- © 2019 American Society of Plant Biologists. All rights reserved.
Abstract
Plants can fully catabolize purine nucleotides. A firmly established central intermediate is the purine base xanthine. In the current widely accepted model of plant purine nucleotide catabolism, xanthine can be generated in various ways involving either inosine and hypoxanthine or guanosine and xanthosine as intermediates. In a comprehensive mutant analysis involving single and multiple mutants of urate oxidase, xanthine dehydrogenase, nucleoside hydrolases, guanosine deaminase, and hypoxanthine guanine phosphoribosyltransferase, we demonstrate that purine nucleotide catabolism in Arabidopsis (Arabidopsis thaliana) mainly generates xanthosine, but not inosine and hypoxanthine, and that xanthosine is derived from guanosine deamination and a second source, likely xanthosine monophosphate dephosphorylation. Nucleoside hydrolase 1 (NSH1) is known to be essential for xanthosine hydrolysis, but the in vivo function of a second cytosolic nucleoside hydrolase, NSH2, is unclear. We demonstrate that NSH1 activates NSH2 in vitro and in vivo, forming a complex with almost two orders of magnitude higher catalytic efficiency for xanthosine hydrolysis than observed for NSH1 alone. Remarkably, an inactive NSH1 point mutant can activate NSH2 in vivo, fully preventing purine nucleoside accumulation in nsh1 background. Our data lead to an altered model of purine nucleotide catabolism that includes an NSH heterocomplex as a central component.
INTRODUCTION
Plants can synthesize purine and pyrimidine nucleotides and use salvage reactions to recycle nucleosides and nucleobases into the nucleotide pool (Stasolla et al., 2003; Zrenner et al., 2006). Unlike most animals, plants are also capable of full nucleotide catabolism, releasing the ring nitrogen (Zrenner et al., 2009; Werner and Witte, 2011) for reassimilation into amino acids. In the current model of purine nucleotide catabolism, a branched network of reactions beginning at 5′AMP and 5′GMP leads to the generation of the purine base xanthine as first common intermediate of all branches (Figure 1; see, e.g., Zrenner et al., 2006; Kopecná et al., 2013; Ashihara et al., 2018). The subsequent oxidation and hydrolysis of the xanthine ring via uric acid into glyoxylate, CO2, and NH4+ occur in a linear sequence of reactions (Werner and Witte, 2011), as fully described in the last decade (Todd and Polacco, 2006; Werner et al., 2008, 2010, 2013; Lamberto et al., 2010; Pessoa et al., 2010; Serventi et al., 2010). If this pathway is interrupted by mutation of urate oxidase (UOX) in Arabidopsis (Arabidopsis thaliana), toxic amounts of uric acid accumulate and interfere with peroxisome maintenance in the embryo and cause severe defects in germination and seedling establishment (Hauck et al., 2014).
Current Model of Purine Nucleotide Catabolism in Arabidopsis.
The figure was redrawn from Yin et al. (2014) and slightly simplified. Similar models of the pathway can be found in Riegler et al. (2011) or in reviews (Stasolla et al., 2003; Zrenner et al., 2006). Metabolites: IMP, inosine monophosphate; XMP, xanthosine monophosphate. Enzymes: AMPD, AMP deaminase; IMPDH, IMP dehydrogenase; GMPS, GMP synthetase; HGPRT, hypoxanthine guanine phosphoribosyltrnasferase; IMPP, IMP phosphatase; XMPP, XMP phosphatase; GMPP, GMP phosphatase (IMPP, XMPP, and GMPP are usually summarized as 5′-nucleotide phosphatase not specifying whether the enzyme[s] are nucleotide specific); IGK, inosine guanosine kinase; GSDA, guanosine deaminase; NSH, nucleoside hydrolase; XDH, xanthine dehydrogenase; UOX, urate oxidase (uricase). Enzyme names are shown in blue if the corresponding enzyme is presumed to be involved but the genetic identity is unclear. The gray box encloses metabolites that can only be catabolized, but not salvaged. Note that the conversion of nucleotides and nucleosides can also be catalyzed by phosphotransferases (data not shown) transferring phosphate from a donor mononucleotide onto an acceptor nucleoside. However, such reactions will not create changes in the total nucleoside/mononucleotide pool sizes; therefore, no net salvage or degradation will occur. Oxo- and amino-substituents on the purine ring are highlighted in red and blue, respectively.
The model of a branched reaction network leading to xanthine from the mononucleotides has been derived from a multitude of studies (Schubert and Boland, 1990), mostly using radiolabeled metabolites added to plant extracts (Atkins, 1981; Shelp and Atkins, 1983), plant tissue preparations, or cell cultures (Ashihara et al., 1997; Katahira and Ashihara, 2006; Deng and Ashihara, 2010; Yin et al., 2014) and tracing the metabolic distribution of the label. Plant cells readily take up nucleobases and nucleosides, but not nucleotides, limiting the choice for radiotracer studies with intact tissues or cells to the nonphosphorylated nucleotide metabolism intermediates. To assess whether the enzymes for the postulated metabolic conversions occur in vivo, enzymatic activities in cell-free preparations were sometimes determined as well (Triplett et al., 1980; Shelp and Atkins, 1983; Atkins et al., 1989; Katahira and Ashihara, 2006).
In recent years, genes encoding some of these enzymes have been identified. Mutants of nucleoside hydrolase 1 (NSH1; formerly uridine hydrolase 1, URH1) of Arabidopsis were shown to accumulate high amounts of xanthosine as well as uridine (Jung et al., 2009; Riegler et al., 2011), demonstrating that this enzyme has a role in purine and pyrimidine nucleoside catabolism. After 5 d of darkness, inosine concentrations were also slightly elevated in the NSH1 mutant compared to the the wild-type control (Jung et al., 2011), consistent with the current model showing either inosine or xanthosine as alternative intermediates of AMP and GMP catabolism (Figure 1). An Arabidopsis mutant of guanosine deaminase (GSDA) described by Dahncke and Witte (2013) accumulated high amounts of guanosine genetically demonstrating that plants catabolize G (guanylates, guanosine, guanine) on the level of guanosine as previously postulated based on radiotracer experiments (Ashihara et al., 1997). By contrast, animals and microbes employ guanine deaminases for the deamination of G and its catabolism (Fernández et al., 2009). In an NSH1 GSDA double mutant, xanthosine could no longer be detected, leading to the notion that most, if not all, xanthosine is derived from guanosine and not from xanthosine monophosphate (XMP; Dahncke and Witte, 2013). However, an earlier study using cell-free extracts of cowpea (Vigna unguiculata) root nodules suggested that XMP rather than GMP is a precursor of purine nucleotide breakdown products (Atkins, 1981), which would imply the existence of an XMP phosphatase (XMPP). Plant genes encoding the putative phosphatases and kinases that metabolically link the nucleotides IMP (inosine monophosphate), XMP, and GMP with the nucleosides inosine, xanthosine, and guanosine in vivo have not been discovered. However, it has been firmly established that especially guanosine and to a lesser extent also inosine, but not xanthosine, can be salvaged into nucleotides and nucleic acids when these nucleosides are administered to plant tissue preparations or cell cultures. This suggests that plants possess an inosine guanosine kinase (IGK) but lack a xanthosine kinase. If nucleosides are not salvaged, they are degraded (Katahira and Ashihara, 2006; Deng and Ashihara, 2010; Yin et al., 2014; Ashihara et al., 2018). Nucleobases can be salvaged as well. Phosphoribosyltransferases replace the pyrophosphate in 5-phosphoribosly-1-pyrophosphate with the nucleobase generating mononucleotides. The gene for a plant hypoxanthine guanine phosphoribosyltransferase (HGPRT) has been cloned and the enzyme characterized. HGPRT has a sixfold lower Km value for guanine compared with hypoxanthine (Liu et al., 2007). It has recently been shown that an HGPRT mutant accumulates guanine (Schroeder et al., 2018), genetically demonstrating that HGPRT actually operates as a guanine phosphoribosyltransferase in vivo. Whether HGPRT also salvages hypoxanthine in vivo is unknown. Similar to guanosine and inosine, guanine and hypoxanthine can be salvaged when administered to plant cells or tissues, but the salvage of guanine is more efficient, whereas hypoxanthine is mostly catabolized. By contrast, xanthine is not salvaged at all (Katahira and Ashihara, 2006; Deng and Ashihara, 2010; Yin et al., 2014; Ashihara et al., 2018).
NSH2, a close homolog to NSH1, is conserved in plants (Kopecná et al., 2013) and is thought to be involved in cytosolic purine nucleoside catabolism. In tissue extracts of Arabidopsis NSH2 mutants, xanthosine and inosine, but not uridine, hydrolytic activity was reduced (Riegler et al., 2011). However, neither xanthosine nor inosine accumulated in plants lacking NSH2 (Jung et al., 2011; Riegler et al., 2011). When NSH2 was strongly overexpressed (>60-fold compared to the the wild type), some, but not all lines showed a two- to threefold increased inosine hydrolase activity in the extract (Jung et al., 2011). Biochemical analysis of NSH2 from vascular plants was hampered by the marked insolubility of the recombinant protein (Jung et al., 2011; Riegler et al., 2011; Kopecná et al., 2013), but an NSH2 homolog from the moss Physcomitrella patens could be obtained as soluble dimeric protein showing strong inosine and xanthosine hydrolase activity and weak activity with adenosine and guanosine (Kopecná et al., 2013). Interestingly, the recombinant NSH1 homolog of Physcomitrella is highly insoluble, whereas NSH1 from vascular plants is a soluble protein readily amenable to biochemical analysis. In summary, these data indicate that NSH2 of vascular plants might be involved in purine nucleoside catabolism, but its role in vivo remains obscure, and it cannot be positioned correctly in the current model (Figure 1).
In this work, a range of single, double, and triple null mutants of genes encoding enzymes involved in purine nucleotide catabolism and salvage of Arabidopsis were generated, and alterations in phenotypes and corresponding metabolite profiles were recorded to elucidate the catabolic pathway of purine nucleotides used in vivo. The integration of these data with the current knowledge led to a modified, more linear model of plant purine nucleotide catabolism. Additionally, we obtained genetic as well as biochemical evidence that NSH2 is an intrinsic component of purine nucleotide catabolism. NSH2 absolutely requires interaction with NSH1 for activation. Our data demonstrate that in vivo purine nucleoside hydrolysis is catalyzed by an NSH1-NSH2 complex and not by NSH1 alone.
RESULTS
Genetic Suppression of the UOX Mutant
A UOX mutant of Arabidopsis accumulates uric acid in all tissues. The particularly high concentration in embryos compromises peroxisome maintenance in this tissue, with drastic consequences for germination and seedling establishment. Many seeds do not germinate, and those that do are usually unable to establish a seedling, unless Suc is supplied from outside (Hauck et al., 2014). This phenotype can be suppressed by crossing the UOX mutant with a mutant of xanthine dehydrogenase (XDH). Because XDH lies metabolically upstream of UOX (Figure 1), the double mutant accumulates xanthine instead of uric acid, which has no obvious deleterious effects to the plant. Consequently, it must be the uric acid and not the lack of UOX or the defective purine catabolism per se that causes the strong phenotypes in uox (Hauck et al., 2014).
To elucidate whether a gene is involved in purine catabolism upstream of UOX, a mutant allele of the gene in question can be crossed into the uox background, assessing the potential suppression of the uox phenotypes and of uric acid accumulation in the double mutant. We have generated double mutants of uox with xdh, hgprt, nsh1, and nsh2 and two alleles of gsda, gsda-1 and gsda-2, as well as the triple mutant uox nsh1 nsh2. All of these mutant alleles have been previously described and represent functional knockouts of the respective genes. Neither the uox phenotypes in germination (Figure 2A; Supplemental Figure 1A) and seedling establishment (Figure 2B; Supplemental Figure 1B) nor the uric acid accumulation in seeds (Figure 2C) was altered in uox hgprt background. In human, where purine catabolism naturally ends with uric acid as the final product, a defect of HGPRT causes uric acid hyperaccumulation, leading to severe phenotypic consequences called the Lesch–Nyhan syndrome. Here, defective guanine and/or hypoxanthine salvage boosts purine nucleobase catabolism. This is apparently not the case in Arabidopsis seeds, because defective HGPRT in the uox background does not lead to further increased uric acid accumulation. The data indicate that neither guanine nor hypoxanthine is an intermediate of purine nucleotide catabolism in Arabidopsis, which has been suggested already for guanine (Dahncke and Witte, 2013), but not for hypoxanthine.
Genetic Suppression of the UOX Mutant Phenotypes.
Several genotypes were analyzed: the wild type (Col-0, white bars) and single mutants (blue bars) of UOX, XDH, HGPRT, GSDA (GSDA-1, GSDA-2), NSH2, and NSH1 as well as double mutants (light orange bars) and triple mutants (dark orange bars) in the uox background. Individual data points (dots) of biological replicates and the mean (bar) are shown. Biological replicates were generated from seeds of different individuals grown side by side. Error bars are sd. The statistical analyses were performed using one-way ANOVA with Tukey’s post test. Different letters indicate significant differences (P < 0.05). nd, not detectable. FW, fresh weight.
(A) Germination rate at 3 and 10 d after imbibition (dai; n = 4 biological replicates).
(B) Seedling establishment rate. Percentage of seedlings established 15 d after sowing (n = 4 biological replicates).
(C) Uric acid content in seeds (n = 5 biological replicates).
The strongest suppression of the uox phenotypes and the uric acid accumulation was observed in the uox xdh genotype, corroborating results from Hauck et al. (2014). The data confirm that a second XDH-like gene found in the Arabidopsis genome (XDH2, At4g34900) is not relevant for the generation of uric acid and that it therefore does not contribute to purine nucleotide catabolism in this plant.
Suppression in the uox nsh1 background was strong as well, abolishing almost all uric acid accumulation in seeds and thereby demonstrating that NSH1 has a central function for cytosolic purine nucleoside hydrolysis. Interestingly, also in the uox nsh2 line a slightly diminished uric acid accumulation was observed in seeds, and the uox phenotypes seemed less severe, indicating that also NSH2 is involved in purine nucleoside catabolism. However, an additive suppressive effect of NSH1 and NSH2 abrogation in the uox background was not observed.
Both uox gsda double-mutant lines showed strong suppression of the uox phenotypes. Note that mutation of GSDA alone was previously found to delay seed germination (Schroeder et al., 2018), as could be confirmed here (Figure 2A). Uric acid accumulation was strongly, but not completely, prevented in seeds of the uox gsda double-mutant lines, indicating that purine nucleotide catabolism depends to a great extent, yet not fully, on the GSDA reaction. The GSDA-independent flow through the pathway might pass through a putative XMPP or through the IMP-inosine-hypoxanthine branch according to the current model (Figure 1).
Purine Nucleoside and Nucleobase Accumulation in Seeds and Seedlings of Mutants in Purine Catabolism
Similar to uric acid accumulation in seeds of the uox background, xanthine accumulation in seeds as well as in seedlings of the xdh background is (1) not altered by mutation of HGPRT (xdh hgprt line); (2) strongly suppressed in an XDH NSH1 mutant; (3) weakly but significantly suppressed in an xdh nsh2 line; and (4) strongly, but not completely, suppressed in an XDH GSDA mutant (Figures 3A and 3B, top left). This independent data set confirms that HGPRT does not seem to have a marked influence on purine catabolism, whereas NSH1 and GSDA play major roles and NSH2 is also important. Again, the results show that the GSDA reaction is partially bypassed either through an XMPP or through an IMPP giving access to the inosine-hypoxanthine branch (Figure 1). The data in Figures 2C and 3 indicate that the flux through the bypass is quantitatively smaller than the flux through GSDA.
Purine Nucleobase and Nucleoside Content in Seeds and Seedlings of the Wild Type, Several Mutants of Genes Involved in Purine Nucleotide Catabolism and Salvage, and Double as Well as Triple Mutants in the xdh Background.
Individual data points (dots) of biological replicates and the mean (bar) are shown. Three biological replicates generated from seeds of different individuals grown side by side were analyzed per genotype and tissue. Error bars are sd. The statistical analyses were performed using one-way ANOVA with Tukey’s post test. Different letters indicate significant differences (P < 0.05). nd, not detectable. FW, fresh weight.
(A) Seeds.
(B) Ten-day-old seedlings.
How is GSDA bypassed? If there was a significant flux through the inosine-hypoxanthine branch, one would expect accumulation of inosine and hypoxanthine in mutants that block this branch. Interestingly, the XDH mutant does not accumulate hypoxanthine (Figure 3), which has also been noted by Brychkova et al. (2008), although xanthine and hypoxanthine are both substrates of XDH (Figure 1). In accordance, blocking XDH activity chemically with allopurinol in soybean (Glycine max) and tobacco (Nicotiana tabacum) resulted only in xanthine, but not in hypoxanthine, accumulation (Fujihara and Yamaguchi, 1978; Boland and Schubert, 1982; Montalbini and DellaTorre, 1995). However, in cowpea nodules, 5% to 10% hypoxanthine with reference to xanthine was detected upon allopurinol treatment (Atkins et al., 1988). Although the general lack of hypoxanthine accumulation in plant tissues without active XDH suggests that hypoxanthine is not an intermediate of purine nucleotide catabolism, hypoxanthine could potentially be salvaged by HGPRT, preventing its accumulation, whereas xanthine is not salvaged (Ashihara et al., 2018). In the xdh hgprt background, where salvage is also blocked, we indeed observed some hypoxanthine accumulation in seeds and seedlings, but the hypoxanthine concentration was always at least one order of magnitude smaller than the xanthine concentration (Figure 3). In seeds, hypoxanthine even accumulated in an xdh gsda background, where high guanosine and in consequence guanine levels were reached. Guanine is the preferred substrate of HGPRT (Liu et al., 2007), probably limiting the enzyme’s availability for hypoxanthine and leading to its accumulation. As shown above (Figures 2 and 3), HGPRT abrogation in the uox or xdh background does not boost the concentration of the purine nucleotide catabolites uric acid or xanthine, in contrast to what is observed in human. These results together with the comparatively small accumulation of hypoxanthine in the xdh hgprt background suggest that hypoxanthine is not a major intermediate of purine nucleotide catabolism in Arabidopsis.
Similarly, the amount of inosine accumulating in the NSH1 mutant was >400 times lower than the amount of xanthosine in seeds and seedlings (Figure 3). However, as hypoxanthine, inosine can be salvaged to some extent according to radiotracer experiments (Ashihara et al., 2018), presumably by a dual-specific IGK (Combes et al., 1989) which has not yet been identified genetically. Salvage might reduce inosine accumulation in the nsh1 background, whereas xanthosine cannot be salvaged (Ashihara et al., 2018). Our data indicate that a dual-specific IGK is also present in Arabidopsis, because upon strong guanosine accumulation in the gsda background of seeds (gsda and xdh gsda lines; Figure 3A) an elevated inosine concentration is consistently observed (Figure 3A), possibly because IGK is fully occupied with guanosine in this situation and therefore not available for inosine. We examined whether the inosine concentration would strongly increase in the nsh1 background, where degradation is blocked, if salvage is also compromised by high guanosine amounts — a situation occurring in nsh1 gsda seeds. However, we observed only a slightly higher inosine concentration in nsh1 gsda seeds versus nsh1 seeds (Figure 4A), indicating that inosine salvage does not play an important role. Therefore, even when salvage is likely compromised, the inosine concentration in seeds remains approximately two orders of magnitude smaller than the concentration of xanthosine in the nsh1 background (Figure 4A). This is also the case in seedlings and the rosette of 4-week-old plants (Figures 4B and 4C). It appears that inosine, similar to hypoxanthine, is not a major intermediate of purine nucleotide catabolism in vivo. Hence, the flux through AMP/GMP catabolism that bypasses GSDA is not routed through an IMPP to access the inosine-hypoxanthine branch but runs probably through an XMPP which generates xanthosine. This concept is strongly supported by the observation that xanthosine accumulation in the nsh1 background is not completely suppressed in seeds and seedlings when GSDA is additionally abrogated (Figures 4A and 4B). The xanthosine remaining in the nsh1 gsda background is very likely generated by the XMPP. Consistently, uric acid and xanthine accumulation in seeds and seedlings of uox and xdh backgrounds, respectively, is highly, but never fully, suppressed by GSDA abrogation (Figures 2 and 3). In rosette leaves of 4-week-old plants, a stronger suppression of xanthosine accumulation in the nsh1 gsda background is observed than in seeds and seedlings (Figure 4C), indicating that here GSDA dominates xanthosine generation with hardly any contribution from XMPP. This is mirrored by a stronger suppression of xanthine accumulation in xdh gsda double mutants in that tissue (Supplemental Figure 2).
Nucleoside Content in Seeds, Seedlings, and Rosettes of NSH1 GSDA Double Mutants and the Respective Single Mutants as Well as the Wild Type.
Individual data points (dots) of biological replicates and the mean (bar) are shown. Four biological replicates generated from seeds of different individuals grown side by side were analyzed per genotype and tissue. Error bars are sd. The statistical analyses were performed using one-way ANOVA with Tukey’s post test. Different letters indicate significant differences (P < 0.05). nd, not detectable. FW, fresh weight.
(A) Seeds.
(B) Seven-day-old seedlings.
(C) Rosettes of 4-week-old plants just before bolting.
Taken together, our data suggest that AMP and GMP degradation in Arabidopsis are both routed through xanthosine and not inosine in vivo. It is likely that inosine and hypoxanthine are derived from other sources. Inosine occurs in transfer RNA (tRNA), and there are six nuclear tRNAs containing inosine in Arabidopsis (Zhou et al., 2014). Upon vacuolar degradation of tRNA, inosine will be released into the cytosol, probably via the equilibrative nucleoside exporter 1 (Bernard et al., 2011). Hypoxanthine accumulates to ∼10-fold higher concentrations than inosine in seeds and seedlings in the respective mutants (Figure 3), indicating that it is not only derived from the hydrolysis of inosine. Hypoxanthine may stem from base excision repair of DNA. It occurs in DNA due to erroneous incorporation of deoxyinosine triphosphate by DNA polymerases and because adenine spontaneously deaminates to hypoxanthine at a low rate (Alseth et al., 2014). Similarly, the usual origin of guanine in plant metabolism might be spontaneous depurination of nucleotides and nucleic acids (An et al., 2014). Hence, the physiological role of HGPRT is to salvage hypoxanthine and guanine from these sources, whereas purine nucleotide catabolism generates xanthine without a hypoxanthine or guanine intermediate and is thereby decoupled from HGPRT activity in Arabidopsis. By contrast, such decoupling is not observed in human, where HGPRT counteracts excessive purine base catabolism, preventing uric acid accumulation and the development of the Lesch–Nyhan syndrome.
The Role of NSH2
There are several indications for a function of NSH2 in purine nucleotide catabolism from our mutant analyses. (1) The accumulation of uric acid in seeds and the resulting macroscopic phenotypes during germination and seedling establishment of the UOX mutant are partially suppressed in a UOX NSH2 double mutant (Figure 2; Supplemental Figure 1). (2) An XDH NSH2 mutant also accumulates less xanthine than the XDH mutant in seeds (Figure 3A), a tendency that is also observed in seedlings (Figure 3B). (3) The xanthosine concentration appears elevated in NSH2 mutant seeds without reaching statistical significance (Figure 3A), but (4) strikingly, xanthosine accumulates strongly in seeds of the XDH NSH2 double mutant (Figure 3A). However, it is known that NSH1, but not NSH2, is essential to prevent high accumulation of xanthosine in vivo (Figure 3; Jung et al., 2011; Riegler et al., 2011). It is possible that the high concentration of xanthine in the xdh background causes product inhibition of NSH1, reducing the efficiency of xanthosine catabolism, which only becomes apparent if additionally, NSH2 is absent. Is NSH2 an activator of NSH1? Or is NSH2 itself a xanthosine hydrolase that is activated by NSH1? In both scenarios, NSH1 and NSH2 might physically interact.
NSH2 Is a Xanthosine and Inosine Hydrolase That Strictly Requires NSH1 for Activity
A detailed enzymatic characterization of purified NSH2 from vascular plants has been hampered mainly by the low solubility of the enzyme when it was expressed in Escherichia coli (Jung et al., 2011; Riegler et al., 2011; Kopecná et al., 2013). Here, we transiently expressed N-terminally Strep-tagged variants of NSH1 and NSH2 in leaves of Nicotiana benthamiana and were able to obtain highly pure and soluble enzymes after affinity purification (Figure 5). The NSH1 preparation was active with xanthosine and uridine, confirming previous findings (Jung et al., 2009; Riegler et al., 2011). By contrast, purified NSH2 showed no hydrolytic activity with these substrates (Table 1).
Purification of Nucleoside Hydrolases with Strep Affinity Chromatography.
(A) Purification of N-terminally Strep-tagged NSH1 and NSH2 transiently expressed in N. benthamiana leaves (top and middle panels, respectively). To boost expression of NSH2, an additional construct containing viral enhancer sequences (see “Methods”) was used (bottom panel). Proteins were detected by immunoblot with StrepTactin-alkaline phosphatase conjugate. CL, clarified cell lysate; NB, protein not bound to the affinity matrix; W, protein in the last washing step; E, elution fraction; B, protein remaining bound to the affinity matrix after elution.
(B) Solubility of transiently expressed NSH1 and NSH2. Clarified cell lysates of N. benthamiana leaves expressing either NSH1 or NSH2 were centrifuged at 100,000g for 1 h. The supernatant and the pellet fractions were analyzed on an immunoblot using StrepTactin-alkaline phosphatase detection.
(C) Purity of affinity (co)purified nucleoside hydrolases. SDS gel loaded with 12.5 μL of the affinity-purified enzymes (elution fraction) that either had been expressed alone as Strep-tagged variants or together with myc-tagged variants. ΔNSH1 and ΔNSH2 are inactive point mutants of NSH1 and NSH2. The gel was stained with colloidal Coomassie blue.
The results of our mutant analyses indicated that NSH1 and NSH2 might physically interact. This notion was experimentally tested by coexpressing N-terminally Strep- and myc-tagged variants of NSH1 and NSH2 in N. benthamiana leaves and assessing whether the myc-tagged variants could be copurified upon affinity purification of the respective Strep-tagged enzymes. We found that NSH1 and NSH2 indeed interact, because myc-tagged NSH1 can be copurified with Strep-tagged NSH2 and vice versa (Figure 6). Interestingly, NSH1 also interacts with itself, whereas NSH2 does not. The lack of activity of NSH2 and its inability to form homomers is surprising, because the NSH2 homolog from the moss P. patens was shown to have xanthosine and inosine hydrolase activity and to form dimers (Kopecná et al., 2013).
Interaction of NSH1 and NSH2 In Planta.
NSH1 and NSH2 were transiently coexpressed as Strep- or myc-tagged variants in N. benthamiana leaves in the combinations indicated. Affinity purification was performed using the Strep tag. Protein expression (top) and protein (co)purification by affinity chromatography (bottom) were assessed on immunoblots developed with either StrepTactin-AP conjugate or anti-myc antibodies. Per lane, 12 μL of clarified leaf extracts or affinity-purified proteins were loaded. AP, affinity purification; suffix -m, myc-tagged; suffix -s, Strep-tagged.
Specific polyclonal antibodies against NSH1 and NSH2 were developed to assess by immunoprecipitation (IP) experiments whether the untagged native proteins interact in Arabidopsis. Unfortunately, we were unable to obtain an antibody for NSH2 that did not also cross-react with NSH1, even after multiple rounds of affinity purification of the antibody preparation. However, a difference in electrophoretic mobility of both enzymes allowed them to be distinguished on immunoblots. The dual-specific anti-NSH1/2 antibody was used to detect both enzymes after the IP performed with the mono-specific anti-NSH1 antibody. The IP experiments using root extracts of hydroponically grown wild-type, nsh1, and nsh2 plants demonstrated that native NSH1 and NSH2 interact with each other in Arabidopsis (Figure 7).
Interaction of Native NSH1 and NSH2 in Roots of Arabidopsis.
Root cell lysates of Col-0, nsh1, and nsh2 were IPed with anti-NSH1 antibody. The presence of the nucleoside hydrolases in the clarified lysate (before IP) and after the IP was detected by immunoblot with anti-NSH1 (top) and dual-specific anti-NSH1/2 antibodies (bottom). The last lane in both panels is an IP control without added root extracts.
The expression of NSH1 and NSH2 was assessed with immunoblots using the anti-NSH1 antibody as well as the dual-specific NSH1/2 antibodies in seedlings, roots, and the rosettes of plants before bolting, and in different tissues of flowering plants (Figure 8). The blots were loaded with extracts obtained from an equal fresh weight of tissue. With both antibodies, NSH1 was detected in all tissues, showing that NSH1 is a ubiquitously expressed enzyme. By contrast, NSH2 could only be detected clearly in roots, seedlings, and flowers, although the dual-specific antibody is as sensitive for NSH2 as for NSH1 (Supplemental Figure 3).
Protein Abundance of NSH1 and NSH2 in Different Tissues at Distinct Developmental Stages.
(A) Immunoblot developed with anti-NSH1 antibody. A 2:1 buffer-to-sample ratio was used for extraction, and 12 μL of sample were loaded per lane. Samples: 10-d-old seedlings grown on agar plates, roots and the rosette of a 4-week-old plant before bolting grown in soil, young leaves (leaves up to position 14), middle leaves (13 to 7), old leaves (6 to 1), cauline leaves, flowers, and siliques of a 6-week-old plant.
(B) As in (A) but developed with the dual-specific NSH1/2 antibody and using the NSH2 mutant additionally to the wild type, as negative control for the NSH2 signal.
In roots, NSH2 often appears only slightly less abundant than NSH1 (Figures 7 and 8B; Supplemental Figure 3). In several other tissues, NSH2 is not sufficiently abundant to be clearly detected in our immunoblots, although public transcriptome data show that NSH2 transcript is ubiquitously present. Crystal structure and size exclusion chromatography data showed that plant NSH proteins form dimers (Kopecná et al., 2013). The higher abundance of NSH1 compared with NSH2 indicates that in vivo a homodimeric NSH1 complex and a heterodimeric NSH1-NSH2 complex will usually be present.
Because of the higher abundance of NSH2 in roots, we investigated whether the abrogation of NSH2 has a marked impact on nucleoside accumulation in that tissue (Figure 9). A small accumulation of xanthosine, but not of inosine or uridine, was observed in the nsh2 background, again confirming the general involvement of NSH2 in xanthosine degradation, but not supporting the idea of an especially prominent role of NSH2 in roots. In the NSH1 mutant, a strong accumulation of xanthosine and uridine and an increase in the inosine concentration were detected, re-emphasizing the central role of NSH1 for the catabolism of these nucleosides.
Nucleoside Concentration in Root Extracts of the Nucleoside Hydrolase Mutants and the Wild Type.
Uridine, inosine, and xanthosine concentrations in root extracts of hydroponically grown Col-0, nsh2, and nsh1 plants. Individual data points (dots) of three biological replicates and the mean (bar) are shown. Error bars are sd. The asterisks indicate significant differences between the mutants and the wild type as determined by unpaired two-tailed t tests (*P < 0.05; ***P < 0.001). FW, fresh weight.
NSH1, NSH2, and the NSH1-NSH2 complex were purified after transient (co)expression in leaves of N. benthamiana, and the specific activities were assessed with 250 µM xanthosine, 250 µM inosine, and 250 µM uridine as substrates (Table 1). NSH1 was most active with uridine, approximately ninefold less active with xanthosine, and showed only very low activity with inosine when this substrate was offered at higher concentrations (Supplemental Table 1). NSH2 did not hydrolyze any of these substrates, irrespective of concentration. Interestingly, the NSH1-NSH2 complex showed a marked activity with xanthosine, much stronger than that of NSH1 alone, and also some activity with inosine and uridine. By determining the kinetic constants, it became clear that the catalytic efficiency (kcat/Km) of the NSH1-NSH2 complex for xanthosine is ∼80-fold higher than the catalytic efficiency of NSH1 alone. This is due to a lower Km and a higher kcat (Table 2). For uridine, the situation is reversed: the catalytic efficiency of the NSH1-NSH2 complex is more than threefold lower than that of NSH1 alone, mainly due to a higher Km of the complex. For inosine, we were unable to determine the kinetic constants of NSH1 with our spectrophotometric method, because the required substrate concentrations would have exceeded the measurable range, but Jung et al. (2009) determined a Km value of 1.4 mM and a kcat value of 8.7 s−1 using enzyme purified from E. coli and a radiometric assay. Hence, with a Km value of 0.6 mM and a kcat value of 42.3 s−1 for inosine, the catalytic efficiency of the NSH1-NSH2 complex is more than an order of magnitude higher than that of NSH1 alone. We conclude that the NSH1-NSH2 complex is a highly efficient xanthosine hydrolase that also has inosine as well as weaker uridine hydrolase activity. The xanthosine/inosine hydrolase activities exceed by far those of the NSH1 homodimer, which is the better uridine hydrolase.
We asked whether a xanthosine/inosine hydrolase activity of NSH2 is activated by the interaction with NSH1, or whether NSH2 is intrinsically inactive, but can boost the xanthosine/inosine hydrolase activity of NSH1. To address this question, active site mutants were generated by changing a catalytically essential Asp to Ala in NSH1 (D29A) and NSH2 (D14A) (Supplemental Figure 4). In the resulting mutants, called ΔNSH1 and ΔNSH2, nucleoside hydrolase activity was abrogated (Table 1; Supplemental Table 1). Interestingly, the ΔNSH1-NSH2 complex was highly active with xanthosine and retained low inosine hydrolase activity, demonstrating that NSH2 is an active xanthosine inosine nucleosidase, which requires the interaction with NSH1 for activation. The NSH1-ΔNSH2 complex also retained some hydrolytic activity for xanthosine and uridine, showing that both enzymes actively contribute to the enzymatic characteristics of the NSH1-NSH2 complex. Xanthosine- and inosine-specific activities are mainly contributed by NSH2, whereas uridine-specific activity appears to reside mainly in NSH1. Nonetheless, NSH1 alone is sufficiently active with xanthosine to prevent strong accumulation of this metabolite in planta (Figures 3, 4, and 9).
To assess whether the tagged variants of NSH1 and NSH2 are functional in vivo, the respective knockout lines were transformed with constructs encoding Strep- and myc-tagged variants of the enzymes. Additionally, the NSH1 mutant was transformed with a construct encoding the Strep-tagged inactive ΔNSH1 (D29A) variant. According to our in vitro data, NSH2 might be activated by ΔNSH1 in this nsh1 background, creating a situation where NSH2 is active but NSH1 is inactive. Using immunoblots, the expression of the proteins in the corresponding transgenic plants was shown (Figure 10A). In the respective lines, xanthosine, inosine, and uridine were quantified in seeds, 7-d-old seedlings, and rosettes of 4-week-old plants just before bolting (Figure 10B). Xanthosine accumulation in the nsh2 background was suppressed by the NSH2 transgenes, and xanthosine, inosine, and uridine accumulation in the nsh1 background was suppressed by the NSH1 transgenes, demonstrating that the myc- and Strep-tagged variants of these enzymes are functional in vivo. Interestingly, xanthosine and inosine accumulation in the nsh1 background was also completely suppressed by the ΔNSH1 transgene. Apparently, the activation of native NSH2 by this inactive ΔNSH1 variant is sufficient for the maintenance of the wild-type xanthosine and inosine homeostasis in seeds, seedlings, and rosettes. This demonstrates that NSH2 is an efficient xanthosine and inosine hydrolase in vivo and shows that NSH2 must be expressed in leaf tissue, although it was difficult to clearly detect it there with our dual-specific antibody (Figure 8B). In seeds, but not in seedlings and rosettes, even the high uridine concentration in the nsh1 background was reduced by activation of NSH2 with ΔNSH1. At such an elevated substrate concentration as found in the nsh1 seed, the weak uridine hydrolase activity of the ΔNSH1-NSH2 complex detected in vitro (Supplemental Table 1) seems to be sufficient to degrade some uridine over time.
Protein Expression and Nucleoside Accumulation in Transgenic Lines Expressing Nucleoside Hydrolase Variants.
(A) Expression of Strep- or myc-tagged NSH1 (NSH1-s, NSH1-m) as well as an enzymatically inactive Strep-tagged NSH1 point mutant (ΔNSH1-s) in the nsh1 background detected in seedlings and roots by immunoblots developed with the anti-NSH1 antibody (top panels). Expression of Strep- or myc-tagged NSH2 (NSH2-s, NSH2-m) in the nsh2 background detected in seedlings and roots by immunoblots developed with the dual-specific anti-NSH1/2 antibody (bottom panels). Seven-day-old seedlings and roots from 4-week-old plants were used.
(B) Xanthosine, inosine, and uridine concentration in extracts from seeds, 7-d-old seedlings, and rosette leaves of 4-week-old plants of the wild type (white), the nucleoside hydrolase single mutants (blue), and the double mutant (orange) and the indicated transgenic lines overexpressing tagged variants of NSH1 and NSH2 (green) as well as the inactive D29A point mutant of NSH1 (ΔNSH1, dark green) in the corresponding mutant backgrounds. Individual data points (dots) of biological replicates (n = 5 for seedlings and rosettes and n = 3 for seeds derived from different mother plants) and the mean (bar) are shown. Error bars are sd. The statistical analyses were performed using one-way ANOVA with Tukey’s post test. Different letters indicate significant differences (P < 0.05). nd, not detectable. FW, fresh weight.
Dark Stress–Induced Nucleotide Catabolism
Although NSH2 is a highly efficient xanthosine hydrolase, its abrogation only had a small influence on xanthosine concentrations in seeds, seedlings, rosettes, and roots (Figures 3, 4, 9, and 10). We wondered whether the impact of NSH2 abrogation would increase in a situation where the flux through purine nucleoside catabolism is higher. Therefore, we exposed the 3-week-old wild-type and different NSH mutant plants to dark stress, known to boost nucleotide catabolism (Jung et al., 2011; Schroeder et al., 2018), and xanthosine, inosine, and uridine concentrations were determined after 2 and 3 d of exposure to darkness. In the wild type and in the nsh2 background, the xanthosine concentration did not increase (Figure 11). By contrast, in the nsh1 background, the xanthosine concentration was increased ∼25-fold by the second day and by ∼65-fold by day 3 to 22.8 µmol/g dry weight (DW). It appears that the requirement for NSH2 is not rising with higher flux through nucleoside catabolism. Interestingly, the massive dark-induced xanthosine accumulation in the nsh1 background was prevented by the inactive ΔNSH1 transgene, showing that the activity of native NSH2, which is induced by ΔNSH1, is sufficient to hydrolyze xanthosine even in situations of high metabolic flux.
Alteration of Nucleoside Content in Nucleoside Hydrolase Variants during Dark Stress.
Xanthosine, inosine, and uridine were quantified in rosettes of 3-week-old plants before dark exposure (day 0) and after 2 and 3 d of darkness. Individual data points (dots) of four biological replicates and the mean (bar) are shown. Error bars are sd. The statistical analyses were performed using one-way ANOVA with Tukey’s post test. Different letters indicate significant differences (P < 0.05). nd, not detectable. DW, dry weight.
The inosine concentration was very low throughout the experiment and only rose by day 3 in the nsh1 background to an ∼65-fold lower concentration as found for xanthosine. However, the ∼0.35 µmol/g DW reached at day 3 was the highest tissue concentration of inosine observed in any of our experiments. The lack of any inosine accumulation up to the second day strongly supports the idea that inosine is not an intermediate of AMP/IMP catabolism. The increase in inosine concentration on the third day in the nsh1 background might be caused by the onset of autophagy-mediated vacuolar tRNA turnover by that time.
We repeated the dark stress experiment with the wild type and the xdh, hgprt, and xdh hgprt lines, quantifying xanthine and hypoxanthine (Figure 12A). Despite an already strongly elevated xanthine concentration in xdh plants at the beginning of the experiment, the xanthine concentration was further increased by dark stress, demonstrating that purine catabolism is operating, whereas the hypoxanthine concentration in the xdh hgprt line, which is blocked in catabolism and salvage, stayed constant. Again, these data strongly suggest that AMP/IMP is not catabolized via hypoxanthine.
Changes in Nucleobase Concentrations during Long-Term and Short-Term Dark Stress.
(A) Xanthine and hypoxanthine were quantified in the wild type and the XDH HGPRT double mutant as well as the respective single mutants using rosettes of 3-week-old plants before dark exposure (day 0) and after 2 and 3 d of darkness. Individual data points (dots) of four biological replicates and the mean (bar) are shown. Error bars are sd. The statistical analyses were performed using one-way ANOVA with Tukey’s post test. Different letters indicate significant differences (P < 0.05). nd, not detectable. DW, dry weight.
(B) Concentration of nucleobases in the 7-day-old wild-type and hgprt seedlings at the end of the night (before) and after 8 h in light or in darkness, when plants were optionally treated with allopurinol at the beginning of the 8-h period. Plants from four growth plates (half-strength MS medium) were analyzed per treatment. Individual data points (dots) of the biological replicates (n = 4 biological replicates) and the mean (bar) are shown. Error bars are sd. On each plate both genotypes were grown and treated in parallel. FW, fresh weight.
One may argue that the permanently high xanthine concentration in the xdh background already disturbs purine catabolism. We therefore performed an additional short-term dark exposure experiment with the wild type and the HGPRT mutant, where we used allopurinol, which strongly inhibits XDH, causing an acute lack of XDH activity after treatment. Long-day-grown seedlings were optionally exposed at the end of the night to allopurinol and then either grown in a normal day or exposed to darkness for 8 h to stimulate purine nucleotide catabolism (Figure 12B). Allopurinol-treated plants accumulated xanthine, but this accumulation was approximately three times stronger in the dark (∼55 nmol g−1) than in the light (∼17 nmol g−1), demonstrating that purine nucleotide catabolism was more active under dark conditions. In the hgprt background, the allopurinol-treated plants also accumulated hypoxanthine, but slightly more in the light compared with darkness (∼5 nmol g−1). This is again consistent with the idea that hypoxanthine is not derived from AMP/IMP catabolism. It also confirms that hypoxanthine is usually processed by both XDH and HGPRT, because it only accumulates if both activities are compromised. As expected, the guanine concentration was higher in the HGPRT mutant than in the wild type, but the concentration did not change over time or by treatment in both genotypes, confirming that guanine is not an intermediate of purine nucleotide catabolism in Arabidopsis.
DISCUSSION
When the data presented here are integrated with the current knowledge about purine nucleotide catabolism in vascular plants, a revised model can be drawn (Figure 13; Supplemental Figure 5). Central differences to the current model (Figure 1) are the absence of an IMPP reaction connecting IMP to inosine and hypoxanthine and the placement of NSH2 in complex with NSH1 on the metabolic map as xanthosine and inosine nucleoside hydrolase. Doubts about the existence of an IMPP in vivo have been raised earlier in work with soybean and cowpea, where purine nucleotide catabolism is highly active in nodules to generate the ureides allantoin and allantoate (Supplemental Figure 5) as export forms of fixed nitrogen. Shelp and Atkins (1983) concluded from work on cell-free extracts of cowpea nodules with labeled IMP that the production of ureides rather occurs via IMP oxidation than IMP dephosphorylation. In soybean nodules, only xanthine, but no hypoxanthine, was observed after allopurinol treatments (Fujihara and Yamaguchi, 1978; Boland and Schubert, 1982), which cannot be explained by preferential hypoxanthine salvage in this tissue, since the nodule is a site of strong purine nucleotide catabolism. A consequence of the absence of IMPP in Arabidopsis is that inosine and hypoxanthine are not intermediates of AMP/IMP catabolism and that both AMP and GMP are catabolized via xanthosine. Dahncke and Witte (2013) postulated that xanthosine is exclusively generated by GSDA in Arabidopsis, because xanthosine could not be detected in an NSH1 GSDA double mutant using HPLC coupled with UV detection. Here, we present several lines of evidence that xanthosine is not exclusively produced from guanosine deamination by GSDA but also from XMP dephosphorylation (Figure 4), which leads us to postulate that an as-yet unidentified XMPP must exist. In cell-free extracts of nodules from cowpea, ureide synthesis was supported significantly better by XMP than by GMP, and GMP was not generated from XMP (Atkins, 1981), indicating that an XMPP is active in this tissue. It is tempting to speculate that AMP is degraded to xanthosine via IMP and XMP using a XMPP and that GMP is catabolized by a GMP phosphatase and GSDA. However, our data consistently show that the GSDA mutant is a surprisingly strong suppressor for the accumulation of purine ring catabolites (Figures 2 to –4), suggesting that the route via the XMPP is less important in Arabidopsis. This might be correct, but one needs to bear in mind that the GSDA mutant accumulates high amounts of guanosine, which could potentially inhibit XMPP. In this scenario, a GSDA mutant would not only lack GSDA activity but also be partially compromised in XMPP activity. Because of these possible complications, the quantitative aspects of our data should not be overinterpreted.
Updated Model of Purine Nucleotide Catabolism in Arabidopsis.
Metabolites with orange background are released in cytosolic (AMP, GMP) or vacuolar (adenosine, guanosine) RNA turnover. Blue, red, and yellow arrows indicate salvage, catabolic, and biosynthetic reactions, respectively. Note that reactions leading from AMP to GMP can be biosynthetic or catabolic depending on whether the degradation of AMP or the generation of GMP from AMP is dominating (dotted box). Metabolites within the box demarcated with a light gray line can only be catabolized and not salvaged. An extended version of the model is presented in Supplemental Figure 5. Enzymes: ADK, adenosine kinase; AMPD, AMP deaminase; IMPDH, IMP dehydrogenase; GMPS, GMP synthetase; HGPRT, hypoxanthine guanine phosphoribosyltransferase; XMPP, XMP phosphatase; GMPP, GMP phosphatase; IGK, inosine guanosine kinase; GSDA, guanosine deaminase; NSH1, nucleoside hydrolase 1; NSH2, nucleoside hydrolase 2; XDH, xanthine dehydrogenase; UOX, urate oxidase. Enzyme names are shown in blue if the corresponding enzyme is presumed to be involved but the genetic identity is unclear.
One puzzling aspect of purine nucleotide catabolism according to the current model (Figure 1) is the parallel presence of catabolic and salvage steps in the same compartment and probably the same tissue. Most of the enzymes shown in Figure 1 are located in the cytosol. For the phosphatases, the location is unknown, whereas an IGK has been partially purified from a mitochondrial preparation in Helianthus tuberosus (Combes et al., 1989), indicating that this enzyme might reside in the mitochondria. Why would IMP and GMP first be degraded to hypoxanthine and guanine, respectively, and then salvaged back to IMP and GMP by HGPRT in a futile cycle? With the discovery of GSDA (Dahncke and Witte, 2013), it became clear that guanine is unlikely an important intermediate of G nucleotide catabolism in Arabidopsis. The data presented here suggest that only xanthine, but neither hypoxanthine nor guanine, is an intermediate of AMP and GMP catabolism. Therefore, in the revised model (Figure 13), HGPRT is completely decoupled from AMP and GMP catabolism. The enzyme salvages guanine, probably generated by spontaneous depurination of nucleotides and nucleic acids. It also salvages hypoxanthine, possibly derived from DNA and nucleotide pool maintenance as well as from vacuolar tRNA turnover releasing inosine, which is hydrolyzed by NSH1-NSH2 to hypoxanthine and ribose. Uptake of soilborne nucleosides from the rhizosphere for salvage or plant nitrogen nutrition is probably another source of inosine (Tokuhisa et al., 2010), possibly explaining why NSH enzymes are most strongly expressed in roots (Figure 8; Riegler et al., 2011). In agreement with pulse-chase studies using radiotracers, our data show that hypoxanthine is alternatively salvaged or degraded in vivo (Figures 3 and 12), although it is unclear which reaction prevails in undisturbed metabolism.
Another interesting complication in purine nucleotide metabolism is that the biosynthetic pathway for GMP and the catabolic pathway for AMP share at least two common enzymes: AMP deaminase and IMP dehydrogenase. How biosynthetic versus catabolic fluxes are regulated is currently unclear, but a metabolic switch might reside on the XMPP and GMP synthetase enzymes, which both use XMP as substrate.
The genomes of plants encode two phylogenetically separated clades of cytosolic nucleoside hydrolases, NSH1 and NSH2 (Kopecná et al., 2013). An assignment of a functional role was relatively straightforward for NSH1 in Arabidopsis, because the corresponding mutant showed significant defects in the catabolism of xanthosine and uridine (Jung et al., 2009; Riegler et al., 2011) and inosine (Figure 3). By contrast, the functional role of NSH2 remained unresolved, although based on in vitro experiments, the enzyme appeared to be involved in xanthosine and inosine hydrolysis, and it was hypothesized that NSH2 might form a heterocomplex with NSH1 (Riegler et al., 2011). We show here that an NSH1 mutant is actually a functional knockout of both NSH1 and NSH2, because NSH2 requires NSH1 for activity, but not vice versa. NSH2 was predicted to be a nucleosidase, which preferentially hydrolyzes xanthosine and inosine, because of two characteristic active site Tyr residues (Y233 and Y238) versus W247 and D252 in NSH1, which are indicative of an enzyme that prefers xanthosine and uridine (Kopecná et al., 2013). Nonetheless, the inosine concentration is not altered and the xanthosine concentration is only slightly elevated in the nsh2 background, which could only be detected with our sensitive mass spectrometry method, but was not noted previously using less sensitive methods. Is NSH1 the more important enzyme and NSH2 only plays a minor role? One needs to look at a situation where NSH2 is the only active enzyme in vivo; this occurs in the NSH1 mutant complemented with the inactive ΔNSH1 variant (Figures 10 and 11). Here, NSH2 alone can suppress xanthosine and inosine (but not uridine) accumulation normally observed in the nsh1 background to the wild-type levels, for xanthosine consistently below the concentration found when only NSH1 is active (Figure 10B, nsh2 background). Note that NSH2 is not overexpressed in this line and that its activation by the mutated ΔNSH1 might not be complete as indicated by our in vitro activity measurements (Table 1). The partial suppression of uric acid and xanthine accumulation in the uox and xdh background by NSH2 mutation (Figures 2 and 3) demonstrates that flux through purine nucleotide catabolism is reduced in the absence of NSH2. We conclude that for xanthosine as well as inosine hydrolysis, NSH2 is at least as effective as NSH1 and that NSH2 is required for complete xanthosine homeostasis and undisturbed flux through purine nucleotide catabolism. It is possible that by modulation of NSH2 amounts, the hydrolytic capacity for distinct nucleosides could be steered, because the NSH1-NSH1 and NSH1-NSH2 complexes have clearly distinct enzymatic characteristics. It remains to be elucidated whether such a regulation actually occurs and under which conditions a regulation of hydrolytic capacity for distinct nucleosides might be required.
METHODS
Cloning and Site-Directed Mutagenesis
The complementary DNAs (cDNAs) of NSH1 and NSH2 were amplified with the primer pairs 1905 + 1906 and 1904 + 1903, respectively (Supplemental Table 2), introducing NcoI and BamHI restriction sites. For protein expression in Nicotiana benthamiana with N-terminal tags, the cDNAs were cloned into pXNS2pat-Strep (V42; Cao et al., 2010) and pXNS2pat-myc (V103), generated from V42 by replacing a XhoI SfoI fragment with the annealed primers N0379 and N0380 encoding a myc tag. NSH2 was further cloned via NcoI and BamHI into pXNS2cpmv-Strep (V90). This vector was created by first cloning a 523-bp XhoI SalI fragment from V69 (Myrach et al., 2017) into the XhoI site of V42 and then amplifying a 189-bp fragment from V69 with primers N0004 and N0005, cutting it with SpeI and XbaI, and cloning it via XbaI into the modified V42. The resulting vector V90 contains 5′ and 3′ untranslated region enhancer sequences of the Cowpea mosaic virus (CPMV) RNA-2, leading to enhanced protein expression (Figure 5A). The NSH1 and NSH2 cDNAs were also cloned into pET30nco-CTH (V48) (Myrach et al., 2017) via NcoI and SmaI for expression of the untagged proteins in Escherichia coli. The cDNAs of NSH1 and NSH2 were amplified with the primers 1905 + P106 and 1903 + P105respectively, to introduce a SmaI site and a STOP codon at the 3′ end.
Site-directed mutagenesis of NSH1 was performed with the primer pairs P468 + P447 and P446 + P469 using the NSH1 cloned in V42 as template. The two fragments were assembled with a NEBuilder HiFi DNA Assembly Cloning kit (New England Biolabs), and the resulting fragment was further amplified with P468 + P469 and cloned into V42 and V103 via NcoI and BamHI. The same procedure was followed to mutate NSH2. The primer pairs used were 0001 + P514 and P513 + 0002.
Plant Material and Cultivation
The following T-DNA mutants of Arabidopsis (Arabidopsis thaliana) from the SALK collection (Alonso et al., 2003), the GABI-Kat collection (Kleinboelting et al., 2012), and the SAIL collection (Sessions et al., 2002) were used: SALK083120, nsh1-1 (At2g36310; Jung et al., 2011); SALK128723, nsh2-1 (At1g05620; Riegler et al., 2011); SAIL305B08, gsda-1 and GK432D08, gsda-2 (At5g28050; Dahncke and Witte, 2013); GK015E03, hgprt-2 (At1g71750; Schroeder et al., 2018); GK049D04, xdh1-2 (At4g34890; Hauck et al., 2014) allele assignment by Ma et al. (2016); and uox-1 (Hauck et al., 2014) allele assignment here. The first mentioning and allele assignments in the literature are cited. The wild type was Col-0.
Double mutants were obtained by crossing uox-1 and xdh1-2, hgprt-2 and uox-1, gsda-1 and uox-1, gsda-2 and uox-1, nsh2-1 and uox-1, nsh1-1 and uox-1, hgprt-2 and xdh1-2, xdh1-2 and gsda-2, xdh1-2 and nsh2-1, xdh1-2 and nsh1-1, gsda-2 and nsh1-1, and nsh1-1 and nsh2-1. Triple mutants were obtained by crossing uox-1 nsh1-1 and uox-1 nsh2-1 and xdh1-2 nsh2-1 and xdh1-2 nsh1-1; the pollen donor is named last, respectively.
Complementation lines for nsh1 were obtained by transforming constructs in the vectors V42 and V103 containing NSH1 cDNAs under the control of the 35S promoter and leading to the expression of N-terminally Strep- or myc-tagged fusion proteins. For complementing nsh2, NSH2 cDNAs cloned into the vectors V90 and V103 were used. A construct in V42 encoding the N-terminally Strep-tagged and mutated D29A version of NSH1 was also transformed into the nsh1 background.
Arabidopsis and N. benthamiana were grown in a controlled growth chamber (at 16-h-light/8-h-dark cycle of 85 μmol m−2 s−1, 22°C day, 18°C night, 60% RH). For metabolic profiling, plants were grown on full nutrient soil or on agar plates prepared with half-strength Murashige and Skoog (MS) media. uox-1 and the double and triple mutants in the uox-1 background were first germinated on plates supplemented with 1% (w/v) Suc and, after seedling establishment, transferred to soil (Hauck et al., 2014). For the allopurinol treatment, plates with 7-d-old seedlings were sprayed with 500 μL of a 200 µM allopurinol solution (Ma et al., 2016).
For the dark stress treatment, plants were grown as described by Schroeder et al. (2018). To harvest root material for the IP experiments, plants were grown hydroponically using an Araponics system (Araponics). Seeds were sown in half-strength MS media with 0.65% phytoagar and placed in the Araponic box filled with deionized water. One week after planting, the Araponic box was filled with nutrient solution formulated as described by Myrach et al. (2017) and renewed every 2 weeks. Plants were grown for 6 weeks in a growth chamber under short-day conditions (Binder 865 with Osram Lumilux lights, 8-h-light/16-h-dark cycle of 60 μmol m−2 s−1, 20°C day, 18°C night, 60% RH).
Protein Purification and Determination of Kinetic Constants
The N-terminally Strep-tagged variants of NSH1 and NSH2 were affinity purified after transient expression in N. benthamiana as described by Werner et al. (2008), but adjusting optical densities (ODs) of the helper Agrobacteria carrying the p19 silencing inhibitor construct to 0.1 and all the other bacterial strains to 0.4. In copurification experiments where Strep- and myc-tagged variants of nucleosidases were coexpressed, the OD was adjusted to 0.2. For activity assays, the heterocomplex was always purified via a Strep-tagged NSH2 enzyme.
The NSH activity was assessed spectrophotometrically using a UV/VIS spectrophotometer (UV-2700; Shimadzu), according to Kopecná et al. (2013) and Parkin (1996). The activity assay was set up at room temperature, in a 0.5-cm quartz cuvette, in a total volume of 1 mL. The reaction buffer was 100 mM HEPES , pH 8, mixed with substrates at specific concentrations. The reaction was initiated by adding 5 to 10 μL of purified NSH1 or NSH2 and 10 to 30 μL of purified NSH1-NSH2 complex. The activity was determined by monitoring the absorption decrease due to the consumption of xanthosine (Δε248 = −3.7 mM−1 cm−1; Kopecná et al., 2013), uridine (Δε280 = −1.8 mM−1 cm−1; Kopecná et al., 2013), and inosine (Δε280 = −0.92 mM−1 cm−1; Parkin, 1996). The protein concentration was determined with a BSA standard curve resolved on an SDS-PAGE gel and stained by colloidal Coomassie Brilliant Blue R 250 (Roth). Band intensities were quantified with an Odyssey Fc Imager (LI-COR). The specific activities of the heterocomplexes were calculated normalizing the activity per milligram of NSH dimer. The kinetic constants were determined using the following concentrations: 0.1, 0.2, 0.4, 0.6, 1.0, and 1.6 mM for uridine; 0.03, 0.06, 0.12, 0.24, 0.30, and 0.45 mM for xanthosine; and 0.25, 0.5, 0.75, 1.0, 1.5, and 2.0 mM for inosine. The curves were recorded three times, and the kinetic parameters were determined by fitting the data to the Michaelis–Menten equation using Prism V4 software (GraphPad).
Electrophoresis and Immunoblotting
SDS gel electrophoresis, immunoblotting for the detection of the Strep-tagged proteins, and Coomassie Brilliant Bl Brilliant Blue R 250 staining were performed as described by Witte et al. (2004). Myc-tagged proteins were detected using a mouse anti–c-myc antibody (catalog no. 11667149001; lot no. 18302300; dilution, 1:400; Roche) and a goat anti-mouse IgG alkaline phosphatase–conjugated secondary antibody (catalog no. A3562; lot no. SLBT8638; dilution, 1:10,000; Sigma-Aldrich). The native NSH1 and NSH2 from Arabidopsis were detected with a rabbit anti-NSH1 antibody (0.2 µg mL−1) and anti-NSH1/2 dual-specific antibody (1 µg mL−1). As secondary antibody, a goat anti-rabbit IgG alkaline phosphatase conjugate (catalog no. A3687; lot no. SLBC3108; dilution, 1:30,000; Sigma-Aldrich) was used.
For the detection of NSH1 and NSH2 in different tissues, the tissues were ground with 2 volumes of protein extraction buffer (100 mM Hepes, pH 8,5 mM EDTA, 15 mM DTT, 100 mM NaCl, and 0.5% [v/v] Triton X-100) and centrifuged at 20,000g at 4°C for 10 min. Samples (12 µL) were loaded on a 10% SDS polyacrylamide gel. The rosette and the root correspond to the entire rosette and root of 4-week-old plants before bolting. The 6-week-old flowering plants were separated into the different tissues according to Schroeder et al. (2018).
Generation of Polyclonal Antisera and Antibody Production
Untagged NSH1 and NSH2 were expressed in E. coli BL21 from cDNAs cloned in vector pET30nco-CTH (V48). Cells were grown in 1.5 liters of Luria-Bertani medium and induced by isopropyl-β-d-thiogalactoside after reaching an OD of 0.5. After 3 h of induction, cells were harvested, the pellet was resuspended in 80 mL of lysis buffer (50 mM Tris-HCl, pH 8.0, 0.25% [w/v] Suc, and 1 mM EDTA, pH 8.0) and vortexed. Then, 200 mg of lysozyme were added in 20 mL of lysis buffer, and the slurry was incubated on ice for 30 min. Cells were disrupted by sonication on ice, and 200 mL of detergent buffer (20 mM Tris-HCl, pH 7.5, 2 mM EDTA, pH 8.0, 200 mM NaCl, 1% [w/v] deoxycholic acid, and 1% [v/v] Nonidet P-40) was added. The lysate was centrifuged at 5000g for 10 min, the supernatant was removed, and the pellet was resuspended in 250 mL of washing buffer (0.5% [v/v] Triton X-100 and 1 mM EDTA, pH 8.0). Centrifugation and resuspension were repeated until a tight pellet was obtained. The pellet was washed in 250 mL of 70% ethanol [v/v], suspended in a small volume of freshly prepared PBS, and used for commercial rabbit antisera production and antibody purification (immunoGlobe).
The initial antibodies against NSH1 and NSH2 showed cross-reactivity for both NSH proteins. Therefore, the anti-NSH1 antiserum was run eight times over a column with immobilized NSH2 to deplete the serum from the anti-NSH2 cross-reacting antibody. The flow-through was then passed over a column with immobilized NSH1, and the eluted fraction represents the specific anti-NSH1 antibody. The dual anti-NSH1/2 was purified from the anti-NSH2 antiserum by running it once over a column with immobilized anti-NSH2, and the eluted fraction was used as antibody.
Immunoprecipitation
IP was performed using the Pierce Classic IP kit using the anti-NSH1 antibody (2 µg). Roots (300 mg) from 6-week-old Arabidopsis plants grown hydroponically in short-day conditions were ground in 540 μL of IP lysis/wash buffer (0.025 M Tris, 0.15 M NaCl, 0.001 M EDTA, 1% [v/v] Nonidet P-40, and 5% [v/v] glycerol, pH 7.4; Pierce Classic IP kit) and 60 μL of protease inhibitor (Complete, EDTA-free, protease inhibitor cocktail; Roche). The cell lysate was clarified at 20,000g for 10 min at 4°C and then treated according to the manufacturer’s instructions.
Liquid Chromatography-Mass Spectrometry Analysis
Metabolite analysis was performed using an Agilent HPLC 1200 system coupled to an Agilent 6460C series triple quadrupole mass spectrometer. Nucleosides, nucleobases, and uric acid were separated by a Polaris 5 C18A column (50 × 4.6 mm; particle size, 5 µm; Agilent). Solvent A was 10 mM ammonium acetate, pH 9.5, to detect (and extract) uric acid and pH 7.5 to detect (and extract) nucleosides and nucleobases. Solvent B was methanol. The flow rate was 0.8 mL min−1 with the following gradient: 0 min, 5% B; 1.5 min, 5% B; 3.5 min, 15% B; 6 min, 100% B; 7 min, 100% B; 7.1 min, 5% B; and 13 min, 5% B. The injection volume was 20 µL. Uric acid was detected in negative mode; purine nucleosides, nucleobases, and uridine were detected in positive mode.
The extraction method from Hauck and Witte (2015) was used and adapted for metabolite extraction. For an efficient disruption of seed tissues, 10 mg of seeds, one 7-mm steel bead, and four 5-mm steel beads were placed in a 2-mL centrifuge vial, frozen in liquid nitrogen, and ground with a mixer mill MM 400 (Retsch) at a frequency of 19 Hz for 4 min 30 s. The sample holders were precooled in liquid nitrogen. Extraction buffer (mobile phase A, 500 µL) was added containing the internal standards (ISTDs) 15N5-guanosine, 15N2-uridine, and 15N4-inosine (Cambridge Isotope Laboratories). Samples were immediately incubated at 95°C for 10 min and then centrifuged at 50,000g for 20 min. To remove any particles, the clarified samples were passed through a micro-spin filter (polyvinylidene difluoride, 0.2 µm; Thermo Fisher Scientific). Before loading, samples were generally diluted 10 times with extraction buffer without ISTDs; for measurement of uric acid, samples were diluted 100 times. For seedlings and roots, 50 mg of fresh material was used and treated in the same way except that grinding was performed with four 5-mm steel beads at 30 Hz. Seedling samples were diluted five times and root samples two times before loading. Rosettes of 4-week-old plants were harvested and frozen at −80°C. The tissue was freeze-dried, and 10 mg of material was extracted as described before, except that 1 mL of extraction buffer was used and the samples were diluted 10 times before measuring.
Guanosine, uridine, and inosine were quantified using the respective ISTD, and all other compounds were quantified by external standards (calibration curve). For external standards, standard solutions of xanthosine (0, 0.03, 0.06, 0.1, 0.5, 1, and 5 µg mL−1), xanthine (0, 0.03, 0.06, 0.1, 0.5, 1, and 5 µg mL−1), hypoxanthine (0, 0.01, 0.03, 0.06, 0.1, 0.5, and 1 µg mL−1), and guanine (0, 0.01, 0.03, 0.06, 0.1, 0.5, and 1 µg mL−1) or uric acid (0, 0.01, 0.05, 0.075, 0.1, 0.25, and 0.5 µg mL−1) were added to the matrix (Col-0 extracts). Guanosine, uridine, xanthosine, xanthine, and guanine were analyzed using the method parameters described by Schroeder et al. (2018). Method parameters for inosine, hypoxanthine, and uric acid are listed in Supplemental Table 3. A signal-to-noise ratio < 10 was considered as not detectable (nd).
Statistical Analysis
All data were analyzed with Prism 8 software. One-way analysis of variance (ANOVA) with Tukey´s post test was used. Different letters represent differences at a significance level of P < 0.05. Statistical analyses results are shown in Supplemental Data Set.
Accession Numbers
Sequence data from this article can be found in the GenBank/EMBL data libraries under the following accession numbers: NSH1 (At2g36310), NSH2 (At1g05620), HGPRT (At1g71750), GSDA (At5g28050), XDH (At4g34890), and UOX (At2g26230).
Supplemental Data
Supplemental Figure 1. Genetic suppression of seed germination and seedling establishment phenotypes of the urate oxidase mutant.
Supplemental Figure 2. Xanthine and guanosine content in rosettes of the XDH GSDA double mutant and the respective single mutants as well as the wild type.
Supplemental Figure 3. Calibration of the anti-NSH1/2 antibody and absolute quantification of NSH1 and NSH2 in roots.
Supplemental Figure 4. Position and function of the Asp mutated to Ala in NSH1 and NSH2.
Supplemental Figure 5. Extended updated model of purine nucleotide catabolism in Arabidopsis.
Supplemental Table 1. Specific activities of nucleoside hydrolases at elevated substrate concentrations for inosine and uridine.
Supplemental Table 2. Primers.
Supplemental Table 3. Method parameters for the LC-MS.
Supplemental Data Set. Statistical analyses results.
Dive Curated Terms
The following phenotypic, genotypic, and functional terms are of significance to the work described in this paper:
Acknowledgments
We thank André Specht and Hildegard Thölke for technical support, Anting Zhu for generating pXNS2cpmv-Strep (V90), Nieves Medina Escobar for generating pXNS2pat-myc (V103), and Lennart Doering for performing the site-directed mutagenesis of NSH1 and NSH2. We also thank Marc Heins, Sebastian Hoffmann, Sue Genschmer, Manuel Maidorn, Robin Meier, Vincenzo Puggioni, Jana Scharnberg, Anne Taraschewski, and Anting Zhu for generating and screening the double and triple Arabidopsis mutants used in this study. This work was financially supported by the Deutsche Forschungsgemeinschaft (grants WI3411/2-1 and WI3411/4-1) and the German Academic Exchange Service (DAAD full PhD fellowship to C.B.).
AUTHOR CONTRIBUTIONS
C.-P.W. conceived and designed this research. C.B. designed and performed the experiments and analyzed the data. C.-P.W. wrote the paper with the assistance of C.B.
- Received November 26, 2018.
- Revised January 28, 2019.
- Accepted February 14, 2019.
- Published February 20, 2019.
References
