Evidence that Human Histidine Triad Nucleotide Binding Protein 3 (Hint3) is a Distinct Branch of the Histidine Triad (HIT) Superfamily

J. Mol. Biol. (2007) 373, 978–989

doi:10.1016/j.jmb.2007.08.023

Evidence that Human Histidine Triad Nucleotide Binding Protein 3 (Hint3) is a Distinct Branch of the Histidine Triad (HIT) Superfamily Tsui-Fen Chou 1 , Jilin Cheng 1 , Ilya B. Tikh 1 and Carston R. Wagner 1,2 ⁎ 1

Department of Medicinal Chemistry, University of Minnesota, 8-174 Weaver Densford Hall, 308 Harvard St. S.E., Minneapolis, MN 55455, USA 2

Department of Chemistry, University of Minnesota, Minneapolis, MN 55455, USA Received 20 June 2007; received in revised form 3 August 2007; accepted 10 August 2007 Available online 21 August 2007

Human Hint3 (hHint3) has been classified as a member of the histidine triad nucleotide (Hint) binding protein subfamily. While Hint1 is ubiquitously expressed by both eukaryotes and prokaryotes, Hint3 is found only in eukaryotes. Previously, our laboratory has characterized and compared the aminoacyl-adenylate and nucleoside phosphoramidate hydrolase activity of hHint1 and Escherichia coli hinT. In this study, hHint3-1(Ala36) and its single nucleotide polymorphism, hHint3-2 (A36G variant), were cloned, overexpressed, and purified. Steady-state kinetic studies with a synthetic fluorogenic indolepropinoic acyl-adenylate (AIPA) and with a series of fluorogenic tryptamine nucleoside phosphoramidates revealed that hHint31 and hHint3-2 are adenylate and phosphoramidate hydrolases with apparent second-order rate constants (kcat/Km) ranging from 102 to 106 s− 1 M− 1. Unlike hHint1, hHint3-1 and hHint3-2 prefer AIPA over tryptamine adenosine phosphoramidate by factors of 33- and 16-fold, respectively. In general, hHint3s hydrolyze phosphoramidate 370- to 2000-fold less efficiently than hHint1. Substitution of the potential active-site nucleophile, His145, by Ala was shown to abolish the adenylate and phosphoramidate hydrolase activity for hHint3-1. However, 0.2–0.4% residual activity was observed for the H145A mutant of hHint3-2. Both hHint3-1 and hHint3-2 were found to hydrolyze lysyl-adenylate generated by human lysyl-tRNA synthetase (hLysRS) by proceeding through an adenylated protein intermediate. hLysRS-dependent labeling of hHint3-1 and hHint3-2 was found to depend on His145, which aligns with the His112 of the Hint1 active site. The extent of active-site His145-AMP labeling was shown to be similar to His112-AMP labeling of hHint1. In contrast to all previously characterized members of the histidine triad superfamily, which have been shown to exist exclusively as homodimers, wild type and the H145A of hHint3-1 were found to exist across a range of multimeric states, from dimers to octamers and even larger oligomers, while wild type and the H145A of hHint3-2 exist predominantly in a monomeric state. The differences in oligomeric state may be important in vivo, because unlike tetracysteine-tagged Hint1, which was found along linear arrays exclusively in the cytoplasm in transfected HeLa cells, tagged Hint3-1 and Hint3-2 were found as aggregates both in the cytosol and in the nucleus. Taken together, these results imply that while Hint3 and Hint1 prefer aminoacyl-adenylates as substrates and catalytically interact with aminoacyl-tRNA synthetases, the significant differences in phosphoramidase activity, oligomeric state, and cellular localization

*Corresponding author. E-mail address: [email protected]. Abbreviations used: Hint, histidine triad nucleotide binding protein; AIPA, indolepropinoic acyl-adenylate; LysRS, lysyl-tRNA synthetase; HIT, histidine triad; AARS, aminoacyl-tRNA synthetase; CEM, human T-lymphoblast leukemia cell line; PBMC, peripheral blood mononuclear cell; DHFR, dihydrofolate reductase; MTX, methotrexate; PPase, pyrophosphatase; SEC, size exclusion chromatography; SLS, static light scattering; ESI, electrospray ionization; MS, mass spectroscopy; HBSS, Hank's balanced salt solution; BAL, (2,3-dimercapto-1-propanol). 0022-2836/$ - see front matter © 2007 Elsevier Ltd. All rights reserved.

hHint3 is a Branch of the HIT Superfamily

979

suggest that Hint3s should be placed in a distinct branch of the histidine triad superfamily. © 2007 Elsevier Ltd. All rights reserved.

Edited by F. Schmid

Keywords: human Hint1; human Hint3-1; human Hint3-2; lysyl-tRNA synthetase; phosphoramidase

Introduction Histidine triad nucleotide binding protein (Hint) belongs to a histidine triad (HIT) superfamily that contains a characteristic C-terminal active-site motif, His-X-His-X-His-XX, where X represents hydrophobic residues.1 HIT superfamily consists primarily of nucleoside phosphoramidases, dinucleotide hydrolases, and nucleotidyl transferases.1 The HIT proteins have been classified into five subfamilies according to their enzymatic function, sequence composition, and structural similarity. The Hint branch is the most ancient and can be found in all three kingdoms of life. In addition to their nucleoside phosphoramidase activity,2–5 Hint1 isolated from human and Escherichia coli has recently been shown to be an efficient aminoacyl-adenylate and acyl-adenylate hydrolase.6,7 The regulation of aminoacyl-tRNA synthetase (AARS)mediated aminoacyl-adenylate formation as well as AARS transcription factor association has been proposed as a possible physiological role for Hints.8–11 Hint1 knockout mice have been shown to exhibit normal embryonic development. However, at 2–3 years of age, both heterozygous and homozygous mice have been found to have an increased susceptibility to the induction of ovarian and mammary tumors by the carcinogen 7,12-dimethylbenzanthracene and to spontaneous tumors.12 Up-regulation of Hint1 and the significantly reduced in vivo tumorigenicity of 5-aza-deoxycytidine-treated human non-small-cell lung cancer cell line NCI-H522 demonstrated that hHint1 might be a tumor suppressor.13 The recent observation that hHint1 is involved in the modulation of apoptosis, independent of its enzymatic activity, suggested a possible mechanism for its tumor suppressor activity.14 Furthermore, the ability of overexpressed Hint1 to inhibit cell growth and activator protein-1 activity in the human colon cancer cell line SW480 supports its potential tumor suppressor function.15 Search results from the Basic Local Alignment Search Tool program indicated that E. coli contains only one hinT gene, whereas four human Hint genes have been identified (hHint1, hHint2, hHint3, and hHint4). Human Hint2 has recently been shown to be a mitochondrial apoptotic sensitizer that is downregulated in hepatocellular carcinomas.16 hHint3 and hHint4 have not been characterized. Nevertheless, upon DNA sequence analysis, we have concluded that hHint3 and hHint4 have been incorrectly annotated and are in fact identical genes†. † The National Center for Biotechnology Information (NCBI), http://www.ncbi.nlm.nih.gov/

Human Hint3 protein was identified as a member of the Hint branch based on its probable activesite sequence similarity to hHint1, even though its total amino acid sequence identity is only 28% (Fig. 1).1 Since it has a slightly higher sequence identity (31%), it has also been classified as a member of the Aprataxin subfamily. This sequence identity is similar to that observed between hHint1 and human fragile histidine triad (Fhit), which is another HIT subfamily. In addition, a linkage between the mRNA expression levels of the Hint3 and Hint1 genes in mice has not been observed.17 Thus, the physiological function and subfamily membership of Hint3 remain ambiguous. Consequently, since little is known of the cellular function of Hint3, we have chosen to compare hHint3, hHint1, and Aprataxin to characterize their catalytic activity and molecular organization. If hHint3 is indeed similar to hHint1, it should be a homodimer and exhibit similar levels of phosphoramidate and acyl-adenylate hydrolase activity. If it is similar to Aprataxin, hHint3 should exhibit Ap4A hydrolase activity in addition to phosphoramidate hydrolase activity.18–20

Results Sequence analysis of hHint3 To isolate and characterize hHint3, PCR was carried out using Hint3-specific primers with cDNA libraries derived from human T-lymphoblast leukemia cell line (CEM) and peripheral blood mononuclear cells (PBMCs). DNA sequence analysis of the PCR products revealed that Hint3 obtained from CEM has the same sequence as gi:21359981 (Ala36)†, whereas the sequence of Hint3 from PBMCs corresponded to previously deposited human genome sequence results (AAH15732, BC015732, CAB92728)†. One additional human Hint3 sequence containing Ala36 and a deletion of the residues 68–75 has been identified†. Hereafter we have named the hHint3 as hHint3-1(Ala36), hHint3-2(Gly36), and hHint3-3 (Ala36, Δ68–75) (Fig. 1)†. Expression and purification of hHint3 proteins Previously, we have developed a protein expression protocol for hHint1 as a fusion protein to E. coli dihydrofolate reductase (ecDHFR).4,7 This approach facilitates the isolation of Hints that are unable to be purified by AMP affinity chromatography by taking advantage of the binding of ecDHFR to a methotrexate (MTX) affinity column. Two potential throm-

hHint3 is a Branch of the HIT Superfamily

980

Fig. 1. Sequence alignment of hHint1 and hHint3s. Yellow shading indicates identity between hHint1 and hHint3s; the conserved histidine residues are underlined. Among 101 alignable sequences, two proteins shared 28% sequence identity and 47% positive similarity. The sequence was downloaded from the NCBI Website†. Multiple sequence alignment was performed using the homology modeling module of the INSIGHTII software package (Accelrys, Inc.).

bin recognition sites were found in hHint3 sequence so the thrombin cleavable linker in pPH70D plasmid was replaced with the PreScission™ protease cleavable linker (GLGGGGGLVPRGTLEVLFQ/ GPLE) where the recognition sequences were underlined with cleavage occurring between Gln and Gly. Human Hint3-1 and -2 could only be isolated as DHFR fusion proteins by MTX affinity chromatography, since initial attempts to use AMP affinity chromatography and His-tag purification were unsuccessful (data not shown). His6-tag hHint3-1 and -2 expressed by E. coli could only be found in inclusion bodies and attempts to refold hHint3-1 and -2 were unsuccessful (data not shown). Therefore, successful isolation of hHint3-1 and -2 as ecDHFR-fusion proteins indicated that ecDHFR was likely to assist correct folding of hHint3-1 and 2 in bacteria. In addition, ecDHFR tag was easily separated from hHint3-1 and -2 by a diethylami-

noethyl ionic exchange column after PreScission™ protease digestion (Supplementary Fig. 1). The catalytic His145 was replaced with Ala for both hHint3-1 and -2 by using site-directed mutagenesis and purified as described for wild type. SDS-PAGE and HPLC–electrospray ionization (ESI) + mass spectroscopy (MS) analysis confirmed the homogeneity of the purified proteins (Supplementary Figs. 2 and 3 and Table 2). Steady-state kinetic characterization of hHint3s The Michaelis–Menten constants kcat and Km were determined for hHint3-1 and -2 with indolepropinoic acyl-adenylate (AIPA)7 and a series of phosphoramidate substrates 5 (Tables 1 and 2 and Supplementary Fig. 4). Hydrolysis of the fluorogenic adenylate revealed similar kcat values for hHint1,7 hHint3-1, and hHint3-2, while the Km values for

Table 1. Comparison of steady-state kinetic parameters of hydrolysis of a fluorogenic adenylate

Enzyme hHint1a hHint3-1 (Ala36) hHint3-2 (Gly36) a

Km (μM)

kcat (s− 1)

kcat/Km (s− 1 M− 1)

Ratio of kcat/Km

0.04 ± 0.002 0.58 ± 0.15 1.5 ± 0.4

1.98 ± 0.02 1.32 ± 0.3 1.77 ± 0.3

(5.3 ± 0.5) × 107 (2.3 ± 0.7) × 106 (1.3 ± 0.7) × 106

1 0.04 0.03

Data for hHint1 were adapted from Ref. 7.

hHint3 is a Branch of the HIT Superfamily

981

Table 2. Comparison of steady-state kinetic parameters of hydrolysis of fluorogenic tryptamine nucleoside phosphoramidate monoesters

kcat (s− 1) Compound no. 1 2 3 4 5

R1 Adenine Guanine Hypoxanthine Uracil Cytosine

Km (μM) hHint1a

hHint3-1 (Ala36) hHint3-2 (Gly36) 0.31 ± 0.01 0.31 ± 0.01 0.33 ± 0.03 0.11 ± 0.001 0.023 ± 0.001

0.52 ± 0.01 0.59 ± 0.01 0.36 ± 0.01 0.14 ± 0.02 0.06 ± 0.03

2.1 ± 0.1 2.3 ± 0.1 2.6 ± 0.04 2.5 ± 0.3 1.2 ± 0.1

hHint3-1 (Ala36) hHint3-2 (Gly36) 16 ± 1 28 ± 2 58 ± 8 164 ± 1 84 ± 2

22 ± 2 29 ± 2 32 ± 2 121 ± 20 181 ± 86

hHint1a 0.13 ± 0.02 0.21 ± 0.02 0.71 ± 0.03 2.2 ± 0.4 2.3 ± 0.4

kcat/Km (s− 1 M− 1) × 10− 3 Compound no. 1 2 3 4 5

R1 Adenine Guanine Hypoxanthine Uracil Cytosine

a

hHint3-1 (Ala36) hHint3-2 (Gly36) 20 ± 2 11 ± 1 5.7 ± 1.3 0.66 ± 0.01 0.27 ± 0.01

23 ± 3 20 ± 2 11 ± 1 1.2 ± 0.4 0.33 ± 0.32

hHint1a 15,000 ± 3000 11,000 ± 1000 3700 ± 300 1200 ± 500 600 ± 200

Data for hHint1 were from Ref. 5.

hHint3-1 and hHint3-2 were 10- and 35-fold greater, respectively. The kcat values for the hHint3-1 and -2 hydrolysis of purine phosphoramidates were found to be 7- and 4-fold less than for hHint1, while the Km values were shown to be 100-fold greater than the value observed for hHint1. Similar to hHint1, hHint3s prefer purine over pyrimidinebased substrates.5 Specific activity of H145A-hHint3 mutants Based on the active-site signature of the HIT superfamily, the putative nucleophile, His145, of hHint3-1 and hHint3-2 was mutated to Ala. Indeed, this mutation abolished hHint3-1 activity completely, whereas the H145A-hHint3-2 mutant was found to exhibit 0.2–0.4% residual activity. The level of activity observed for the H145A-hHint3-2 mutant was similar to the residual activity found for the H112A-hHint1 mutant (Fig. 2 and Table 3). Adenylation of Hint proteins by hLysRS Recently, we have shown that lysyl-adenylate generated by LysRS is a native substrate for human and E. coli Hint1.6 During the reaction, the activesite nucleophile, His112, of hHint1 is adenylated in situ. Therefore, the ability of hHint3-1 and -2 to utilize lysyl-adenylate as a substrate was examined by incubations with hLysRS, lysine, and [α-32P]ATP. As shown in Fig. 3a, the intensity of the labeling of hHint3-1 (Ala36) is similar to that observed for hHint1 (lanes 2 and 4). Consistent with the ability of

pyrophosphate to act as a product inhibitor of AARS by either promoting the reverse reaction21 or blocking product release,22 we chose to mimic in vivo aminoacylation conditions23,24 by adding inorganic pyrophosphatase (PPase) to the reaction mixture. Enhanced labeling of both hHint3-1 and -2 by factors of 3 and 2, respectively, was observed upon addition of PPase (Fig. 3b). However, unlike hHint1, a very small amount (1/70 to 1/50) of active-site independent labeling was observed in the absence of hLysRS, with [α-32P]ATP, [α-32P]GTP, and [γ-32P] ATP, indicating that hHint3s may have cryptic phosphatase activity (Fig. 3b and Supplementary Fig. 5b). Size exclusion chromatography of hHint3s Superdex 75 and 200 size exclusion chromatography (SEC) columns with an in-line fluorescence detector were employed to examine the apparent molecular weight of recombinant purified proteins. As shown in Fig. 4a, hHint3-1 was found to exist in solution as monomeric, dimeric, tetrameric, octameric, and larger species, while the H145A mutant was found to form larger amounts of the higher molecular weight assemblies in solution (Fig. 4c). In the case of hHint3-2 and its H145A mutant, the monomeric species was shown to be predominant in solution followed by smaller amounts of dimeric, tetrameric, and octameric species (Fig. 4b and d). SEC with an in-line static light-scattering (SLS) detector was employed to assess the apparent molecular mass of the H145A-hHint3-1 species.

hHint3 is a Branch of the HIT Superfamily

982

in the far-UV region (190–260 nm) and the results were expressed as the mean residue ellipticities. As can been seen in Fig. 6, the CD spectra for hHint3-1 (black square) and the H145A-hHint3-1 (yellow diamond) were found to be slightly different compared to the 190- to 210-nm region. In contrast, for hHint3-2, wild-type, and the H145A mutant, spectra were found to be superimposable and consistent with a nearly identical combination of α-helical and β-sheet secondary structural elements, whereas when both wild-type hHint3-1 and hHint32 were compared, the CD spectra displayed only modest differences in their secondary structural composition. HPLC-ESI+ MS analysis To ensure the observed difference of hHint3-1 and -2 in their oligomerization state and secondary structure composition were not due to impurities from the protein preparation, HPLC-ESI+ MS analysis was employed to demonstrate purity and average molecular masses of recombinant purified proteins in this study (Supplementary Fig. 3 and Table 2). Only monomeric hHint3 was observed with a molecular mass within 0.001% of the calculated mass. Intracellular localization of hHint3s

Fig. 2. Residual phosphoramidate and adenylate hydrolase activity of H145A-hHint3-2. (a) H145AhHint3-1 or (b) H145A-hHint3-2 (500 pmol) was incubated with 50 μM substrate, TpAd (blue), and GpAd (yellow). (c) Reactions of the fluorogenic adenylate (AIPA, 50 μM) with H145A-hHint3-1 (black) or H145A-hHint3-2 (red) were monitored for 30 min at 25 °C.

The molecular masses for the large oligomers ranged from 1000 to 1800 kDa (Fig. 5). Our apparent inability to observe hHint3s as either solely monomers or as a well-defined oligomer, therefore, contrasts markedly with Hint1s, which are found as native homodimers, unless engineered to be monomeric.7 Secondary structure analysis To assess potential differences in hHint3 secondary structure, CD spectra of the wild-type and H145A mutants of hHint3-1 and -2 were determined

HeLa cells were transfected with plasmids expressing hHint1, hHint3-1, and hHint3-2 as fusion proteins to the CCPGCC tetracysteine peptide. After incubation with the cell-permeable arsenical, FlAsH,25,26 overexpression of each protein could be readily observed (Fig. 7a–d). hHint1 was found to localize in cytoplasmic linear array, while hHint3-1 and hHint3-2 were found to localize in the cytoplasm as amorphous aggregates and microaggregates (Fig. 7a–c). In contrast to hHint1, small amounts of nuclear localized hHint3-1 and hHint3-2 could be observed (Fig. 7b and c).

Discussion The highly conserved Hint1s are dimeric proteins and have been shown to be efficient nucleoside phosphoramidases 2–5 and aminoacyl-adenylate hydrolases.6,7 hHint2, which is 60% identical with hHint1, is also a homodimeric protein and nucleoside phosphoramidase.16 Hint3, however, is not found in prokaryotes. When compared to other Table 3. Summary of the specific activity (nmol s− 1 nmol− 1) of the recombinant purified H112A-hHint1, H145A-hHint3-1 and H145A-hHint3-2 Substrate

H112A-hHint1

H145A-hHint3-1

H145A-hHint3-2

AIPA TpAd TpGd

(2.0 ± 0.1) × 10− 4 (3.0 ± 0.3) × 10− 4 (6.6 ± 0.1) × 10− 4

0 0 0

(4.5 ± 0.2) × 10− 3 (1.0 ± 0.1) × 10− 3 (4.0 ± 0.1) × 10− 4

hHint3 is a Branch of the HIT Superfamily

983

Fig. 3. Adenylation of Hint proteins by hLysRS. 32P-Labeling reaction was carried out with hLysRS (0.9 μM) in buffer A containing lysine (28.5 μM) and [α-32P]ATP (0.9 μM) at 23 °C for 1 min with or with out PPase at 23 °C for 1 min followed by addition of Hint proteins (3.6 μM) for 1 min. (a) A similar amount of hHint1 (lane 2) and hHint3-1 (lane 4) was labeled. (b) hLysRS dependent labeling is on the active-site His145 and PPase enhanced the labeling intensity.

members of the HIT family, little sequence identity was observed, with the exception of hHint1 (28%) and Aprataxin (31%). Aprataxin is an AMP-lysine phosphoramidase and Ap3A and Ap4A hydrolase that resides in the nucleus and is associated with DNA repair.19,20 Although a cellular function for hHint114 and hHint216 has been recently proposed to involve in the regulation of apoptosis, the physiological role of hHint3 remains a mystery. Therefore, because of the inability of sequence analysis to unambiguously place hHint3 into one of the branches of the HIT superfamily, we hypothesized that hHint3 may constitute a unique branch. Consequently, we have cloned, developed the first expression system, and characterized the oligomeric state, enzymatic activity, and cellular location of hHint3. Isolation of hHint3 DNA from cDNA libraries prepared from PBMCs and CEM by PCR with genespecific primers revealed two variants of hHint3. A single nucleotide polymorphism from 106 to 108 in

the 107 base of the codon produced the point variants Ala36 and Gly36. Both genes were cloned and several attempts to purify them using the AMP affinity column2,4 and His-tag approach failed due to the insolubility of overexpressed hHint3-1 and hHint3-2 proteins in E. coli (data not shown). When both hHint3s were expressed as ecDHFR fusion proteins,27 milligram quantities of both enzymes could be obtained. To determine whether hHint3s can hydrolyze synthetic acyl-adenylate (AIPA) and nucleoside phosphoramidates, steady-state kinetics were carried out with the fluorescence-based assay as previously described.5 For hydrolysis of the adenylate, the kcat/Km values for hHint3-1 and hHint3-2 are 20- to 40-fold less than that for hHint1 (Table 1). While the Km values were one order of magnitude higher than that of hHint1, the kcat values were found to be similar. For hydrolysis of purine phosphoramidates, Km values for hHint3-1 and hHint3-2 were 100-fold higher than that for hHint1, while the

984

hHint3 is a Branch of the HIT Superfamily

Fig. 4. Protein oligomerization analysis by SEC with in-line fluorescence detector. SEC chromatograms from the Superdex 75 Column of wild-type hHint3-1 (a) and hHint3-2 (b). The SEC chromatograms from the Superdex 200 Column of wild-type and H145A-hHint3-1 (c) and H145A-hHint3-2 (d). Data were recorded with a fluorescence detector (excitation at 285 nm and emission at 340 nm).

kcat values are 4- and 7-fold lower. Therefore, the kcat/Km values for the hHint3s are 1000-fold lower than those for hHint1 (Table 2). When substrate specificity was compared, hHint3s were found to be similar to hHint1 (Table 2, Supplementary Fig. 4) with a preference for purine over pyrimidine

phosphoramidate substrates.5 Substitution of the putative hHint3-1active-site His145 with Ala completely abolished its hydrolase activity. Unexpectedly, the same mutant for hHint3-2 retained 0.2% to 0.4% activity when compared to wild type (Fig. 2 and Table 3). Thus, unlike hHint1, hHint3s exhibited

Fig. 5. SEC with in-line multiangle SLS detector. SEC chromatograms from the Superdex G200 column equipped with UV (absorbance of 280 nm) and SLS detector was employed to analyze H145A-hHint3-1 (50 μM, 200 μL). Molecular masses of protein eluted at 16.5 to 20.7 min range from 1000 to 1800 kDa.

Fig. 6. CD spectroscopy. CD Spectra of the wild-type and H145A mutants of hHint3-1 and -2 were collected in the far-UV region from 190 to 260 nm at 23 °C with protein concentration of 7.5 μM. The results are expressed as the mean residue ellipticities. hHint3-1 (black squares) and the H145A-hHint3-1 (yellow diamonds) were found to be slightly different in the region compared to the 190- to 210-nm region. In contrast, for the wild-type and the H145A-hHint3-2 mutant, spectra were found to be superimposable.

hHint3 is a Branch of the HIT Superfamily

985

Fig. 7. Localization of hHint1 and hHint3 proteins in live cells. Fluorescence confocal microscopy of HeLa live cells stably transfected with CCPGCC tetracysteine-tagged (a) hHint1, (b) hHint3-1, (c) hHint3-2, and (d) nontransfected cells were labeled with FlAsH-EDT2 (2 μM) in HBSS for 1 h at 23 °C (left panels), then washed with BAL (250 μM) twice before imaging. The nuclei were stained with Hoechst 33342 dye (1 μg/mL) for 15 min at 23 °C (middle panels). Right panels show merged images. Fluorescence images were acquired using a Fluoview FV1000 microscope (Olympus).

a distinct preference for acyl-adenylate-based substrates over phosphoramidates with very little difference in the catalytic efficiencies conferred by the polymorphic side-chain variability at position 36. To evaluate whether lysyl-adenylate generated by hLysRS could be a substrate for hHint3-1 and hHint3-2, adenylation of the enzymes by incubation with hLysRS, lysine, and [α-32P]ATP was examined as described previously.6,7 As can be seen in Fig. 3a, a similar extent of hHint3s and hHint1 labeling in the presence of hLysRS was

observed. In addition, lysine-dependent adenylation of hHint3-1 by hLysRS was consistent with the requirement of lysyl-adenylate for the observed hHint3-1 labeling (Supplementary Fig. 5a). Replacement of the putative active-site nucleophile, His145, abolished labeling of both hHin3-1 and hHint3-2 (Fig. 3b). However, the incubation of [32P]ATP and [32P]GTP with either hHint3 or the counterpart H145A mutants revealed some residual labeling at apparently an alternative site (Supplementary Fig. 5b). As observed for hHint1,6 the removal of pyropho-

986 sphate by PPase during the reaction enhanced labeling of hHint3s. Moreover, in contrast to Aprataxin, but similar to hHint1, neither hHint31 nor hHint3-2 exhibited measurable Ap3A and Ap4A hydrolase activity (T.-F. Chou and C. R. Wagner, unpublished data). Thus, both hHint3s and hHint1 are able to hydrolyze lysyl-AMP generated by hLysRS. Whether other aminoacyladenylates generated by AARSs will be substrates for Hints with different specificities remains to be determined. The apparent molecular weight of hHint3-1 and hHint3-2 was examined by SEC and static light scattering. These experiments demonstrated that wild-type and the H145A mutant of hHint3-1 tend to form oligomers, while those for hHint3-2 were found to predominantly exist as monomers (Figs. 4 and 5). The exclusive homodimeric structure of Hints appears to be violated by hHint3. To confirm the observed oligomerization was not due to an unforeseen impurity, the purity of the hHint3-1 and hHint3-2 was verified by HPLC-ESI MS and no contaminating proteins were observed (Supplementary Fig. 3 and Table 2). To further investigate the possible contribution of the observed different oligomerization states between hHint3-1 and hHint3-2, CD spectroscopy was used to compare the secondary structures of wild-type and H145A mutant hHint3s. As shown in Fig. 6, only minor differences in the secondary structure were observed. Thus, although the sidechain polymorphism and active-site histidine substitutions dramatically affect the oligomeric state of hHint3, these differences appear not to translate into significant differences between the proteins. Stable expression of each hHint3 in HeLa cells demonstrated that they are localized in both the cytoplasm and, to a lesser degree, the nucleus. In contrast, although previous immunofluorescence studies with fixed cells suggested hHint1 could be found both in the cytoplasm and in the nucleus,28 we were only able to demonstrate cytoplasmic localization with live cells, which is consistent with the results of previous hHint1-EGFP localization.14 The observed nuclear localization or hHint3s is reminiscent of Aprataxin. Although a rationale for the effect of Ala and Gly at position 36 on the oligomerization state of the proteins remains to be determined, the cellular localization appears not to be affected. Thus, although hHint3 appears to be similar to hHint1, the ability of hHint3 to assemble into multimeric oligomers and to significantly prefer acyl-adenylates over phosphoramidates establishes a clear demarcation between Hint3 and Hint1. Nevertheless, although hHint1 and hHint3 differ in their cellular distribution, they are both able to hydrolyze lysyl-AMP generated by hLysRS by proceeding through an adenylated-histidine intermediate (Fig. 8). Since LysRS is found only in the cytoplasm, it would be available to interact with both proteins. Moreover, despite a slightly greater

hHint3 is a Branch of the HIT Superfamily

Fig. 8. Proposed mechanism of lysyl-AMP Hydrolysis. Both hHint1 and hHint3 hydrolyze lysyl-AMP generated by LysRS by proceeding through an enzyme His-AMP intermediate. In both cases, active enzyme is regenerated by hydrolysis of the enzyme intermediate.

sequence similarity to Aprataxin than Hint1, the inability of hHint3 to hydrolyze Ap4A suggests that hHint3 and Aprataxin are not members of the same HIT branch. Consequently, we propose that Hint3 is neither a member of the Hint1 nor the Aprataxin branch, but instead should be placed in a distinct branch of the HIT superfamily.

Materials and Methods Plasmid construction Plasmids expressing hHint3-1 and -2 were constructed by replacing the hamster polymorphic N-acetyltransferase 2 (NAT2) cDNA in pPH70D with the desired open reading frames.27 Human Hint3-1 and Hint3-2 cDNA were obtained after PCR amplification with primers, Hint3F and Hint3R containing XhoI and XbaI sites, and cDNA libraries were prepared from the CEM and PBMCs, respectively. The cDNA libraries were generated as described.4 All primer sequences are listed in Supplementary Table 1. The double-digested PCR products and pPH70D were ligated using T4 DNA ligase (Invitrogen) to result in the plasmids pJLCH3-1 and pJLCH3-2. The thrombin cleavable linker was exchanged with PreScission™ protease cleavable linker (GLGGGGGLVPRGTLEVLFQ/GPLE) by PCR and ligation using primers PresF and PresR to create pJLCH3-1-Pres and pJLCH3-2Pres. pJLCH3-1-Pres and pJLCH3-2-Pres plasmids were subjected to site-directed mutagenesis using primer H145AF and H145AR to create active-site His145 to Ala mutants. Protein expression and purification To avoid E. coli hinT contaminant, pJLCH3-1-Pres and pJLCH3-2-Pres plasmids were transformed into BB2 (echinT disrupted BL21 star strain).4 A single colony was inoculated into LB medium (10 mL) containing carbenicillin (50 μg/L) and chloramphenicol (15 μg/L) and grown at 37 °C overnight. Starter culture (10 mL) was inoculated into a 2-L flask containing LB medium (1 L), ampicillin (100 μg/L), and chloramphenicol (15 μg/L). Cell culture was grown at 37 °C until OD600 reached 0.5 at which IPTG was added to a final concentration of 0.2 mM. Cells were harvested after 4-h growth at 37 °C by centrifugation at 6000g for 30 min at 4 °C. The fusion proteins were purified with MTX affinity chromatography followed by cleavage with PreScission™ Protease (10 U/mg, GE Healthcare) to

hHint3 is a Branch of the HIT Superfamily release the native hHint3 protein and ecDHFR. The protease digest reaction mixture was loaded onto a diethylaminoethyl ionic exchange column to purify hHint3 from ecDHFR as described.7 Steady-state kinetic measurement with the fluorescent assay Steady-state kinetic studies were carried out with fluorogenic adenylate and phosphoramidate substrates as described.5,7 Excitation wavelength was set at 280 nm, fluorescence emission was measured at 360 nm, and excitation and emission slits were set at 10 nm for concentration of the adenylate substrate ranging from 0.5 to 2 μM in Hepes buffer (20 mM, pH 7.2) or 5 nm for concentration ranging from 10 to 50 μM in Hepes buffer containing 1 mM MgCl2. The fluorescence intensity was monitored for 2 min to obtain the baseline and to allow the temperature to stabilize at 25 °C, and then enzyme (2 or 50 pmol) was added to initiate reactions. The Michaelis– Menten constants, kcat (s− 1) and Km (μM), were determined by JumpIN nonlinear regression. Variants represented standard deviations of the fit. Specific activity of the active-site mutants Hydrolysis of tryptamine adenosine, guanosine phosphoramidates and the adenylate (50 μM) by active-site His to Ala mutants of hHint1, hHint3-1, and hHint3-2 (500 pmol) was determined at 25 °C in Hepes buffer. The specific activity was expressed in nmol s− 1 nmol− 1. Lysyl-AMP-dependent adenylation of hHint3-1 (Ala) and -2 (Gly) by hLysRS Adenylation reaction was carried out as described previously.6,7 hLysRS (0.9 μM) was incubated with [α-32P]ATP (0.9 μM, 800 Ci/mmol, MP Biomedicals) in buffer A (7 μL, 25 mM Tris–HCl, pH 7.8, 100 mM NaCl, 2 mM MgCl2, 1 mM DTT, 28.5 μM lysine) with or without yeast inorganic PPase (0.1 U) at 23 °C for 1 min followed by addition of proteins (3.6 μM) and incubation for an additional 1 min. The reaction was terminated by the addition of SDS sample buffer (4×, 4 μL, Invitrogen). The reaction mixture was boiled for 10 min and the proteins were separated by SDS-PAGE and electroblotted onto a polyvinylidene difluoride membrane. Labeled proteins were visualized by subjecting dried polyvinylidene difluoride membranes to autoradiography with a storage phosphor screen for 14 h, followed by scanning with a Strom 840 Phosphorimager. Quantitation of the intensity of the 32P signal was carried out with ImageQuant software (GE Healthcare). Size exclusion chromatography The apparent molecular weight of recombinant purified proteins was analyzed by analytical gel filtration chromatography on Superdex 75 or 200 size exclusion columns (GE Healthcare). The proteins were eluted with P500 buffer (0.5 M NaCl, 50 mM potassium phosphate, 1 mM EDTA, pH 7.0, filtered through a 0.02μm filter) as described.4 Retention times were monitored by protein fluorescence (excitation 285 nm, emission 340 nm) with an in-line FP1520 fluorescence detector (Jasco).

987 Molecular mass from static light scattering H145A-hHint3-1 (50 μM, 200 μL) was subjected to the Superdex 200 SEC and eluted at a flow rate of 0.5 mL/min with P500 buffer and monitored by an in-line multiangle light-scattering detector with a power 690-nm argon laser light source (DAWN EOS, Wyatt Technology), a refractive index detector (Altex), and a UV detector (Beckman Gold 166). Data were collected across a range of angles from 14.5° to 163.3°: 14.5°, 25.9°, 34.8°, 42.8°, 51.5°, 60°, 69.3°, 79.9°, 100.3°, 110.7°, 121.2°, 132.2°, 142.5°, 152.5°, and 163.3°. The instrument was calibrated with a molecular standard bovine serum albumin sample (1 mg/mL, 500 μL). Molecular masses were calculated with the ASTRA software package.29 Circular dichroism spectroscopy CD spectra of proteins were obtained at 23 °C with a J710 spectropolarimeter (Jasco). Proteins at concentrations of 7.5 μM in sodium phosphate buffer (10 mM, pH 7.2) were analyzed in a quartz cuvette with path length of 1 mm, and spectra were accumulated and averaged over nine scans. Subtraction of buffer background from the protein spectrum was performed by using Excel program. HPLC-ESI+ MS analysis The molecular mass of purified proteins (12.5 μM, 5 μL) was analyzed by HPLC-ESI+ MS using an Agilent 1100 capillary HPLC ion trap MS system operated in the ESI+ mode. The spectra were obtained by performing full−scan MS within the m/z range of 100–1500. Chromatographic separation was achieved with an Agilent Zorbax 300 SBC3 column (150 mm × 0.5 mm, 5 μm) eluted at a flow rate of 15 μL/min. The mobile phase consisted of 0.05% trifluoroacetic acid in water (solvent A) and 0.05% trifluoroacetic acid in acetonitrile (solvent B). The elution program started at 30% B for 5 min, followed by a linear increase to 80% B in 25 min. Cellular localization of Hints using fluorescence microscopy To overexpress hHint1 and hHint3s as the CCPGCC tetracysteine-tagged proteins in mammalian cells, their PCR products (primer sequence is listed in Supplementary Table 1) were subcloned into the pDONR™221 and then into the Mammalian Lumio™ Gateway® Vector (nLumio-pcDNA6.2) according to manufacturer's instructions (Invitrogen). HeLa cells were transfected with hHint1, hHint3-1, or hHint3-2-nLumio-pcDNA6.2 plasmids using Lipofectamine 2000 (Invitrogen). Stable cell lines were generated using the antibiotic blasticidin (10 μg/mL). Cells were grown on 35-mm black dishes (Bioptech Inc) for live cell image 1 day before fluorescence microscopy. Transfected cells were labeled with FlAsH-EDT2 (2 μM, Invitrogen) in Hank's balanced salt solution (HBSS) containing glucose (1 mg/mL) for 1 h at 23 °C, then washed with BAL (250 μM, 1.5 mL) twice before imaging. To label nuclei, cells were incubated with Hoechst 33342 dye (1 μg/mL, Invitrogen) for 15 min at 23 °C followed by washing with HBSS (2 mL) twice. Fluorescence images were acquired using a Fluoview FV1000 microscope (Olympus). Fluorescein isothiocyanate and 4′,6-diamidino-2-phenylindole lasers

988 were used to observe FlAsH and Hoechst 33342 dye labeled cells, respectively.

Acknowledgements We thank Xiaodan Liu (University of Minnesota) for help with the HPLC-ESI+ MS experiments, Dr. Karin Musier-Forsyth (The Ohio State University) for providing us the expression plasmid for hLysRS, and John Oja and Jerry Sedgewick (Biomedical Image Processing Laboratory, University of Minnesota) for assistance with confocal experiments.

Supplementary Data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/ j.jmb.2007.08.023

References 1. Brenner, C. (2002). Hint, Fhit, and GalT: function, structure, evolution, and mechanism of three branches of the histidine triad superfamily of nucleotide hydrolases and transferases. Biochemistry, 41, 9003–9014. 2. Bieganowski, P., Garrison, P. N., Hodawadekar, S. C., Faye, G., Barnes, L. D. & Brenner, C. (2002). Adenosine monophosphoramidase activity of Hint and Hnt1 supports function of Kin28, Ccl1, and Tfb3. J. Biol. Chem. 277, 10852–10860. 3. Krakowiak, A., Pace, H. C., Blackburn, G. M., Adams, M., Mekhalfia, A., Kaczmarek, R. et al. (2004). Biochemical, crystallographic, and mutagenic characterization of Hint, the AMP-lysine hydrolase, with novel substrates and inhibitors. J. Biol. Chem. 279, 18711–18716. 4. Chou, T.-F., Bieganowski, P., Shilinski, K., Cheng, J., Brenner, C. & Wagner, C. R. (2005). 31P NMR and genetic analysis establish hinT as the only Escherichia coli purine nucleoside phosphoramidase and as essential for growth under high salt conditions. J. Biol. Chem. 280, 15356–15361. 5. Chou, T.-F., Baraniak, J., Kaczmarek, R., Zhou, X., Cheng, J., Ghosh, B. & Wagner, C. R. (2007). Phosphoramidate pronucleotides: a comparison of the phosphoramidase substrate specificity of human and E. coli histidine triad nucleotide binding proteins (Hint1). Mol. Pharm. 4, 208–217. 6. Chou, T.-F. & Wagner, C. R. (2007). Lysyl-tRNA synthetase-generated lysyl-adenylate is a substrate for histidine triad nucleotide binding proteins. J. Biol. Chem. 282, 4719–4727. 7. Chou, T.-F., Tikh, I. B., Horta, B. A., Ghosh, B., de Alencastro, R. B. & Wagner, C. R. (2007). Engineered monomeric human histidine triad nucleotide binding protein 1 hydrolyzes fluorogenic acyl-adenylate and lysyl-tRNA synthetase-generated lysyl-adenylate. J. Biol. Chem. 282, 15137–15147. 8. Razin, E., Zhang, Z. C., Nechushtan, H., Frenkel, S., Lee, Y. N., Arudchandran, R. & Rivera, J. (1999). Suppression of microphthalmia transcriptional activity by its association with protein kinase C-interacting protein 1 in mast cells. J. Biol. Chem. 274, 34272–34276.

hHint3 is a Branch of the HIT Superfamily 9. Lee, Y. N. & Razin, E. (2005). Nonconventional involvement of LysRS in the molecular mechanism of USF2 transcriptional activity in Fc{varepsilon}RIactivated mast cells. Mol. Cell. Biol. 25, 8904–8912. 10. Yannay-Cohen, N. & Razin, E. (2006). Translation and transcription: the dual functionality of LysRS in mast cells. Mol. Cells, 22, 127–132. 11. Lee, Y. N., Nechushtan, H., Figov, N. & Razin, E. (2004). The function of lysyl-tRNA synthetase and Ap4A as signaling regulators of MITF activity in FcepsilonRIactivated mast cells. Immunity, 20, 145–151. 12. Li, H., Zhang, Y., Su, T., Santella, R. M. & Weinstein, I. B. (2006). Hint1 is a haplo-insufficient tumor suppressor in mice. Oncogene, 25, 713–721. 13. Yuan, B. Z., Jefferson, A. M., Popescu, N. C. & Reynolds, S. H. (2004). Aberrant gene expression in human non small cell lung carcinoma cells exposed to demethylating agent 5-aza-2′-deoxycytidine. Neoplasia, 6, 412–429. 14. Weiske, J. & Huber, O. (2006). The histidine triad protein Hint1 triggers apoptosis independent of its enzymatic activity. J. Biol. Chem. 281, 27356–27366. 15. Wang, L., Zhang, Y., Li, H., Xu, Z., Santella, R. M. & Weinstein, I. B. (2007). Hint1 inhibits growth and activator protein-1 activity in human colon cancer cells. Cancer Res. 67, 4700–4708. 16. Martin, J., Magnino, F., Schmidt, K., Piguet, A.-C., Lee, J.-S., Semela, D. et al. (2006). Hint2, a mitochondrial apoptotic sensitizer downregulated in hepatocellular carcinoma. Gastroenterology, 130, 2179–2188. 17. Korsisaari, N., Rossi, D. J., Luukko, K., Huebner, K., Henkemeyer, M. & Makela, T. P. (2003). The histidine triad protein Hint is not required for murine development or Cdk7 function. Mol. Cell. Biol. 23, 3929–3935. 18. Seidle, H. F., Bieganowski, P. & Brenner, C. (2005). Disease-associated mutations inactivate AMP-lysine hydrolase activity of Aprataxin. J. Biol. Chem. 280, 20927–20931. 19. Kijas, A. W., Harris, J. L., Harris, J. M. & Lavin, M. F. (2006). Aprataxin forms a discrete branch in the HIT superfamily of proteins with both DNA/RNA binding and nucleotide hydrolase activities. J. Biol. Chem. 81, 13939–13948. 20. Ahel, I., Rass, U., El-Khamisy, S. F., Katyal, S., Clements, P. M., McKinnon, P. J. et al. (2006). The neurodegenerative disease protein aprataxin resolves abortive DNA ligation intermediates. Nature, 443, 713–716. 21. Kern, D., Lorber, B., Boulanger, Y. & Giege, R. (1985). A peculiar property of aspartyl-tRNA synthetase from Bakers' yeast: chemical modification of the protein by the enzymatically synthesized aminoacyl adenylate. Biochemistry, 24, 1321–1332. 22. Jakubowski, H. (1999). Misacylation of tRNALys with noncognate amino acids by lysyl-tRNA synthetase. Biochemistry, 38, 8088–8093. 23. Khvorova, A., Motorin, Y. & Wolfson, A. D. (1999). Pyrophosphate mediates the effect of certain tRNA mutations on aminoacylation of yeast tRNA(Phe). Nucleic Acids Res. 27, 4451–4456. 24. Wolfson, A. D. & Uhlenbeck, O. C. (2002). Modulation of tRNAAla identity by inorganic pyrophosphatase. Proc. Natl Acad. Sci. USA, 99, 5965–5970. 25. Adams, S. R., Campbell, R. E., Gross, L. A., Martin, B. R., Walkup, G. K., Yao, Y. et al. (2002). New biarsenical ligands and tetracysteine motifs for protein labeling in vitro and in vivo: synthesis and biological applications. J. Am. Chem. Soc. 124, 6063–6076. 26. Machleidt, T., Robers, M. & Hanson, G. T. (2007).

hHint3 is a Branch of the HIT Superfamily Protein labeling with FlAsH and ReAsH. Methods Mol. Biol. 356, 209–220. 27. Sticha, K. R., Sieg, C. A., Bergstrom, C. P., Hanna, P. E. & Wagner, C. R. (1997). Overexpression and largescale purification of recombinant hamster polymorphic arylamine N-acetyltransferase as a dihydrofolate reductase fusion protein. Protein Expression Purif. 10, 141–153.

989 28. Klein, M. G., Yao, Y., Slosberg, E. D., Lima, C. D., Doki, Y. & Weinstein, I. B. (1998). Characterization of PKCI and comparative studies with FHIT, related members of the HIT protein family. Exp. Cell Res. 244, 26–32. 29. Wyatt, J. P. (1993). Light scattering and the absolute characterization of macromolecules. Anal. Chim. Acta, 272, 1–40.