Three genes showing distinct regulatory patterns encode the asparagine synthetase of sunflower (Helianthus annuus)


  • María Begoña Herrera-Rodríguez,

    1. Departamento de Biología Vegetal, División de Fisiología Vegetal, Universidad de Córdoba, Avda. San Alberto Magno s/n, E−14071 Córdoba, Spain;
    Search for more papers by this author
  • Susana Carrasco-Ballesteros,

    1. Departamento de Biología Vegetal, División de Fisiología Vegetal, Universidad de Córdoba, Avda. San Alberto Magno s/n, E−14071 Córdoba, Spain;
    Search for more papers by this author
  • José María Maldonado,

    1. Departamento de Fisiología Vegetal y Ecología, Unidad de Fisiología Vegetal, Facultad de Biología, Universidad de Sevilla, Avda, Reina Mercedes 6, E−41012 Seville, Spain
    Search for more papers by this author
  • Manuel Pineda,

    1. Departamento de Bioquímica y Biología Molecular. Universidad de Córdoba, Campus Rabanales, Edif. C-6, 1a Planta, E−14071 Córdoba, Spain;
    Search for more papers by this author
  • Miguel Aguilar,

    1. Departamento de Bioquímica y Biología Molecular. Universidad de Córdoba, Campus Rabanales, Edif. C-6, 1a Planta, E−14071 Córdoba, Spain;
    Search for more papers by this author
  • Rafael Pérez-Vicente

    Corresponding author
    1. Departamento de Biología Vegetal, División de Fisiología Vegetal, Universidad de Córdoba, Avda. San Alberto Magno s/n, E−14071 Córdoba, Spain;
    Search for more papers by this author

Author for correspondence: Rafael Pérez-Vicente Tel: +34 95 7218692 Fax: +34 95 7218939∖ Email:


  •  Asparagine metabolism in sunflower (Helianthus annuus) was investigated by cDNA cloning, sequence characterization and expression analysis of three genes encoding different isoforms of asparagine synthetase (AS, EC
  •  The AS-coding sequences were searched for in leaves, roots and cotyledons by using a methodology based on the simultaneous amplification of different cDNAs. Three distinct AS-coding genes, HAS1, HAS1.1 and HAS2, were identified.
  • HAS1 and HAS1.1 are twin genes with closely related sequences that share some regulatory features. By contrast, HAS2 is a singular sequence that encodes an incomplete AS polypeptide and shows an unusual regulation. The functionality of both the complete HAS1 and the truncated HAS2 proteins was demonstrated by complementation assays. Northern analysis revealed that HAS1, HAS1.1 and HAS2 were differentially regulated dependent on the organ, the physiological status, the developmental stage and the light conditions.
  •  Asparagine synthetase from sunflower is encoded by a small gene family whose members have achieved a significant degree of specialization to cope with the major situations requiring asparagine synthesis.


Asparagine plays a prominent role in nitrogen transport and storage in plants (Sieciechowicz et al., 1988; Lea et al., 1990). The transfer of the amide group from glutamine (or ammonia) to aspartate, catalysed by asparagine synthetase (AS, EC, is the main route for asparagine biosynthesis (Lea et al., 1990). A complete biochemical characterization of plant AS is lacking because of its extreme instability in vitro (Huber & Streeter, 1985; Ta et al., 1989). Therefore, most of the data on plant AS are derived from molecular studies.

Different numbers of genes have been reported to encode AS in plants. A single gene is found in the majority of species studied. However, some leguminous plants contain two closely related genes that show similar regulatory patterns (Tsai & Coruzzi, 1990; Waterhouse et al., 1996; Hughes et al., 1997; Osuna et al., 2001). An exception is Arabidopsis thaliana, which contains three different AS genes (ASN1, ASN2 and ASN3; Lam et al., 1998). All plant AS genes identified to date are homologous to asnB, the gene encoding the glutamine-dependent AS from Escherichia coli (Scofield et al., 1990).

Data from the characterization of cDNA clones suggest that plant AS proteins are composed of 579–591 amino acids with a predicted molecular mass of about 65 kDa. (Shi et al., 1997). Plant AS polypeptides have been divided into group I, group II (Lam et al., 1998), type I and type II (Osuna et al., 2001) according to their position in phylogenetic trees derived from sequence comparisons. However, neither the nature of such sequence differences nor their physiological significance was determined.

Increased levels of asparagine have been observed in carbon-starved tissues from different plant species (Sieciechowicz et al., 1988). Asparagine has a higher nitrogen to carbon (N : C) ratio than glutamine (2 : 4 vs 2 : 5). These data support the role of asparagine as the preferred compound for nitrogen transport and storage under conditions of limited carbon supply, such as in the absence of light. A typical feature of plant AS genes is an enhanced expression in the darkness that is repressed by transfer to light. Both phytochrome and the carbon status of the tissue mediated the light repression of ASN1 from Arabidopsis (Lam et al., 1994).

Nevertheless, asparagine is also required to support the growth of plants in the light. It is essential for protein synthesis in aerial organs. It is also implicated in photorespiration (Ta et al., 1984; Sieciechowicz et al., 1988), a process that occurs in well-illuminated leaves. Moreover, asparagine synthesis in the light has been acknowledged as a mechanism to detoxify high levels of ammonia (Givan, 1979; Sieciechowicz et al., 1988).

Ordinary AS genes that are repressed under illumination cannot account for a direct supply of the asparagine required in the light. An exception to this is Arabidopsis ASN2, a gene that is not repressed but induced by light (Lam et al., 1998). This finding presents Arabidopsis as the only species with AS genes suitable to satisfy a direct demand of asparagine for both light (ASN2) and dark (ASN1) processes (Lam et al., 1994, 1998). Another AS gene not repressed by light was also detected in Phaseolus vulgaris (PVAS1); however, its expression is restricted to roots and to cotyledons at specific stages of germination, and is undetectable in leaves (Osuna et al., 2001).

While the possession of only light-repressed AS genes is the most common situation in plants, the opposite has also been documented. Lotus japonicus contains two AS genes, LJAS1 and LJAS2, not repressed by light, but is devoid of AS genes with enhanced expression in darkness (Waterhouse et al., 1996).

The involvement of asparagine in translocation of the nitrogen fixed by root nodules (Schubert, 1986) has prompted a more detailed study of its biosynthesis in legumes than in non-legume species such as sunflower, where the role of AS in metabolism is still poorly understood. Previous steps of nitrogen utilization by sunflower (i.e. the reduction of nitrate to ammonia and the incorporation of ammonia into glutamine) have been characterized by our group (de la Haba et al., 1988, 1992; Agüera et al., 1990; Maldonado et al., 1990; Montenegro et al., 1998, 1999). We now report on the use of nitrogen one step further. In this paper, we describe the identification, cDNA characterization and expression analysis of HAS1, HAS1.1 and HAS2, three distinct and differently regulated AS genes whose existence suggests a rich metabolism of asparagine in sunflower.

Materials and Methods

Plant culture conditions and sample collection

Sunflower plants (H. annuus L.) from the isogenic cultivar HA-89 (Semillas Cargill SA, Sevilla, Spain) were grown under a 16-h photoperiod with irradiance of 200 µmol m−2 s−1 PAR (provided by Sylvania cool white F72T12/CW/VHO, 160 W fluorescent lamps supplemented with Mazda 60 W incandescent bulbs) and day/night temperature and relative humidity regimes of 25°C/19°C and 70%/80%, respectively. Seeds were germinated in plastic trays containing a 1 : 1 (v : v) mixture of perlite and vermiculite. Seedlings were irrigated daily with a nutrient solution containing 10 mm KNO3 (Hewitt, 1966) and cultured in the trays for up to 25 d. For longer culture periods, plants were transferred individually to pots containing Composana substrate (Compo GmbH, Vchte, Germany).

Cotyledons at germination were collected from 3-d-old-seedlings, 7 d before their full expansion. Senescing cotyledons, having lost 40% of their maximum chlorophyll content (0.7 mg g−1 fresh weight), were obtained from 20-d-old plants, which also provided roots and stems (epicotyls). Immature primary leaves (2–3 cm long) were collected on day 11. Mature, fully expanded primary leaves (9 cm long) were collected on day 20. Dark-adapted leaves were obtained from 20-d-old plants after a 48-h dark treatment. Immature inflorescences (1.5–2 cm diameter) and flowers were obtained from 55- and 70-d-old plants, respectively. All plant samples were frozen by immersion in liquid nitrogen immediately after collection and stored at −80°C until use.

Isolation of cDNA clones from three different sunflower AS genes

Single-stranded cDNA with an anchor sequence at the 5′ end was prepared by priming 5 µg of total RNA with the Qt oligonucleotide (Table 1), as described by Frohman (1994). To amplify the 3′ end of different AS cDNA clones, nested degenerate primers AS2 and AS1, designed to fit conserved regions of plant AS, were paired to nested primers QoR and QiR, designed to fit the anchor sequence Qt (Table 1, Fig. 1). Two consecutive nested amplifications were carried out with the GeneAmp polymerase chain reaction (PCR) Core kit (Perkin-Elmer, Foster City, CA, USA). Each reaction mix contained: 1 unit of Taq DNA polymerase, 200 µm of each of the four dNTPs and 2 mm MgCl2 in 10 mm Tris-HCl/50 mm KCl, pH 8.3 buffer plus 1 µm of each of the external primers AS2 and QoR and 1 µl of the cDNA solution for the first PCR reaction, or 1 µm of each of the internal primers AS1 and QiR and 1 µl of a 1 : 19 dilution of the first PCR products for the second PCR reaction. The PCR reactions were performed in a DNA Thermal Cycler 480 (Perkin-Elmer) using the following 40-cycle profile: 60 s at 94°C, 30 s at 60°C and 180 s at 72°C. Single-primer PCR reactions were performed as a control to identify non-specific PCR products.

Table 1.  Primers used for cDNA synthesis, amplification and labelling
Figure 1.

Polymerase chain reaction (PCR) amplification of sunflower asparagine synthetase (AS)-coding cDNAs. (a) Schematic model of the AS cDNA with indication of the position of primers AS1, AS2, QoR and QiR. The 5′ and 3′ untranslated sequences are represented by lines. (b) Samples of cDNA from leaves (L), roots (R), cotyledons of germinating seeds (CG), senescing cotyledons (SC), and leaves of plants adapted to darkness (DL) were amplified with primers AS2 and QoR, and then reamplified with primers AS1 and QiR (lanes marked with a C). Single-primer PCR reactions with AS1 or QiR were performed to detect nonspecific products (lanes AS1 or QiR). Reactions that did not render detectable fragments are not shown. Bands corresponding to AS-coding cDNA fragments are marked with an asterisk. The PCR products were separated in a high-resolution agarose gel and stained with ethidium bromide. Five microlitres from each amplification reaction (total 50 µl) were loaded per lane. Gel image is shown in negative. Position and length of DNA size markers are indicated.

The products of the second PCR (AS1/QiR) were blunt-ended and separated by electrophoresis. The positive fragments Fas1 and Fas2 were gel-purified and cloned in pBluescript SK(–) (Stratagene, La Jolla, CA, USA) by standard methods (Sambrook et al., 1989).

Sixty randomly selected Fas1 clones from dark-adapted leaves and 60 randomly selected Fas2 clones from cotyledons of germinating seedlings were chosen for restriction analysis. The AS1/QiR inserts from individual Fas1 and Fas2 clones were digested with different restriction enzymes. Two different restriction patterns were observed among the Fas1 clones digested with EcoRV, HaeIII, MluI, or MspI, which led to the identification of a new AS cDNA (Fas1.1).

Partial (3′ end) AS cDNAs Fas 1, Fas 1.1 and Fas2 were sequenced and pairs of nested gene-specific 3′ primers were designed to amplify their corresponding 5′ ends. Gene-specific primers 3GS1 and 3GS1n were designed for Fas1, 3GS11 and 3GS11n for Fas1.1, and 3GS2 and 3GS2n for Fas2 (Table 1). These 3′ primers were located at a distance from the beginning of Fas1, Fas1.1 and Fas2 such that the 5′ and 3′ ends of the cDNAs would overlap a minimum of 130 nucleotides. cDNA fragments of 1570, 1632 and 1568 bp, corresponding to the 5′ end of HAS1, HAS1.1 and HAS2, were obtained by 5′ rapid amplification of cDNA ends-polymerase chain reaction (RACE-PCR) by using the Marathon kit (Clontech, Palo Alto, CA, USA). The sequences of the 5′ and 3′ ends were aligned, allowing the reconstruction of the full-length cDNAs from HAS1, HAS1.1 and HAS2. In addition, full-length cDNAs from HAS1 and HAS2 were amplified directly using 5′ and 3′ gene-specific primers (GEX5H1 and GEX3H1 for HAS1, and GEX5H2 and GEX3H2 for HAS2; Table 1). These full-length cDNAs were cloned in pBluescript SK(–)for sequencing.

Sequencing and sequence analysis

The cDNA inserts in pBluescript SK(–) were sequenced by using the AmpliTaq DyeDeoxy Terminator Cycle Sequencing kit (Perkin-Elmer). Two to three independent clones of each cDNA type were sequenced to avoid PCR-induced mutations. Computer applications used for sequence analysis were those included in the Wisconsin software package (version 9.1-UNIX, 1997) from Genetics Computer Group (Madison, WI, USA) and the Lasergene software package (version 3.0) from DNASTAR (Madison, WI, USA). Multiple sequence alignments were edited and displayed with the Multiple Sequence Alignment Editor & Shading Utility from Genedoc (v. 2.4).

Complementation of an Escherichia coli asparagine auxotroph

HAS1 and HAS2 coding regions were obtained by direct amplification of cDNA with 5′ and 3′ gene-specific primers (GEX5H1, GEX5H2, GEX3H1 and GEX3H2), and cloned in pGEX-KG (Guan & Dixon, 1991) to obtain pHAS1ex and pHAS2ex constructs.

The E. coli asparagine auxotroph strain ER (asnA, asnB, thi-1, relA, spoT1) from the Genetic Stock Center (New Haven, CI, USA) was transformed with the constructs pHAS1ex, pHAS2ex or with the empty vector pGEX-KG, as a control. The ER cells carrying pHAS1ex, pHAS2ex or pGEX-KG and XL1-Blue wild-type cells (Stratagene) were grown at 28°C, in M9 minimal medium (Sambrook et al., 1989) without asparagine. pHAS1ex and pHAS2ex expression was induced by adding 1 mm isopropyl β-d-thiogalactoside (IPTG) to the medium. M9 medium was supplemented with ampicillin (100 µg ml−1) when needed. Bacterial growth was determined by measuring the absorbance of the cultures at 550 nm.

Analyses of DNA and RNA

Total RNA was isolated from different organs by selective precipitation with LiCl, according to the method of Manning (1991). Genomic DNA was obtained from the LiCl supernatant by precipitation with ethanol.

Gene-specific probes for HAS1, HAS1.1 and HAS2 were prepared from their 3′ untranslated regions. Fragments containing the 3′ untranslated regions were excised from Fas1, Fas1.1 and Fas2 cDNAs with EcoRV, HpaI and BbsI, respectively. Single-stranded probes were then prepared by PCR with α32P-dATP, as described by Konat et al. (1994). Primers used for PCR labelling were lab1, lab11 and lab2 (Table 1) for HAS1, HAS1.1 and HAS2, respectively. The HAS1-specific probe encompassed 1765–1941 nt of the cDNA sequence, the HAS1.1-specific probe encompassed 1817–2017 nt of the cDNA sequence and the HAS2-specific probe encompassed 1696–1819 nt of the cDNA sequence.

An AS conserved probe was obtained from a 1 : 1 mix of fragments from the coding region of HAS1 (nt 1393–1764) and HAS2 (nt 1391–1695) that were labelled with α32P-dCTP by random priming with the oligolabellling kit from Pharmacia (Uppsala, Sweden).

Seven micrograms of DNA were digested with XbaI, EcoRV or HindIII, electrophoresed in agarose gels and blotted onto a nylon filter for Southern analysis. Medium stringency hybridization was performed at 65°C, overnight, as described previously (Sambrook et al., 1989) and filters were washed three times (15 min each) in 1× standard saline citrate (SSC), 0.1% (w : v) sodium dodecyl sulphate (SDS) at 65°C.

Samples of total RNA from different organs were separated by denaturing electrophoresis in formaldehyde agarose gels and blotted onto nylon filters according to standard procedures (Sambrook et al., 1989). Hybridization was performed at 42°C, overnight, in solutions containing 50% formamide and the corresponding gene-specific probe. After hybridization, filters were washed twice for 15 min in 0.2× SSC, 0.1% (w : v) SDS at 65°C. Results of the Southern and Northern analyses were revealed by autoradiography after exposing X-ray films (Kodak X-Omat AR) to the filters for 6 d and 5 d at −80°C, respectively.


Cloning and sequence analysis of three AS-coding cDNAs from sunflower

An intensive search for AS-coding sequences was performed by amplifying cDNA samples from leaves, roots, senescing cotyledons and cotyledons from germinating seeds, as well as from leaves from plants adapted to darkness. Degenerate primers against conserved regions were used to amplify sequences from different AS-coding genes. The PCR strategy and results are shown in Fig. 1.

Two positive fragments named Fas1 (582 bp) and Fas2 (462 bp) were simultaneously amplified from senescing cotyledons (Fig. 1). The analysis of their sequences showed that Fas1 and Fas2 corresponded to the 3′ ends of two different AS-coding genes that were named HAS1 and HAS2, respectively (Fig. 2). Fragments with the same sequence as Fas2 were also amplified from cotyledons of germinating seeds, leaves, dark-adapted leaves and roots. Fragments with the same sequence as Fas1 were amplified only from root and dark-adapted leaves, in addition to senescing cotyledons (Fig. 1).

Figure 2.

Alignment of the 3′ ends of the asparagine synthetase (AS)-coding genes HAS1, HAS1.1 and HAS2. Sequences shown were obtained from the cDNA fragments Fas1, Fas1.1 and Fas2, respectively. Arrowhead indicates the beginning of the 3′ variable region. Stop codons are blocked in black. A putative ancestral stop codon is boxed. Putative polyadenylation signals are underlined with a single line (far upstream element) or appear in bold type (near upstream element).

Under the PCR conditions employed, the amplification of a single band does not imply the amplification of a single cDNA. Several cDNAs for AS with a similar size but different sequences can be simultaneously amplified and grouped together in a single band. To assess this possibility, the sequence homogeneity of 60 independent clones of Fas1 obtained from dark-adapted leaves was checked by restriction analysis. The detection of Fas1 clones showing different restriction patterns with the same enzymes allowed identification of a third AS-coding sequence named Fas1.1 (622 bp) that had been amplified simultaneously with Fas1 cDNA from dark-adapted leaves, and was electrophoretically indistinguishable from it. The sequence of Fas 1.1 derived from the 3′ end of a new AS-coding gene named HAS1.1 (Fig. 2). More than 60 independent Fas2 clones were also checked with 20 restriction enzymes and no sequence heterogeneity was detected.

The sequences of the Fas1, Fas 1.1 and Fas2 partial cDNAs were used to design gene-specific primers to amplify the corresponding 5′ ends of each cDNA by 5′ RACE-PCR. Fragments encompassing the 5′ ends were amplified and sequenced, allowing the reconstruction of three full-length cDNAs whose properties are summarized in Table 2.

Table 2.  Summary of sequence properties of the asparagine synthetase (AS) gene family from sunflower
Gene nameGenBank accession no.cDNA length (bp)Polypeptide length (aa)Molecular mass (kDa)Isoelectric pointC-terminal variable regionGln-binding domain

A final PCR reaction was performed with gene-specific primers located at the ends of each reconstructed cDNA that allowed the direct amplification of full-length cDNAs. Sequence comparison between the reconstructed and the directly amplified full-length cDNAs allowed us to discard the possibility that reconstructed cDNAs could be hybrids of 5′ and 3′ ends from different genes.

In the 3′ untranslated regions of HAS1, HAS1.1 and HAS2, several cis-acting elements that may control message termination and processing or stability were identified. Both T-rich far upstream elements (FUEs) and near upstream AATAAA elements (NUEs) involved in polyadenylation (Hunt, 1994; Li & Hunt, 1997) were found (Fig. 2).

The deduced sequences of HAS1, HAS1.1 and HAS2 polypeptides were aligned with AS amino acid sequences from 14 plant species plus the E. coli glutamine-dependent AS asnB (Fig. 3; to save space, only part of the alignment is displayed) and the dendrogram in Fig. 4 was constructed. The three sunflower AS exhibited a high overall sequence identity with the rest of the plant AS examined. The minimum sequence homology value found for the comparison of sunflower HAS1.1 to maize (Zea mays) AS corresponded to 77% of sequence identity at the amino acid level.

Figure 3.

Sequence alignment of the polypeptides encoded by HAS1, HAS 1.1 and HAS2 with others deduced asparagine synthetases (AS). The AS polypeptides shown are sunflower HAS1, HAS1.1 and HAS2, Arabidopsis ASN1, ASN2 and ASN3, Sandersonia AS (SAND), rice AS (ORY) and the glutamine-dependent AS from Escherichia coli (asnB). Identical amino acids are indicated by dots; dashes denote gaps introduced to maximize the homology of the compared sequences. Essential residues from the glutamine-binding domain are marked with an open inverted triangle. Residues proposed to facilitate the binding of aspartate are marked with a dot. An inverted black triangle marks the residues proposed to be responsible for the anchoring of the AMP moiety. The pyrophosphate-binding region is marked with asterisks. Unique positions specific for the class II AS from Fig. 4 are in bold type and underlined. The arrowhead indicates the beginning of the C-terminal variable region.

Figure 4.

Dendrogram of relationships among different asparagine synthetase (AS) polypeptides. The dendrogram was obtained after a multiple sequence comparison. The sequences compared were from Pisum sativum 1 (AS1; X52179) and 2 (AS2; X52180), Lotus japonicus 1 (LJAS1; X89409) and 2 (LJAS2; X89410), Glycine max 1 (SAS1; U77679) and 2 (SAS2; U77678), Arabidopsis thaliana 1 (ASN1; L29083), 2 (ASN2; AF095453) and 3 (ASN3; AF095452), Helianthus annuus 1 (HAS1; AF190728), 1.1 (HAS1.1; AF263432) and 2 (HAS2; AF190729), Phaseolus vulgaris 1 (PVAS1; AJ33522) and 2 (PVAS2; AJ009952), Vicia faba (VFAS1; Z72354), Medicago sativa (U89923), Brassica oleracea (X84448), Elaeagnus umbellata (AS; AF061740), Triphysaria versicolor (AS; AF014055), Asparagus officinalis (X67958), Sandersonia aurantiaca (SAND1; AF005724), Oryza sativa (D83378), Zea mays (X82849) and Escherichia coli (asnB; J05554).

HAS1, HAS1.1 and HAS2 polypeptides contained the proposed glutamine-binding site that is shared by all NH2-terminal nucleophilic (Ntn) amidotransferases (Zalkin & Smith, 1998), a family of enzymes, represented by the glutamine phosphoribosyl pyrophosphate amidotransferase (GPRPPA; EC, that uses glutamine as amide donor. All the essential residues of the glutamine-binding site Cys1–Arg26–Gly27–Gly32–Arg73–Pro86–Asn101–Gly102–Asp127 (GPRPPA numbering; Zalkin & Smith, 1998), were conserved in the sunflower polypeptides (Fig. 3). Essential residues for the binding of other substrates were also conserved in HAS1, HAS1.1 and HAS2. Threonine residues 317 and 318, and arginine residue 320 (HAS1 numbering) were proved to take part in aspartate-binding in E. coli asnB (Boehlein et al., 1997a). Cysteine 524 (HAS1 numbering) was also suggested to facilitate the reaction between aspartate and the enzyme-bound ATP in asnB (Boehlein et al., 1997b). The amino acids occurring between serine residues 234 and 239 (HAS1 numbering) are a conserved motif in enzymes that hydrolyse ATP to AMP and pyrophosphate, such as GMP synthase (Mäntsälä & Zalkin, 1992) or argininosuccinate synthetase (Ratner, 1973; Surh et al., 1988), and constitute a pyrophosphate-binding element (Richards & Schuster, 1998). This segment was conserved in the sunflower AS polypeptides (Fig. 3). Finally, four amino acid residues (Leu233, Val269 Ser343 and Gly344; HAS1 numbering) have been recognized as the points of anchoring of the AMP moiety in asnB (Larsen et al., 1999). Those residues were conserved in the sunflower polypeptides, although valine 269 has been replaced by an isoleucine in HAS2 (Fig. 3).

All the essential residues cited above were also conserved in the AS from A. thaliana, rice and Sandersonia aurantiaca (Fig. 3), as well as in the rest of the plant AS listed in Fig. 4 (data not shown), with the following exceptions: threonine 317 from the aspartate-binding site has been replaced by serine in the pea (Pisum sativum) AS2 polypeptide, cysteine 524 by lysine in the AS from Sandersonia, valine 269 from the AMP-binding site is replaced by isoleucine in ASN2 and ASN3 from Arabidopsis and in the polypeptides from rice and maize, and serine 343 by cysteine in Phaseolus PVAS1.

Against the background of a high overall homology, several unique features differentiated the sunflower AS from each other. HAS1 and HAS 1.1 that encoded two extremely closely related polypeptides (92% sequence identity), differed from one another because of the 3′ variable region (Lam et al., 1994), a poorly conserved stretch encoding the last 35 amino acids from the C-terminal end that is specific for plant AS (Figs 2 and 3). On the other side, HAS2 diverged from HAS1 and HAS1.1 both in size and sequence (Table 2; Fig. 3). The main difference between HAS2 and the other sunflower AS was the absence of the C-terminal variable region due to an early stop codon in the sequence of HAS2 (Figs 2 and 3). A single-nucleotide deletion may explain the origin of this stop codon that produces the truncated HAS2 polypeptide (Fig. 2). A putative ancestral stop codon is found 74 nucleotides downstream from the actual stop codon (Fig. 2). The translation of the mRNA up to this putative ancestral stop codon would give a polypeptide containing the C-terminal variable region. This finding strongly suggests that HAS2 has derived, by a single-nucleotide deletion, from a longer sequence that possessed the C-terminal variable region.

HAS2 also differed from HAS1 and HAS1.1 by about 100 divergent residues scattered along its entire length (Fig. 3). It shares a greater sequence identity with Arabidopsis ASN2 (91%) and ASN3 (90%) and with rice AS (88%) than with the other AS from sunflower (84%).

The dendrogram derived from the comparison of different plant AS polypeptides showed two major dendritic groups (Fig. 4). A large dendritic group (class I) containing most of the AS compared, including HAS1 and HAS1.1, and a small dendritic group (class II) formed by HAS2 plus the AS from rice, maize and Arabidopsis (ASN2 and ASN3) (Fig. 4). Unique sequence features specific to class II sequences lay in seven positions: Lys131, Phe157, Ala163, Ser165, Leu187, Thr264 and Ile267 (HAS2 numbering, Fig. 3). The residues in those positions were not repeated in the other plant AS. Residues that occupy the corresponding positions in the rest of the plant AS were also significantly conserved.

Existence of additional AS-coding genes in sunflower

The existence of additional AS genes was investigated by Southern analysis of the sunflower genome. A probe for AS-conserved regions was prepared as a tool for detecting fragments from any AS-coding gene present in the sunflower genome. This probe was obtained from conserved sequences of both HAS1 and HAS2, and was hybridized to total digested DNA under medium stringency conditions in order to broaden its hybridization spectrum. Gene-specific probes for HAS1, HAS1.1 and HAS2 were prepared from unique regions of their 3′ untranslated sequences.

Hybridization of the conserved probe to DNA digested with either XbaI, EcoRV or HindIII resulted in a multiple banding pattern, as expected from the occurrence of several AS genes in the sunflower genome (Fig. 5). All the bands obtained with the conserved probe could be identified, although with varying intensities, in the banding patterns obtained with the gene-specific probes. This is generally accepted as an indication of the absence of additional genes to the ones the gene-specific probes were made for. However, under the medium stringency conditions used, multiple bands were also obtained with the probe for HAS1.1. A strongly hybridizing band plus two to four weakly hybridizing bands were obtained for each hybridization with the HAS1.1-specific probe, indicating the existence of fragments homologous to the 3′ untranslated sequences of HAS1.1. (Fig. 5). Those weakly hybridizing bands were not detected under high stringency conditions (results not shown). The fact that those bands were detected by the conserved probe (Fig. 5, XbaI digestion) indicates that they also contain AS conserved regions. This strongly suggests the existence of additional AS genes, closely related to HAS1.1, in the sunflower genome. A very low expression level of those genes, below the PCR detection limit, may explain the failure of our approach to amplify them.

Figure 5.

Southern analysis of sunflower DNA. Seven micrograms of total DNA were digested with (a) XbaI, (b) EcoRV or (c) HindIII, separated by agarose gel electrophoresis and blotted onto a nylon filter. Four identical replica filters were hybridized with gene-specific probes for HAS1 (1), HAS1.1 (1.1) HAS2 (2) and with a conserved probe (C). Hybridization was performed under medium stringency conditions to allow the conserved probe to bind sequences from different AS genes. The position and length of DNA size markers are indicated.

Complementation of an asparagine auxotroph mutant strain of E. coli

To test the functionality of the cloned sequences, the regions encoding the proteins HAS1 and HAS2 were inserted in-frame in the pGEX-KG expression vector (Guan & Dixon, 1991). The new constructs (pHAS1ex and pHAS2ex) were transformed and expressed into the E. coli auxotroph ER strain (asnA, asnB, thi-1, relA, spoT1) lacking AS activity (Felton et al., 1980). As expected, growth of E. coli ER transformed with the empty vector was very poor when cultured in a medium without asparagine (Fig. 6). Higher growth levels were obtained in the same medium when the ER strain was transformed with either pHAS1ex or pHAS2ex (40% and 75% of maximum growth level, respectively). However, none of the transformants, carrying either pHAS1ex or pHAS2ex, achieved the growth of the wild-type E. coli used as a positive control. This is, nevertheless, the expected result since transformation with a single heterologous AS gene (pHAS1ex or pHAS2ex) only partly overcomes the lack of two homologous AS genes (asnA and asnB) from E. coli.

Figure 6.

Complementation of an Escherichia coli asparagine auxotrophic strain (ER). Mutant strain ER was transformed with the constructs pHAS1ex, pHAS2ex, or with the empty vector pGEX-KG, as a control. Bacterial strains were inoculated to 0.1 unit of absorbance at 550 nm, at zero time, in minimal media with (filled bars) or without (open bars) 1 mm isopropyl β-d-thiogalactoside (IPTG). Bacterial growth was determined 24 h after inoculation by measuring the absorbance at 550 nm of the culture. Growth percentages represent the mean of three independent experiments with differences below 10% among them. Maximal growth levels obtained by the wild-type E. coli XL1-Blue strain in minimal medium corresponded to 1.40 (+ IPTG) and 1.58 (– IPTG) absorbance units.

Patterns of expression of HAS1, HAS1.1 and HAS2 in sunflower

The abundance of transcripts from each one of the three AS genes from sunflower was determined in a variety of organs under different physiological conditions and developmental stages. Selected organs included immature and mature leaves, leaves of dark-adapted plants, roots, stems (epicotyls), cotyledons from germinating seedlings, senescing cotyledons, immature inflorescences and flowers.

According to the abundance of its transcripts, the expression of HAS1.1 in plants cultured under standard photoperiodic conditions was restricted to roots. However, HAS1 was also expressed in senescing cotyledons, immature inflorescences and flowers. Transcripts from HAS2 were detected in all organs tested, being especially abundant in cotyledons during germination and in immature leaves.

Darkness greatly increased the level of HAS1 and HAS1.1 transcripts in leaves, but not that of HAS2 (Fig. 7).

Figure 7.

Relative levels of HAS1, HAS1.1 and HAS2 mRNA in organs under different physiological and developmental conditions. Twenty micrograms of total RNA from roots (R), epicotyls (E), cotyledons from germinating seedlings (CG), senescing cotyledons (SC), immature leaves (IL), fully expanded mature leaves (ML), immature inflorescences (II), flowers (F) and mature leaves from plants adapted to darkness for 48 h (DL) were separated by electrophoresis, stained with ethidium bromide (bottom) and blotted onto a nylon filter. The filter was hybridized sequentially to probes specific for HAS1, HAS1.1 and HAS2. The old probe was stripped off the filter before each new hybridization. Results were revealed by autoradiography after 5 d of exposure. Hybridization bands correspond to an mRNA of about 2 kb.

Differences in transcript accumulation also depended on the developmental stage. HAS1 was expressed both in immature inflorescences and in flowers, although a greater accumulation of transcripts was found in the mature organ. By contrast, HAS2 transcripts were much more abundant both in the immature inflorescences and immature leaves than in mature leaves and flowers (Fig. 7). Transcripts from HAS1.1 could not be detected in those organs.


Identification of three genes encoding functional asparagine synthetases in sunflower

Most plant species have been found to contain a single AS gene. Model plant A. thaliana was first reported to contain a single AS gene that was isolated by heterologous hybridization (ASN1; Lam et al., 1994). A second approach based on the complementation of a yeast asparagine auxotroph revealed two additional genes for AS in this crucifer (ASN2 and ASN3, Lam et al., 1998). The fact that a plant with a small genome such as Arabidopsis contained three genes for AS led us to expect multiple AS genes in sunflower. Our PCR-based approach permitted us to examine cDNA sequences from several organs in different physiological situations, and led to the identification of three different AS genes: HAS1, HAS1.1 and HAS2. Subsequent Southern analysis (Fig. 5) provided data indicating the existence of additional HAS1.1-related genes in the sunflower genome.

Multigene families are a common feature of plant genomes and have been reported for other enzymes of nitrogen metabolism (i.e. glutamine synthetase, EC This enzyme is encoded by four, five and six isogenes in Arabidopsis, sunflower and maize, respectively (Peterman & Goodman, 1991; Li et al., 1993; Montenegro et al., 1998). The identification of three genes for AS in sunflower and in Arabidopsis strongly suggests that multiple AS genes occur in most species of plants.

All three AS cDNAs from sunflower contained an open reading frame encoding an AS polypeptide very similar to previously isolated plant AS in length, sequence (Fig. 3) and other molecular parameters (Table 2) (Tsai & Coruzzi, 1990; Lam et al., 1994; Chevalier et al., 1996; Waterhouse et al., 1996; Hughes et al., 1997; Shi et al., 1997; Lam et al., 1998; Osuna et al., 1999, 2001). Moreover, polypeptides coded by HAS1 and HAS2 have proved their functionality by complementing an asparagine auxotroph strain of E. coli (Fig. 6).

Sequence characterization of HAS1, HAS1.1 and HAS2

The biochemical characterization of plant AS is very poor. In addition to an extreme in vitro instability, the presence of endogenous inhibitors and contaminating asparaginase in plant extracts have made AS activity extremely difficult to assay (Streeter, 1977; Joy et al., 1983; Huber & Streeter, 1985; Ta et al., 1989). In the absence of direct kinetic data, the possession of the glutamine-binding triad consisting of three conserved residues (Cys1, Asp29 and His101; Mei & Zalkin, 1989), has been the main criterion to identify a plant AS as a glutamine-utilizing amidotransferase (Lam et al., 1994; Chevalier et al., 1996; Waterhouse et al., 1996; Eason & King, 1997; Hughes et al., 1997; Shi et al., 1997; Lam et al., 1998; Eason et al., 2000). However, further evidence has demonstrated that His101 and Asp29 are too distant from Cys1 to participate in catalysis (Smith et al., 1994; Zalkin & Smith, 1998), thus challenging the reliability of the triad as a criterion to identify glutamine-utilizing enzymes.

On the basis of X-ray structural models and mutagenesis experiments, nine invariant residues have been identified as the distinctive fingerprints of the glutamine-binding domain of Ntn amidotransferases (Zalkin & Smith, 1998). Those invariant residues were conserved in the sequence of all sunflower AS (Cys1–Arg30–Gly31–Gly36–Arg49–Pro61–Asn74–Gly75–Asp99, HAS1 numbering; Fig. 3), as well as in the rest of plant AS (Figs 3 and 4). This finding allows the classification of all the AS examined in this paper as genuine glutamine-utilizing Ntn-amidotransferases, including AS2 from Pisum sativum, where the triad essential residues are not conserved (Tsai & Coruzzi, 1990; Lam et al., 1994). Data from direct kinetic studies, when available, will confirm this.

HAS1 and HAS1.1 are twin genes encoding polypeptides of nearly the same length (591 aa and 589 aa, respectively) that share a high sequence identity (92%). Nonetheless, a higher number of amino acid residues in common with the rest of the plant AS examined (result not shown) suggests that HAS1 is evolutionarily older than HAS1.1.

The identification of highly homologous, almost redundant, AS genes has been the rule in plants containing two AS genes, such as Pisum sativum (AS1 and AS2; Tsai & Coruzzi, 1990), Lotus japonicus (LJAS1 and LJAS2; Waterhouse et al., 1996), Glycine max (SAS1 and SAS2; Hughes et al., 1997) and Phaseolus vulgaris (PVAS1 and PVAS2, Osuna et al., 1999, 2001).

The third AS gene found in sunflower, HAS2, encodes a singular polypeptide that has greater similarity to other plant AS than to HAS1 and HAS1.1. In fact, HAS2, together with Arabidopsis ASN2 and ASN3 and rice and maize AS, is classified in a separate dendritic group (Fig. 4).

The division of plant AS into two main groups, one consisting of the AS from rice and maize (the monocot group) and another including the rest of the plant AS available to the authors (the dicot group), was first reported by Shi et al. (1997). The inclusion of ASN2 and ASN3 in the monocot group by Lam et al. (1998) demonstrated that the separation of this group was not due to properties exclusive to monocot AS. Groups were named as group II, containing the AS from rice and maize plus ASN2 and ASN3, and group I, containing 10 AS from plants of the Liliaceae, Brassicaceae and Fabaceae (Lam et al., 1998). We have extended the comparison by including representatives from distantly related families such as Asteraceae (sunflower), Elaeagnaceae (Elaeagnus), and Orobanchaceae (Triphysaria), and confirmed the existence of two major classes of plant AS by identifying and cloning new members from both classes.

Yet another division into type I and type II AS has been established to differentiate plant AS (Osuna et al., 2001). Our extended sequence comparison confirms this division and shows that both types I and II are subgroups that involve, exclusively, closely related legume AS, all of which are included in the higher group I (Fig. 4). Since some confusion may arise from the similarity of the names used for the different kinds of plant AS (type I and II, group I and II), we suggest a change in the nomenclature to class I and II for those AS belonging to the two major groups, and type a and b for the legume-specific types of AS.

Despite constituting a separate group, class II AS did not differ significantly from class I AS in terms of primary structure. In fact, the number of residues specific to class II AS is small (Lys131; Phe157; Ala163; Ser165; Leu187; Thr264 and Ile2, HAS2 numbering). The similarity between class I and class II AS is significant enough to propose that their genes have derived from the duplication of a common ancestor. This should have happened before the separation of plants into monocot and dicot species, since both classes of AS have been isolated from both kinds of plants (Fig. 4). More recent duplications of class I genes might have originated the two closely related types of legumes AS, as well as the twin genes HAS1 and HAS1.1 from sunflower.

A unique feature of the polypeptide encoded by HAS2 is the lack of the C-terminal variable region, a poorly conserved sequence stretch for which no essential roles have been reported. Despite that, it is maintained in most of the plant AS, suggesting some unknown function. Among the other plant AS compared in this paper (Fig. 4), only Sandersonia AS (Eason & King, 1997) lacks the C-terminal variable region (Fig. 3).

Plant AS seem to have evolved from bacterial AS, as suggested by their high degree of sequence identity. Since bacterial AS possess no C-terminal variable region, this region might have been acquired by plant AS, most probably by extending the coding region into part of the 3′ untranslated sequences.

Except for the case of sunflower, no direct evidence has been presented to discard the functionality of this region. In this case, the ability of the truncated HAS2 polypeptide to complement an asparagine auxotroph strain of E. coli clearly indicates the absence of essential catalytic functions in this part of the protein. However, a role of the C-terminal variable region in promoting interactions with other polypeptides or cell components cannot be excluded.

Differential regulation of HAS1, HAS1.1 and HAS2

In addition to a different primary structure, HAS1, HAS1.1 and HAS2 genes showed distinct patterns of regulation according to the organ, the physiological status, the developmental stage and the light conditions.

The expression of HAS1 and HAS1.1 under standard light conditions was confined to a few organs. However, their transcripts accumulated to high levels in leaves of plants adapted to darkness (Fig. 7).

Limitation of the expression to particular organs and repression by light are the two most frequent traits of plant AS genes. Negative effects of light on the expression of AS genes have been reported in P. sativum, Nicotiana tabacum, Nicotiana plumbaginifolia (Tsai & Coruzzi, 1991), Z. mays (Chevalier et al., 1996), Medicago sativa (Shi et al., 1997), Arabidopsis (Lam et al., 1998) and P. vulgaris (Osuna et al., 2001).

Light repression of HAS1 and HAS1.1, as well as most of plant AS genes, is consistent with the role of asparagine in nitrogen transport and storage when carbon supply is limited (Sieciechowicz et al., 1988). However, this pattern of expression does not explain how asparagine is supplied for protein synthesis in well-illuminated leaves and for photorespiration (Ta et al., 1984; Sieciechowicz et al., 1988).

HAS2 did not follow the typical expression pattern of a plant AS since its transcripts were widely distributed in a variety of organs and its expression was not repressed by light (Fig. 7). This unusual pattern strongly suggests that the isozyme encoded by HAS2 is the direct supplier of asparagine in well-illuminated tissues.

Differences in the regulation by light between class I and II AS has been demonstrated in Arabidopsis (Lam et al., 1998), the other species where both classes of AS genes have been isolated. Light repressed the expression of class I ASN1 but induced the expression of class II ASN2. This resembles the situation found in sunflower except that class II HAS2 is unaffected rather than induced by light. The gene encoding AS in rice is another non-light-repressible class II gene (Nakano et al., 2000). The finding of non-light-repressible class II genes in distantly related species such as Arabidopsis, rice and sunflower suggests a wide distribution of this type of gene, thus encouraging a deeper search for class II genes in species where class I genes have already been found. For that purpose, a nonclassical, high-resolution sequence searching methodology such as ours is recommended.

Not all class II genes lack repression by light, as deduced from the absence of AS transcripts in illuminated tissue of maize (Chevalier et al., 1996). Similarly, although most of the class I genes are repressible by light, there are some exceptions such as Lotus LJAS1 and LJAS2 (Waterhouse et al., 1996) and Phaseolus PVAS1 (Osuna et al., 2001), showing that expression in the light is a valuable trait that has been positively selected in the evolution of both classes of AS genes.

The members of the AS gene family from sunflower exhibit a considerable degree of specialization that is not limited to specific response to light. HAS1 and HAS2 were also developmentally regulated (Fig. 7). The fact that the level of HAS2 transcripts was always much higher in young than in mature tissues from both flowers and leaves indicates a role of HAS2 in the construction of these organs rather than in its maintenance. HAS1 appears to have a function complementary to that of HAS2 in floral tissue, since its transcripts accumulate in mature flowers rather than in immature inflorescences (Fig. 7).

Moreover, synthesis of asparagine required for different situations involving nitrogen mobilization, such as germination or senescence, is assisted by different AS genes in sunflower. According to the data of transcript accumulation, HAS2 is expected to play a major role in the synthesis of asparagine during germination, with no significant collaboration from HAS1 and HAS1.1. Conversely, the asparagine formed during cotyledon senescence is expected to be produced primarily by the activity of HAS1 and HAS2, with little contribution from HAS1.1 (Fig. 7).

The expression of sunflower AS genes have been measured in terms of transcript abundance, which may not correlate with the actual levels of AS protein. However, changes in AS mRNA abundance are physiologically relevant in Arabidopsis, as they parallel the level of free asparagine in the plants (Lam et al., 1998).

Class I AS genes have been reported to be expressed in pea cotyledons during germination (Tsai & Coruzzi, 1990), in asparagus (Asparagus officinalis) spears during harvest-induced senescence (Davies & King, 1993) and in senescing flowers from Sandersonia (Eason et al., 2000). However, data on the activity of class I and II AS genes in response to various situations of nitrogen mobilization, or to different developmental stages to compare with the data obtained from sunflower have not yet been presented.

The finding of a small AS gene family, whose members are differentially regulated, supports the existence of a complex and finely regulated asparagine metabolism in sunflower. Major aspects of this metabolism, such as asparagine synthesis in germination and senescence, in young and mature tissue and in light and dark conditions can be accounted for by the AS genes described here. However, a more precise understanding of the in vivo role of each sunflower AS will require the determination of the relative levels of all AS polypeptides and their location in the tissue; this work is in progress.


This work was funded by a grant from Dirección General de Enseñanza Superior e Investigación Científica (DGESIC & BXX 2000–0289) and by Plan Andaluz de Investigación (PAI, group CV-0159), Spain.