSII	AACAATAATAAGGAGATAAGAA	MPLLV…SPGSGGGSDYKDDDDKSAWSHPQFEK	> 98%

^*Trace activity detected for His₆-hNTAN1 through addition of final purified product (10μg total protein) to 1mM N¹-AII for 90 min at 37°C in 50mM Tris-HCl, 150mM NaCl, pH 7.5. Reactions were quenched with 12% TCA (w/v) and then evaluated by HPLC for the presence of a peak corresponding to the expected product AII. For all other plasmids, the presence of hNTAN1 activity could be detected through the addition of final purified product (10–100ng total protein) to 100μM N¹-AII for 1 min at 37°C in the same buffer and evaluated similarly.

Expression and purification of recombinant hNTAN1

E. coli BL21(DE3) cells transformed with pNTAN2FLAGSII were cultured overnight at 37°C in 2xYT medium supplemented with 30μg/mL kanamycin and used to inoculate 500mL fresh medium (1:100 dilution). When the absorbance at 600nm (A₆₀₀) reached 0.5–0.7, the cells were transferred to 25°C and allowed to equilibrate for 20 min, at which point the culture was supplemented with IPTG to a final concentration of 1mM to induce protein expression. After 16 hr incubation at 25°C, the cells were harvested by centrifugation at 10,000× g for 10 min. hNTAN1 was subsequently purified by tandem strepII affinity chromatography and immobilized anti-FLAG as follows: cell pellets were resuspended in 15mL buffer W (100mM Tris-HCl, 150mM NaCl, 1mM EDTA, pH 8) on ice, and then lysed by three passes through a French pressure cell. Debris was removed by centrifugation at 40,000× g for 45 min and the resulting supernatant (soluble fraction) was decanted and applied to a 5mL polypropylene column (Pierce) packed with 300μL of strep-tactic superflow high capacity resin (33). The column flow-through was collected and re-applied to the column for 3 total passes. The resin was then washed with 30 bed volumes buffer W and 10 bed volumes buffer W2 (buffer W containing 100μM D-Desthiobiotin) before the resin was incubated with 10 bed volumes buffer E (buffer W containing 2.5mM D-Desthiobiotin) and the eluant was collected. All purification steps were performed at 4°C. The elution fraction was applied to an Amicon Ultra 10K MWCO filter, concentrated to a final volume of 1mL in buffer W, and then incubated with 200μL anti-FLAG M2 magnetic beads for 1 hr at room temperature with gentle rotation. The magnetic beads were then collected using a magnetic separator (Dynal MPC-S, Invitrogen) to remove the supernatant, and washed with 30 bed volumes buffer W before the purified protein was collected by competitive elution with 5 bed volumes of buffer W containing 100μg/mL FLAG peptide by gentle rotation for 10 min at room temperature. The final eluant was applied to an Amicon Ultra 10K MWCO filter, concentrated to a final volume of 100μL in storage buffer (50mM Tris-HCl, 150mM NaCl, 1mM EDTA, 0.5mM dTT pH 7.5) and stored at 4°C. Procedures for the purification of recombinant hNTAN1 fused to either a His₆ or a strepII affinity tag can be found in the Supporting Information. Protein concentrations were determined using a calculated extinction coefficient of 32,430 M⁻¹cm⁻¹ (34).

Activity assays

For kinetic analyses, hNTAN1 (40–50nM purified enzyme) was incubated with substrate (concentrations from 0–5 × K_M) at 37°C in assay buffer (50mM Tris-HCl, 150mM NaCl, pH 7.5) at a total volume of 100μL, and at desired times reactions were quenched with 5μL of 12% (w/v) trichloroacetic acid (TCA). Product formation was determined by HPLC as follows: For the substrate Asn¹Val⁵-angiotensin II (N¹-AII), an aliquot of the quenched reaction was brought to a final volume of 100μL with assay buffer and then injected into a Phenomenex C18 reverse-phase column followed by elution with: 5% ACN, 95% H₂O for 2 min, increasing to 21% ACN, 79% H₂O over 3 min, then increasing to 24.7% ACN, 75.3% H₂O over 11 min, and finally increasing to 95% ACN, 5% H₂O over 3 min before returning to 5% ACN, 95% H₂O over 3 min. Both solvents contained 0.1% trifluoroacetic acid (TFA). The product concentration was determined using the integration area at 280nm and a standard curve generated under the same conditions with Val⁵-angiotensin II (AII). For the substrates L-Asn and L-Gln, following quenching, reaction products were derivatized with o-phthalaldehyde (OPA) and analyzed by HPLC as described previously (32). A description of the HPLC program employed to evaluate activity for the substrate N¹-DIAP can be found in the Supporting Information. All reactions were done at least in triplicate and the observed rates were fit to the Michaelis-Menten equation using Kaleidagraph (Synergy).

The effect of pH on hNTAN1 activity was determined in 50mM Bis-Tris, pH 6.0–6.5, 50mM Tris-HCl, pH 7.0–8.5, or 50mM CAPSO, pH 10, and 150mM NaCl. Enzyme was preincubated in buffer and reactions were initiated by the addition of preincubated enzyme (1nM) to 25μM N¹-AII at 37°C in the appropriate pH buffer. After 1.5 min, reactions were quenched and analyzed by HPLC as described earlier.

The effects of various divalent salts, EDTA and D-desthiobiotin on hNTAN1 activity were determined by preincubating enzyme (20nM) in assay buffer at 4°C for 3 hr with either EDTA (1mM), D-desthiobiotin (3mM), or a divalent salt (1mM). Reactions were initiated by the addition of the preincubated enzyme (1nM) to 100μM N¹-AII in assay buffer at 37°C. After 4 min, the reactions were quenched and analyzed by HPLC as described earlier.

Circular Dichroism (CD)

Far-UV CD spectra were measured at 200–260nm using a J-815 CD spectrometer (Jasco, Easton, MD) equipped with a temperature-controlled cell holder set to room temperature, using 15μM enzyme in 50mM Tris-HCl, 150mM NaCl, 1mM EDTA, 0.5mM dTT, pH 7.5. CD thermal melting curves were generated by obtaining the CD spectra at 210nm from 20–80°C at a rate of 1°C/min in 50mM Tris-HCl, 150mM NaCl, 1mM EDTA, 0.5mM dTT, pH 7.5. Data points were collected at 5°C intervals and T_M values were subsequently determined as the inflection point of sigmoidal curves fit to each data set using Kaliedagraph.

RESULTS

Construction of a synthetic hNTAN1 gene

The human NTAN1 gene (GenBank accession no: NM_173474) contains a number of rare E. coli codons, including 13 such Arg codons (AGA, AGG, and CGA). The presence of rare codons has been shown to be detrimental to the expression of several recombinant proteins (35). A synthetic hNTAN1 gene fused to a 5′ sequence encoding a His₆ affinity tag was constructed and expressed in E. coli BL21 (DE3). Western blot analysis with an anti-His tag antibody as described previously (36) revealed the presence of a band with the expected M.W. (~35kDa) in both the soluble and insoluble fractions (Figure S1 of the Supporting Information). However, the purified His₆-hNTAN1 enzyme was found to be minimally active towards the substrate N¹-AII (Table 1). In contrast, fusion of a His₆ affinity tag at the C-terminus of hNTAN1 led to the expression of enzyme (Figure S2 of the Supporting Information) with much higher specific activity towards N¹-A^II, indicating that an unoccluded N-terminus may be important for catalysis. In subsequent studies, it was found that hNTAN1 activity was inhibited by Ni²⁺ or Co²⁺ and therefore, for purification purposes, the C-terminal His₆ tag was replaced with a strepII peptide affinity tag (pR1NTANSII) and accordingly, EDTA was included in all purification buffers.

To improve the expression yield, the RBS calculator (37) was used to design an optimized ribosomal binding site for more efficient translation initiation. Upon expression of hNTAN1-strepII using the engineered RBS sequence from plasmid pR2NTANSII, the expression yield of hNTAN1 increased 2–3-fold, while protein purity following strepII affinity chromatography concomitantly increased from 30% to approximately 70% (Table 1) as determined by SDS-PAGE (Figure S3 of the Supporting Information). To further improve the purity of hNTAN1 following affinity chromatography, we constructed a plasmid encoding a tandem C-terminal FLAG-strepII affinity tag (pNTANFLAGSII). With this construct, the eluant obtained from the strepII affinity chromatography step was loaded onto magnetic beads with an immobilized anti-FLAG monoclonal antibody, and eluted with buffer containing a large excess of synthetic FLAG. Using this tandem affinity chromatography procedure, we obtained ~0.2mg/L recombinant hNTAN1 that was determined to be > 98% homogeneous by SDS-PAGE (Figure 2).

Figure 2. SDS-PAGE analysis of hNTAN1 following tandem affinity purification.

1, hNTAN1 was expressed with C-terminal tandem FLAG-strepII affinity tag from pNTANFLAGSII and purified as described in detail in “Experimental Procedures”. Equivalent product homogeneity was observed following analogous purification of hNTAN1 expressed from pNTAN2FLAGSII (Figure S4). 2, M.W. marker.

ESI-MS analysis of the purified hNTAN1 revealed the presence of two major peaks of 37,208 ± 12 and 37,020 ± 12 m/z at a ~3:1 intensity ratio (Figure 3A). The M.W. of the latter peak was in agreement with the predicted M.W. of the enzyme (37,026 Da). The 37,208 ± 12 peak was consistent with an hNTAN1 species containing the initiating L-Met and an L-Gly at position 2 that was introduced during cloning (37,214 Da). Although the E. coli methionine aminopeptidase (MAP) is highly efficient in removing the initiating L-Met residue when L-Gly occupies P1′ of a protein substrate, its catalytic efficiency is severely compromised when the P2′ residue is L-Pro (38), as was the case in the hNTAN1 enzyme expressed from plasmid pNTANFLAGSII. To prevent the accumulation of unprocessed L-Met-L-Gly-hNTAN1, we generated a modified construct (pNTAN2FLAGSII) in which the L-Gly at +2 was eliminated. MS analysis of the purified protein obtained from cells transformed with the pNTAN2FLAGSII revealed a single peak of the expected M.W. for hNTAN1 in which the initiating L-Met had been removed by MAP (m/z of 37,022 ± 12) (Figure 3B). It is worth noting that E. coli MAP is highly efficient in removing the initiating L-Met residue when L-Pro occupies P1′ and L-Leu occupies P2′ (38) given that native hNTAN1 contains L-Pro at +2 and L-Leu at +3.

Figure 3. Deconvoluted ESI-MS spectra of hNTAN1.

(A) hNTAN1 expressed from pNTANFLAGSII and isolated using tandem affinity purification (See “Experimental Procedures”). (B) hNTAN1 expressed from pNTAN2FLAGSII and isolated with identical tandem affinity purification protocol. Samples were analyzed in 50mM Tris-HCl, 150mM NaCl, 1mM EDTA, 0.5mM dTT, pH 7.5. Calculated masses of observed species: hNTAN1 – 37,026 Da; L-Met-L-Gly-hNTAN1 – 37, 214 Da.

Biochemical characterization of hNTAN1

The kinetics for hydrolysis of N-terminal L-Asn angiontensin (N¹-A^II) by hNTAN1 (Table 2) purified from either pNTANFLAGSII or pNTAN2FLAGSII (Table 1) were determined by HPLC (Figure S5). In both cases, the data could be fit to a Michaelis-Menten model (Figure S6). hNTAN1 exhibited nearly identical K_M values (K_M = 31μM; pNTANFLAGSII and K_M = 35μM; pNTAN2FLAGSII) independent of the expression construct; however, hNTAN1 expressed from pNTAN2FLAGSII in which the L-Gly at +2 was eliminated, displayed a nearly 3-fold greater k_cat (k_cat = 2.3s⁻¹; pNTANFLAGSII and k_cat = 6.6s⁻¹; pNTAN2GLAGSII). This result indicates that the presence of the conserved L-Pro residue (Figure S7 of the Supporting Information) at the N-terminus of hNTAN1 is important for catalysis and is corroborated by the ESI-MS analysis described above.

Table 2. Summary of kinetic parameters for hydrolysis of N-terminal L-Asn angiotensin II (N¹-AII) by hNTAN1.

Kinetic parameters were derived from nonlinear fit of the Michaelis-Menten equation to initial rate measurements. Reactions were carried out at 37°C in 50mM Tris-HCl, 150mM NaCl, pH 7.5. The parenthetical in the enzyme column depicts whether purified hNTAN1 was expressed from pNTANFLAGSII (+ Gly2) or from pNTAN2FLAGSII (− Gly2).

Enzyme	KM (μM)	kcat (s−1)	kcat/KM (M−1s−1)
hNTAN1 (+Gly2)	31 ± 4	2.3 ± 0.07	7.4 ± 1 × 104
hNTAN1 (−Gly2)	35 ± 3	6.6 ± 0.1	1.9 ± 0.2 × 105

hNTAN1 was unable to hydrolyze Q¹-AII, Ac-N¹-AII, L-Asn, and L-Gln, indicating that the recombinant human enzyme displayed a substrate specificity analogous to that of the isoform previously purified from porcine liver (30). Moreover, hNTAN1 was unable to hydrolyze N¹-DIAP (data not shown), suggesting the possibility that NTAN1-mediated deamidation of the DIAP1 substrate in vivo (28) may be additionally dependent on interactions between NTAN1 and other (unidentified) protein factors.

The effect of pH on hNTAN1 activity with N¹-AII is shown in Figure 4, which revealed that the enzyme exhibited an activity optimum between pH 6.0–7.5, followed by a sharp decline in activity at alkaline pH. Our approximation for pH dependence of hNTAN1 activity suggests that the protonation of a single group within the enzyme is crucial for catalysis, resulting in a pH-dependence curve with a single descending limb pKa of 8.2 ± 0.06.

Figure 4. Effect of pH on hNTAN1 activity.

hNTAN1 (20nM) was preincubated at the appropriate pH for 1.5min at 25°C. Reactions were initiated by the addition of the preincubated enzyme (1nM) to 25μM N¹-AII at 37°C in the identical preincubation buffer. After 1.5 min, reactions were quenched and then analyzed by HPLC as described in “Experimental Procedures”. Activity was evaluated as the initial velocity of the reaction normalized by hNTAN1 concentration. All reactions were done in triplicate. Buffers used were: 50mM Bis-Tris, pH 6.0–6.5, 50mM Tris, pH 7.0–8.5, and 50mM CAPSO, pH 10. All buffers contained 150mM NaCl.

As summarized in Table 3, low micromolar concentrations of Zn²⁺ and Cu²⁺ completely inhibit hNTAN1 activity, while 50μM Ni²⁺, Co²⁺, or Mn²⁺ only partially inhibited the hydrolysis of N¹-AII. Further, Mg²⁺, Ca²⁺, EDTA, and D-desthiobiotin did not impart any noticeable effects on the activity of hNTAN1.

Table 3. Effects of divalent metals, EDTA, and D-desthiobiotin on hNTAN1 activity.

hNTAN1 (20nM) was preincubated in 50mM Tris-HCl, 150mM NaCl, pH 7.5 at 4°C for 3 hr with either EDTA (1mM), D-desthiobiotin (3mM), or a divalent salt (1mM). Reactions were initiated by addition of the preincubated enzyme (1nM) to 100μM N¹-AII in 50mM Tris-HCl, 150mM NaCl, pH 7.5 at 37°C. After 4 min, reactions were quenched and then analyzed by HPLC as described in “Experimental Procedures”.

Effector	% Activity Remaining
MgCl2	100
CaCl2	100
CoCl2	58 ± 2.8
MnSO4	40 ±14
NiSO4	39 ± 4.2
CuCl2	0 *
ZnSO4	0
EDTA	100
D-desthiobiotin	100

^*Area of the product peak was less than or equal to that observed in an HPLC trace of N¹-AII only at the equivalent reaction conditions.

Phylogenetic conservation and mutagenesis of conserved Cys residues in hNTAN1

Given prior validation that the functionally similar enzymes S. cerevisiase Nta1 (9) and murine NTAQ1 (11) each possess an essential catalytic cysteine residue, we explored the possibility that hNTAN1 contained an analogous activity-dependent cysteine residue. A Position Specific Iterative BLAST (PSI-BLAST) (39) search was carried out with the complete amino acid sequence of hNTAN1 to determine the phylogenetic conservation across each of the enzyme’s five cysteines, as those residues exhibiting the highest degree of conservation would be implicated as having a higher likelihood for functional and/or structural essentiality. Sequence alignment of hNTAN1 with the PSI-BLAST results (n = 106) revealed two distinct cysteine conservation groupings: (i) C75 and C160 (70% and 66% Cys frequencies respectively), and (ii) C89, C93, and C118 (28%, 22%, and 13% Cys frequencies respectively) (Figure 5). The C75 and C160 in hNTAN1 were substituted with L-Ala and the relative hydrolysis of N¹-AII by the respective protein mutants was determined (Table 4). While replacement of C160 with L-Ala had no effect on the catalytic activity, the C75A variant enzyme was completely inactive. This finding indicates that C75 is essential for enzymatic activity and likely to constitute the catalytic nucleophile. The substitution of C75 with other nucleophilic amino acids, namely L-Ser or L-Thr, failed to restore activity.

Figure 5. Phylogenetic conservation of hNTAN1 Cys residues.

A PSI-BLAST search was carried out with the complete amino acid sequence of hNTAN1 (UniProtKB number: Q96AB6) at EMBL-EBI (http://www.ebi.ac.uk/Tools/sss/psiblast/). The “UniProt Knowledgebase” was selected as the Protein Database and the PSI-BLAST threshold was set to 10. An alignment of hNTAN1 with the PSI-BLAST results (n = 106) was generated with ClustalW2 (40) and used to evaluate the degree of conservation across each Cys residue in hNTAN1. Each plot depicts the residue identities and corresponding frequencies observed at the hNTAN1 Cys position denoted at the top of each plot. The italicized percentage value within each plot indicates the frequency of Cys (highlighted by red outline) across all sequences at the given hNTAN1 position.

Table 4. Relative activity of hNTAN1 mutants.

Reactions of hNTAN1 (40–50nM) with 100μM N¹-AII were carried out at 37°C in 50mM Tris-HCl, 150mM NaCl, pH 7.5. After 1 min, reactions were quenched and then analyzed by HPLC as described in “Experimental Procedures.”

Enzyme	% Relative Activity (v. WT)
WT	100
C160A	100
C75A	0 *
C75S	0
C75T	0

^*Area of the product peak was less than or equal to that observed in an HPLC trace of N¹-AII only at the equivalent reaction conditions.

Analysis of the CD spectrum of recombinant hNTAN1 indicated 30% helical, 20% sheet, 20% turn, and 30% random structure (Table 5). The mutation C75A resulted in no obvious change in CD spectra or secondary structure composition (Figure 6), further supporting the suggestion that the inactivity of this variant was strictly due to the removal of the potentially nucleophilic thiol group and not to changes in folding.

Table 5.

Calculated secondary structure of WT and C75A hNTAN1 from CD spectra.

hNTAN1 variant	% Helix	% Sheet	% Turn	% Random
WT	30	20	20	30
C75A	30	20	20	30

Figure 6. Far-UV CD analyses of WT and C75A hNTAN1.

CD spectra of each sample (15μM) were recorded at 25°C in 50mM Tris-HCl, 150mM NaCl, 1mM EDTA, 0.5mM dTT, pH 7.5. The spectra shown reflect 3 accumulations per sample. WT hNTAN1 (●). hNTAN1-C75A (○). mdeg, millidegrees.

DISCUSSION

A number of biochemical and genetic studies over the past twenty years have begun to elucidate the significance of the N-end rule pathway in numerous physiological processes and have revealed many of the critical components involved in this subset of the UPS-dependent proteolytic quality control system. Proteolysis by the N-end rule pathway is dependent upon distinct types of degradation signals (degrons). In vertebrates, N-terminal L-Asn and L-Gln can function as tertiary destabilizing residues of the N-end rule pathway through enzymatic deamidation by either NTAN1 (Asn-specific) or NTAQ1 (Gln-specific) to yield the secondary destabilizing N-terminal L-Asp and L-Glu, respectively.

Almost 20 years ago, a mammalian monomeric enzyme purified from the cytosolic fraction of porcine liver cells (30) was shown to exhibit L-Asn-specific Nt amidase activity. More recently, a murine cDNA that complemented a S. cerevisiase Δnta1 mutant was identified, though the protein was not characterized enzymologically (31). Interestingly, the murine gene that was able to complement the S. cerevisiase Δnta1 mutant displays only 10% amino acid sequence identity to the recently described murine L--Gln-specific Nt amidase, NTAQ1 (11), and only 12% amino acid sequence identity to the S. cerevisiase Nta1 (9), which deamidates both L-Asn and L-Gln at the N-terminus of a protein substrate.

Because of its high degree of sequence homology to other mammalian NTAN1s (e.g. murine and porcine) a putative NTAN1 gene from Homo sapiens (hNTAN1) has been assigned, though direct experimental evidence of its activity and biochemical characterization had been lacking. Here, we report that bacterial expression and kinetic characterization of hNTAN1. Expression of hNTAN1 in E. coli at preparative levels was achieved through the use of a codon optimized gene and an optimal RBS and translation initiation region for efficient protein synthesis. Two lines of evidence suggested that the N-terminus of hNTAN1 is important for its enzymatic activity. First, enzyme expressed with an N-terminal epitope tag exhibited substantially lower activity relative to enzyme expressed with a C-terminal epitope tag, as evaluated by deamidation of N¹-AII (Table 1). Second, the presence of an N-terminal L-Met-L-Gly dipeptide, which partially interfered with processing by the E. coli methionine aminopeptidase (MAP), resulted in an enzyme preparation containing an approximately 3:1 ratio of L-Met-L-Gly-hNTAN1 to MAP-processed hNTAN1 with an exposed L-Pro at the N-terminus. This mixed preparation had nearly 3-fold lower specific activity relative to a preparation of the authentic human enzyme. Thus, we concluded that the exposure of the conserved L-Pro residue at the N-terminus is important for the function of the enzyme.

Kinetic analysis revealed that hNTAN1 is highly selective for the hydrolysis of N-terminal peptidyl L-Asn, but fails to deamidate free L-Asn or L-Gln, N-terminal peptidyl L-Gln (in Q^I-AII), or acetylated N-terminal peptidyl L-Asn (in Ac-N¹-AII). The human enzyme displayed approximately 10-fold lower K_M and 100-fold lower k_cat for the N¹-AII peptide relative to the results reported by Stewart et al. (30) for the porcine enzyme under similar assay conditions. However, we note that the total yield of the purified porcine enzyme reported in that study was less than 2μg, raising the possibility that inaccuracies in the quantification of very small amounts of protein used in that study may have led to an overestimation of the k_cat value. Nonetheless, it is also possible that the difference in catalytic activity may be due to the slight differences in the primary sequence between the human and porcine NTAN1 isoforms (92% amino acid sequence identity), or perhaps because the recombinant human enzyme may be lacking a factor present in liver tissue.

The recombinant hNTAN1 displayed a pH optimum in the range 6.0–7.5, with a sharp decline in activity upon shifting to more alkaline pH, indicating that protonation of a single group is important for catalysis. The pH dependence of hNTAN1 and its sensitivity to the alkylating agent iodoacetamide suggest that a reactive thiolate may have an important role in catalysis. PSI-BLAST blast analysis of hNTAN1 homologues revealed two highly conserved L-Cys residues. Site-specific mutagenesis subsequently demonstrated that the substitution of C75 with L-Ala completely abolishes activity, but does not affect the secondary structure of the enzyme as determined by circular dichroic spectra. Further, neither L-Ser nor L-Thr could replace C75 and concomitantly maintain enzyme function, revealing that this residue is absolutely required for catalysis. Therefore, similar to both the yeast Nta1 and the recently reported NTAQ1, the human NTAN1 appears to possess a critical nucleophilic L-Cys.

Although NTAN1 has been implicated in the modulation of protein turnover via the N-end rule pathway, very little is known with respect to: (i) the exact molecular mechanisms or interactive processes underlying the neurological defects observed in Ntan1^−/− mice (10), (ii) the catalytic mechanism and nature of the active site structure of NTAN1, which shares little sequence similarity to other amidases, and (iii) the identity of actual physiological substrates targeted by the enzyme.

While it is conceivable to scan the proteome for amino-terminal sequences of the type L-Met-L-Asn (or L-Met-L-Gln), the prediction of physiological substrates targeted by NTAN1 (or NTAQ1) is inherently complicated due to the potentially vast contribution of endoproteolysis in generating N-terminal degrons of these types. This point was illustrated by the discovery of NTAN1-mediated degradation of DIAP1 in the regulation of apoptosis (28). DIAP1 is cleaved by a caspase to generate a C-terminal fragment with N-terminal L-Asn, which is then targeted through the NTAN1-ATE1 pathway in ultimately generating a primary destabilizing N-terminal L-Arg for subsequent substrate protein degradation. To date, DIAP1 remains the only known substrate of N-terminal L-Asn deamidation. However, we describe here that hNTAN1 was unable to deamidate the N-terminal L-Asn of a synthetic 9mer corresponding to the N-terminus of the caspase-mediated C-terminal fragment (in N¹-DIAP), suggesting the possibility that NTAN1-mediated deamidation of the DIAP1 substrate in vivo may be dependent on additional interactions between the enzyme and other (unidentified) protein factors.

The enzymological characterization of N-end rule pathway components such as hNTAN1 should assist in the elucidation of their physiological roles, allow for the in vitro reconstruction of the pathway itself, and subsequently, aid in the discovery of potential physiological substrates targeted by its components. Further, recombinant expression of the human homologues of these components could present an opportunity to specifically investigate mechanisms and features of human pathological states associated with protein misfolding and aggregation, some of which may be associated with the N-end rule pathway.

Abbreviations

Ac-N¹-AII

N-acetyl-Asn¹Val⁵-angiotensin II (Ac-NRVYVHPF)

ACN

acetonitrile

ATE1

arginyltransferase 1

AII

Val⁵-angtiotensin II (DRVYVHPF)

CD

circular dichroism

DIAP1

Drosophila inhibitor of apoptosis 1

ESI-MS

electrospray ionization-mass spectrometry

IPTG

isopropyl β-D-1-thiogalactopyranoside

MAP

methionine aminopeptidase

N¹-AII

Asn¹Val⁵-angiotensin II (NRVYVHPF)

N¹-DIAP

9mer corresponding to the N-terminus of the caspase-mediated C-terminal fragment of DIAP1 (NNTNATQLF)

Nta1

N-terminal amidase 1

Nt-amidase

N-terminal amidohydrolase

NTAN1

N-terminal asparagine amidohydrolase

NTAQ1

N-terminal glutamine amidohydrolase

PSI-BLAST

Position-specific iterative BLAST

Q¹-AII

Gln¹Val⁵-angiotensin II (QRVYVHPF)

Ub

ubiquitin

UPS

ubiquitin-proteasome system