Identification of a nicotinamide/nicotinate mononucleotide adenylyltransferase in Giardia lamblia (GlNMNAT).

Giardia lamblia is an intestinal protozoan parasite that causes giardiasis, a disease of high prevalence in Latin America, Asia and Africa. Giardiasis leads to poor absorption of nutrients, severe electrolyte loss and growth retardation. In addition to its clinical importance, this parasite is of special biological interest due to its basal evolutionary position and simplified metabolism, which has not been studied thoroughly. One of the most important and conserved metabolic pathways is the biosynthesis of nicotinamide adenine dinucleotide (NAD). This molecule is widely known as a coenzyme in multiple redox reactions and as a substrate in cellular processes such as synthesis of Ca2+ mobilizing agents, DNA repair and gene expression regulation. There are two pathways for NAD biosynthesis, which converge at the step catalyzed by nicotinamide/nicotinate mononucleotide adenylyltransferase (NMNAT, EC 2.7.7.1/18). Using bioinformatics tools, we found two NMNAT sequences in Giardia lamblia (glnmnat-a and glnmnat-b). We first verified the identity of the sequences in silico. Subsequently, glnmnat-a was cloned into an expression vector. The recombinant protein (His-GlNMNAT) was purified by nickel-affinity binding and was used in direct in vitro enzyme assays assessed by C18-HPLC, verifying adenylyltransferase activity with both nicotinamide (NMN) and nicotinic acid (NAMN) mononucleotides. Optimal reaction pH and temperature were 7.3 and 26 °C. Michaelis-Menten kinetics were observed for NMN and ATP, but saturation was not accomplished with NAMN, implying low affinity yet detectable activity with this substrate. Double-reciprocal plots showed no cooperativity for this enzyme. This represents an advance in the study of NAD metabolism in Giardia spp.

Introduction

Nicotinamide adenine dinucleotide (NAD) is a key molecule in cellular metabolism [1]. Its central role as a coenzyme in REDOX reactions has been known since its function as an acceptor and donor of hydride ions was established [2], which is an essential characteristic for the maintenance of cellular REDOX balance and energy production. In addition, NAD is used as substrate in non-REDOX reactions, e.g., deacetylation by sirtuins [3], mono and poly-ADP-ribosylation [4], [5] and the synthesis of key second messengers in Ca2+ mobilization such as nicotinic acid-adenine dinucleotide phosphate (NAADP), cyclic ADP-ribose (cADP), ADP-ribose (ADPR) [6] and O-acetyl-ADP-ribose (OAADPR) [7]. These non-REDOX reactions are part of fundamental cellular processes, e.g., gene expression regulation, DNA repair, chromatin stability, calcium mobilization, circadian rhythm control and cell death [2], [8]. When used as a coenzyme in oxidation–reduction reactions, NAD exists in its oxidized form NAD+ and in its reduced form NADH, maintaining a constant concentration. However, in non-REDOX reactions in which NAD undergoes chemical breakdown, it is consumed, reducing its concentration.

Given this scenario, the cell must have NAD biosynthetic pathways to maintain a constant source of this molecule. There are two main NAD biosynthetic pathways (Fig. 1): one de novo pathway from quinolinic acid (QA) and a salvage pathway from nicotinamide (NAM), nicotinic acid (NA) and nicotinamide riboside (NR). Both the de novo pathway and the salvage pathway converge in NMNAT catalytic activity, which involves the Mg2+-dependent reversible condensation [9], [10] of ATP with NMN (nicotinamide mononucleotide) or NAMN (nicotinate mononucleotide), producing NAD (nicotinamide adenine dinucleotide) or NAAD (nicotinate adenine dinucleotide), respectively [11]. Ubiquitous in nature, this enzyme has been found in all organisms from all three kingdoms of life (archaea, eubacteria, eukaryotes) in which it has been sought, justifying the importance of its function in the cell.

Fig. 1

Representation of NAD biosynthesis. Enzymes involved are denoted by numbers: 1. QAPRT (quinolinic acid phosphoribosyltransferase, EC. 2.4.2.19). 2. NAPRT (nicotinic acid phosphoribosyltransferase, EC. 2.4.2.11). 3. Nicotinamidase (EC. 3.5.1.19). 4. NAMPRT (nicotinamide phosphoribosyltransferase, EC. 2.4.2.12). 5. NRK (nicotinamide riboside kinase, EC. 2.7.1.22). 6. NMNAT (nicotinamide/nicotinic acid mononucleotide adenylyltransferase, EC. 2.7.7.1/18). 7. NAD synthetase (EC. 6.3.5.1). Images drawn with ChemSketch Freeware (ACD/Labs v14.01).

The present study aims to examine NAD metabolism in the parasitic protozoan pathogen Giardia lamblia (also called Giardia intestinalis or Giardia duodenalis) through the identification, cloning, expression, purification and enzymatic validation of nicotinamide/nicotinate mononucleotide adenylyltransferase (NMNAT), which was identified using bioinformatics tools.

Materials and methods

Identification and bioinformatic validation of the putative sequence

Multiple sequence alignment of 16 NMNAT sequences reported in the UniProtKB database (http://www.uniprot.org/) was performed using the MUSCLE [12], [13] algorithm in the CLC Sequence Viewer v6.9 (Additional Alignments plugin v.1.4.5) program. The resulting consensus sequence of 244 aa was used to search the GiardiaDB [14] (http://giardiadb.org/giardiadb/) using the BLAST algorithm. The sequences with the highest score and lowest E-value were subsequently validated by finding conserved domains and their putative function. Additionally, a 3-dimensional model of the putative protein was obtained (I-TASSER server [15], [16]) to compare it with the structure of crystallized and reported NMNATs.

Culture of Giardia lamblia

Parasites (WB clone C6) were cultured in borosilicate tubes at 37 °C in TYI-S-33 medium (Diamond, 1978). Subsequently, the parasites were separated from the medium by centrifugation (1500 rpm, 10 min) and washed three times with cold sterile PBS. The washed parasites were resuspended in 6 mL of sterile PBS and counted in a Neubauer chamber. The parasites were centrifuged again, and the resultant pellet was stored at −20 °C to subsequently extract genomic DNA [17].

Genomic DNA extraction

Cells (7.5 × 107) were resuspended in 200 µL of sterile PBS. From these cells, genomic DNA was extracted using an Easy-DNATM kit (Invitrogen). The extracted DNA was resuspended in 100 µL of TE buffer and stored at 4 °C. Purity and concentration were evaluated by a UV–Vis spectrophotometer (Spectronic Genesys 5) at 260 and 280 nm.

PCR amplification of the coding region

The NMNAT coding region of Giardia lamblia was amplified from genomic DNA using the following primers: 5'-caccATGCCCTGTCCGCCGGT-3′(forward) and 5′-TTATTGGAAACTAGGGGGTG-3 (reverse). The amplification conditions were as follows: 3 mM MgCl2, 0.5 µM primers, 0.2 mM dNTPs, 2.5 U of Taq polymerase (Amplitaq) and 100 ng of gDNA template. PCR was performed in a final volume of 25 µL in a GeneAmp® PCR system 2400 Perkin Elmer™ thermal cycler. The optimal thermal cycle used was as follows: initial denaturation (94 °C, 5 min) followed by 30 denaturation cycles (94 °C, 45 s), annealing (55 °C, 45 s), and extension (72 °C, 1 min) and a final extension at 72 °C for 7 min. The PCR product was evaluated by horizontal electrophoresis in a 1.5% agarose gel in 0.5X TBE buffer. Staining was performed with 0.05% ethidium bromide for 20 min [18]. The gel was visualized on a Molecular Imager® Gel Doc™ XR image analyzer including Quantity One Basic 4.6.3 software from BIO-RADTM.

Cloning of the amplified coding region into an expression vector

The PCR product (glnmnat) was ligated into a pET100/D-TOPO® vector (Champion™ pET Directional TOPO® Expression Kits from Invitrogen™). The resulting recombinant plasmid (pET100/D-TOPO-glnmnat) was transformed by heat shock at 42 °C into competent E. coli cells (OneShot® TOP10 Competent Cells) [18]. The transformation was confirmed by colony PCR using the primers and thermal program previously indicated.

Expression vector purification

The pET100/D-TOPO-glnmnat expression vector was purified from a 5-mL culture (LB medium + ampicillin) of E. coli TOP10 cells previously transformed by the alkaline lysis procedure [18]. Purity and concentration were evaluated using a UV–Vis spectrophotometer (Spectronic Genesys 5) at 260 and 280 nm.

Transformation into expression strains

The previously extracted recombinant plasmid pET100/D-TOPO-glnmnat was transformed into chemically competent E. coli BL21 cells (Invitrogen) by heat shock at 42 °C [18].

Expression of the recombinant protein His-GlNMNAT

Isopropyl-β-d-1-thiogalactopyranoside (IPTG) was added to a 250-mL culture of E. coli BL21 (OD600nm = 0.550) for a final concentration of 1 mM. Induction was performed at 20 °C with constant stirring for 20 h. After the induction, the culture was centrifuged at 6000 rpm for 10 min at 4 °C, obtaining a bacterial pellet with a wet weight of 1.96 g. The bacterial pellet was resuspended in lysis buffer (50 mM NaH2PO4, 300 mM NaCl, 10 mM imidazole, pH 8.0) with 5 mL of buffer per gram of wet solid. Lysozyme was added to a final concentration of 1 mg/mL, as well as ribonuclease (Promega) and protease inhibitor cocktail (Sigma, P8340) at a dilution of 1:400. The mixture was homogenized by vortex stirring and incubated at 4 °C for 2 h. Subsequently, 6 ultrasonic pulses were applied for 15 s followed by a rest of 15 s. The soluble and insoluble fractions were separated by centrifugation at 12,000 rpm for 20 min at 4 °C [18].

SDS-PAGE and immunoblotting

The soluble and insoluble cell lysate fractions were evaluated by 12% polyacrylamide gel electrophoresis under denaturing conditions (SDS-PAGE) (Coomassie blue staining) [18]. A replica of the gel was transferred to a nitrocellulose membrane (0.45 µm thick; ThermoScientific) in transfer buffer (25 mM Tris base, 192 mM glycine, and 10% methanol v/v) at 200 mA for 2 h. The membrane was blocked overnight with a 5% w/v skim milk solution in TBST. Immunodetection of the recombinant protein was performed using a histidine tag with the primary antibody anti-His-mAb (1:3000, Amersham) and secondary antibody anti-mAb-HRP (1:2000); 4-chloro-naphthol (Sigma) and hydrogen peroxide were used as developing agents [19].

Nickel-affinity purification

Nickel-affinity resin solution (Ni-NTA, Qiagen) was used at 100 µL per milliliter of soluble protein extract. The resin was equilibrated in lysis buffer (50 mM NaH2PO4, 300 mM NaCl, 10 mM imidazole, pH 8.0) and incubated with soluble extract overnight on ice; the supernatant with unbound proteins was collected by centrifugation at 3000 rpm for 3 min. Washes were performed with wash buffer (50 mM NaH2PO4, 300 mM NaCl, 20 mM imidazole, pH 8.0) to obtain an absorbance at 280 nm of 0.000 ± 0005. The elution of the recombinant protein was performed with elution buffer (50 mM NaH2PO4, 300 mM NaCl, increasing concentrations of imidazole 50 mM/80 mM/250 mM, pH 8.0) on ice overnight. Protein purification was verified by SDS-PAGE and western blotting under previously described conditions [20].

Evaluation of enzyme activity

An in vitro assay was conducted to determine the activity of the purified recombinant protein (Table 1). For the negative control, the elution buffer was used instead of the recombinant protein eluate. For the positive control, an eluate with recombinant Homo sapiens NMNAT-3 protein (His-HsNMNAT-3) was used. The mixtures were incubated at 37 °C for 30 min, after which the reaction was stopped with 1.2 M HClO4 at 4 °C. The protein was precipitated by centrifugation at 12,000 rpm for 3 min at 4 °C. The supernatant was neutralized with 1 M K2CO3 on ice. A total of 120 µL of supernatant was taken and diluted to 240 µL with MQ H2O [21].

Table 1

Reaction mixture for enzyme assays.

Reagent	Volume (µL)	Final concentration
HEPES-KOH pH 7.5 (100 mM)/MgCl2 (40 mM)	25.0	25 mM/10 mM
ATP (12.5 mM)	10.5	1.31 mM
NMN/NAMN (50 mM)	2.5	1.25 mM
His-GlNMNATa	62.0	60 ng/µL

For the positive control, His-HsNMNAT-3 was used (final concentration in reaction mix: 25 ng/uL). For the negative control, soluble fraction of non-transformed E. coli B21 (DE3) cells was used.

The resulting product of the in vitro reaction was evaluated by reverse phase HPLC. A Phenomenex Luna C18 (250 mm × 4.60 mm, 5 µm) column was used. As mobile phases, 0.12 M potassium phosphate buffer pH 6.0 and methanol were used. The separation was carried out in a gradient (0–8 min 80% buffer 20% water, 8–12 min 0–20% MeOH 80% buffer, 12–14 min 20% MeOH 80% buffer, 14–18 min 20% MeOH 80% buffer, 18–20 min 80% buffer 20% water) with a constant flow of 1.5 mL/min, for 20 min per run. The analytes, whose retention times were determined from standards, were detected by spectrophotometry at 254.4 nm using a diode array detector (DAD).

Determination of kinetic constants

Optimal reaction parameters (pH, °T and time) were determined. Then, NAD production rate was evaluated as a function of substrate concentrations (ATP, NMN and NAMN). For each reaction, 1.9 µg of recombinant GlNMNAT were used. ATP concentration window: 0.2, 0.5, 0.8, 1.2, 1.5, 2, 4, 6, 8 and 10 mM, with NMN in excess (10 mM). NMN concentration window: 0.25, 0.5, 1, 2, 3, 4, 6 and 8 mM, with ATP in excess (10 mM). NAMN concentration window: 0.25, 1, 2, 3, 4, 6, 8, 10 and 12 mM, with ATP in excess (10 mM). Triplicates were done for each point in the windows. The data was processed by means of non-linear (two-parameter rectangular hyperbola) and linear regression (Hanes–Woolf). A double-reciprocal plot was made to check if the enzyme had cooperativity.

Results and discussion

Fig. 2 shows a multiple sequence alignment of 16 NMNATs of different organisms from which a consensus sequence of 244 aa was obtained. When searching with this sequence using BLASTP in the GiardiaDB (Giardia Assemblage A isolate WB), a sequence of 227 aa was found with a calculated size of 25.4 kDa and encoded by a coding DNA sequence (CDS) of 684 bp from chromosome 3 (Score: 75.5, E-Value: 2e-16, Gene ID: GL50803_13672/XP_001704508.1/GI:159108475, no introns, named GlNMNAT-A). Another sequence of 249 aa was identified, with a calculated size of 28.1 kDa and encoded by a CDS of 750 bp from chromosome 5 (Score: 77.4, E-value: 6e-17, Gene ID: GL50803_92618/XP_001706814.1/GI:159113174, no introns, named GlNMNAT-B). Although NMNAT identity for GlNMNAT-B has been attributed with bioinformatics tools and qualitative in vitro ADH-coupled enzyme tests performed by us, we have not included more information about this putative isoenzyme because it is less stable in the conditions we have tested for direct assays. For the remaining discussion, GlNMNAT-A will be considered as GlNMNAT.

Fig. 2

Multiple sequence alignment of 16 homologous NMNAT proteins from phylogenetically divergent organisms with GlNMNAT isoenzymes. The percentage of conservation is displayed throughout the sequence in bars. Alignment was done with the ClustalO algorithm in the CLC Sequence Viewer v7.0.2 program (CLCBio A/S, Additional Alignments plugin v.1.5.1).

Fig. 3 shows an alignment of the three human isoenzymes and the two sequences found for GlNMNAT, where the ATP motifs for recognition and binding are depicted (GxFxPx[H/T]xxH and ISSTxxR) [10]. This GlNMNAT sequences share 20–30% of identity with the human NMNATs and 28% between themselves.

Fig. 3

Alignment between the three NMNAT human isoenzymes and the two NMNAT isoenzymes from . Conservation percentage per residue position is observed in pink bars. The multiple alignment was done with ClustalO (CLC Sequence Viewer v7.0.2 program, CLCBio A/S, Additional Alignments plugin v.1.5.1). Identity percentages calculated with BLASTP algorithm (NCBI) and shown in the adjacent table. ATP active site motif for recognition and binding in N-terminus and C-terminus regions depicted in red (GxFxPx[H/T]xxH) and violet respectively (ISSTxxR) [10]. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

The results of the CD database (conserved domains, NCBI) showed that the sequence of 227 aa has the conserved domains of nicotinamide/nicotinate mononucleotide adenylyltransferase (cd09286, E-value: 1.97e–26; cd02165, E-value: 7.03e−26). Additionally, the InterProScan 4 (EMBL-EBI) tool revealed that the protein is a possible nicotinamide mononucleotide adenylyltransferase (PTHR12039, E-value: 3.1e–124; TIGR00482, E-value: 2.7e–46). The GO (gene ontology) terms predicted were GO:0009435 (NAD biosynthetic process) and GO:0016779 (nucleotidyltransferase activity).

Using the I-TASSER server, a tertiary structure model was built for the sequence found. Fig. 4a shows the first of the 5 models built by the server with the best quality parameters (RMSD, C-score, and TM-score). The Rossmann-type characteristic fold [22] of proteins that bind to nucleotides (βαβαβ motif) is observed. Fig. 4b shows a 3D alignment with the tertiary structure of HsNMNAT-3 (1nuq_A, 252 aa) generated from X-ray diffraction data (XRD) and recorded in the PDB protein structure database.

Fig. 4

GlNMNAT tertiary structure model. (A) Three-dimensional GlNMNAT model (I-TASSER [15], [16]). A Rossmann-type fold, characteristic of nucleotide-binding proteins, is observed. C-score: −5 < 0.95 < 2. TM-score: 0.84 ± 0.08 > 0.5. RMSD: 3.7 ± 2.5 Å. (B) Overlay of the model with the structure of the HsNMNAT-3 protein (1nuq_A [9], 252 aa, in yellow). Alignment performed with the UCSF Chimera v1.8 program [23]. RMSD between 171 atom pairs: 0.670 Å. Molecular graphics and analyses were performed with the UCSF Chimera package. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Several structural similarities are observed in various alpha helices and beta sheets of the overlapping proteins. Additionally, the I-TASSER server showed predictions for protein function. The EC number predicted from comparisons with proteins recorded in the databases was 2.7.7.1/18, consistent with adenylyltransferase function.

Based on the above observations, an in vitro characterization of the protein was performed. Fig. 5 shows the amplification of the NMNAT coding region of Giardia lamblia (glnmnat) with a size a little larger than 650 bp, which is consistent with the value of 684 bp reported in databases.

Fig. 5

Amplification of the NMNAT coding region of . (1) Ladder 100 bp (fermentas). (2) Negative control (without DNA template). (3) Amplification of glnmnat from genomic DNA. Agarose gel in 1.5% TBE, ethidium bromide staining at 0.05%.

The amplified sequence was cloned into the expression vector pET100/D-TOPO, which contains the upstream sequence for the T7 promoter and is inducible with IPTG. The expressed recombinant protein is joined at its N-terminal region by a histidine tag to facilitate affinity purification and to perform immunodetection assays. Fig. 6 shows an SDS-PAGE gel and immunoblotting of the induced protein in total cell lysate and in soluble and insoluble lysate fractions produced by ultrasonic pulses. All induced fractions show a single band with a size between 25.0 and 35.0 kDa, which is consistent with the calculated size of the protein of 25.4 kDa plus 4.1 kDa from the additional 36 aa in the N-terminal region corresponding to the His-tag, the enterokinase cleavage site and the X-press epitope, for a total of 29.5 kDa. The strong band in the insoluble fraction indicates that the recombinant protein is expressed in high amounts, folding incorrectly and aggregating in great part.

Fig. 6

Expression of recombinant His-GlNMNAT. (A) SDS-PAGE, Coomassie blue staining. (B) Immunoblotting on nitrocellulose membrane. (1) Molecular weight marker (MWM). (2) Uninduced BL21 cells. (3) Total cells induced. (4) Insoluble fraction. (5) Soluble fraction.

Fig. 7 shows the results of the SDS-PAGE separation and the immunodetection of the recombinant protein in the different fractions obtained during partial purification from the soluble fraction. Using nickel-affinity resin, it was possible to concentrate and separate the recombinant protein from most of the other soluble proteins, with a few higher molecular weight proteins present in the eluate. The protein was specifically detected in the eluate by immunoblotting, with undetectable amounts of the protein in the washes and in unbound proteins. The eluate was used for in vitro activity assays to experimentally verify the adenylyltransferase function of the recombinant protein. The eluate was stored at −80 °C and supplemented with glycerol to a concentration of 20% v/v for preservation.

Fig. 7

Enrichment of His-GlNMNAT recombinant protein from the soluble fraction by nickel affinity. (A) SDS-PAGE, Coomassie blue staining. (B) Immunoblotting on nitrocellulose membrane. (1) Molecular weight marker (MWM). (2) Soluble fraction. (3) Unbound proteins. (4) Wash 15. (5,6) Eluates 50 mM Imidazole. (7,8) Eluates 80 mM Imidazole. (9,10) Eluates 250 mM Imidazole.

Fig. 8 shows the results of the direct enzyme assay using nicotinamide mononucleotide (NMN) and nicotinic acid mononucleotide (NAMN) as substrates. Adenylyltransferase activity was detected with both substrates, showing the respective production of NAD and NAAD.

Fig. 8

Results of a direct enzyme assay using nicotinamide mononucleotide (NMN) or nicotinic acid mononucleotide (NAMN) as a substrate. NMN: (A) Positive control (B) Negative control (C) Reaction mixture with His-GlNMNAT. NAMN: (D) Positive control (E) Negative control (F) Reaction mixture with His-GlNMNAT.

Optimal pH and temperature were 7.3 and 26 °C respectively (graphs not shown). Additionally, a 30 min time was chosen for the subsequent end-point reactions (constant rate of NAD production detected between 0 and 60 min, graph not shown). For ATP and NMN, Michaelis–Menten curves were obtained, but for NAMN saturation was not observed, implying that the enzyme has low affinity yet detectable activity with this substrate in vitro. Double-reciprocal plots from NAD-production rates varying ATP and NMN concentrations (graphs not shown) showed no cooperativity of the enzyme. Table 2 summarizes the kinetic constants for the substrates that reached saturation. Fig. 9 shows the enzyme preparation used for the kinetic experiments, with an estimated purity of 76% based on densitometry and a 131-fold enrichment based on enzyme activity.

Table 2

Summary of kinetic constants for ATP and NMN.

	ATP		NMN
	RNL	HW	RNL	HW
Km (mM)	1.48	1.54	2.53	2.72
V max (mM∗min−1)	9.92E-02	9.82E-02	1.21E-01	1.24E-01
K cat (s−1)	2.44	2.41	2.97	3.04
K cat/Km (s−1∗mM−1)	1.65	1.56	1.17	1.12

NLR: non-linear regression. HWR: Hanes-Woolf regression. Temperature: 26 °C. pH: 7.5. Enzyme concentration: 6.79E-04 mM. Enzyme activity unit: 37.3 pkat/µg.

Fig. 9

GlNMNAT enzyme preparation for kinetic experiments. Purity estimated by means of densitometry using the Image J software. Enzyme enrichment calculated taking in account the activity of the soluble fraction and the eluate. Activity in pmol∗s−1∗µg−1 or pkat∗µg−1.

In comparison with the human cytosolic isoenzyme NMNAT-2, the GlNMNAT is less efficient in terms of affinity for its substrates (KmATP: 0.0889 [24], 0.107 [25], 0.204 [26]; KmNMN: 0.0213 [24]) and turnover (Kcat: 8.8 [24]), hence the parasitic need of G. lamblia for NAD from its host despite its ability to synthesize it from NMN. Based on this, Giardia would use the NMN salvage pathway for NAD biosynthesis, as well as direct intake of this molecule from the host, discarding a priori the salvage pathway from nicotinic acid mononucleotides and the more complex kynurenine pathway for de novo NAAD biosynthesis which would be very demanding for such a metabolically limited organism.

Conclusion

From bioinformatics and experimental evidence, the identification and functional validation of nicotinamide/nicotinate mononucleotide adenylyltransferase in Giardia lamblia (GlNMNAT) is reported in this paper, which represents an advance in the study of the metabolism of nicotinamide and adenine dinucleotide (NAD) in Giardia spp.

Conflict of interest

We certify that there is no conflict of interest with any organization regarding the material discussed in the manuscript.