• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of emborepLink to Publisher's site
EMBO Rep. Mar 2004; 5(3): 297–303.
Published online Feb 13, 2004. doi:  10.1038/sj.embor.7400093
PMCID: PMC1299004
Scientific Report

Structure of the conserved domain of ANAC, a member of the NAC family of transcription factors


The structure of the DNA-binding NAC domain of Arabidopsis ANAC (abscisic-acid-responsive NAC) has been determined by X-ray crystallography to 1.9 Å resolution (Protein Data Bank codes 1UT4 and 1UT7). This is the first structure determined for a member of the NAC family of plant-specific transcriptional regulators. NAC proteins are characterized by their conserved N-terminal NAC domains that can bind both DNA and other proteins. NAC proteins are involved in developmental processes, including formation of the shoot apical meristem, floral organs and lateral shoots, as well as in plant hormonal control and defence. The NAC domain does not possess a classical helix–turn–helix motif; instead it reveals a new transcription factor fold consisting of a twisted β-sheet surrounded by a few helical elements. The functional dimer formed by the NAC domain was identified in the structure, which will serve as a structural template for understanding NAC protein function at the molecular level.

Keywords: NAC domain, transcription factor, DNA binding, abscisic acid response, Arabidopsis thaliana, crystal structure


The NAC protein family comprises a variety of plant proteins that are identifiable by the presence of a highly conserved N-terminal NAC domain, accompanied by diverse C-terminal domains. NAC is an acronym derived from the names of the three genes first described as containing the domain, namely NAM (no apical meristem), ATAF1,2 and CUC2 (cup-shaped cotyledon) (Souer et al, 1996; Aida et al, 1997). NAC proteins appear to be widespread in plants. For example, the genome of Arabidopsis thaliana is estimated to contain at least a hundred NAC-encoding genes, whereas no examples have been found thus far in other eukaryotes (The Arabidopsis Genome Initiative, 2000; Riechmann et al, 2000). NAC proteins have been implicated in transcriptional control in a variety of plant processes. Functions include involvement in the development of the shoot apical meristem and floral organs, and in the formation of lateral roots (Souer et al, 1996; Aida et al, 1997; Xie et al, 2000). Furthermore, NAC proteins have been implicated in responses to stress and viral infections (Xie et al, 1999; Ren et al, 2000; Collinge & Boller, 2001).

Several NAC proteins have been shown to function as transcription activators (Ren et al, 2000; Xie et al, 2000; Duval et al, 2002). The DNA-binding ability of two of these proteins, NAC1 and AtNAM, has been localized to the NAC domain, whereas the C-terminal regions constitute transcriptional activation domains (Xie et al, 2000; Duval et al, 2002). Several NAC genes are hormone inducible (Hoth et al, 2002; Xie et al, 2002; Greve et al, 2003). NAC domains have also been implicated in interactions with other proteins such as viral proteins (Xie et al, 1999; Ren et al, 2000) and RING finger proteins (Xie et al, 2002; Greve et al, 2003). For example, interactions between the auxin-inducible NAC1 protein, involved in lateral root formation, and the RING domain protein SINAT5 attenuate the auxin signal through ubiquitination and degradation of NAC1 (Xie et al, 2002). Recently, a NAC protein from Arabidopsis (ANAC) was identified as an interaction partner of another RING protein (Greve et al, 2003). An examination of interactions between different RING domains and ANAC suggested that RING and NAC interactions may also regulate pathways controlled by the plant stress hormone abscisic acid. The nuclear localization of ANAC and its primary structure characteristics are consistent with a role as a transcription factor.

NAC proteins are thus emerging as important proteins in plant development and biology. Here, the first X-ray crystal structure of a NAC domain, the NAC domain of ANAC, is presented. The structure is used to evaluate previous suggestions on function and interactions, including DNA binding by NAC proteins. The structure provides the framework for understanding the functions of NAC proteins at the molecular level.

Results and Discussion

Overall Structure of the ANAC NAC Domain

The overall structure of the NAC domain monomer consists of a very twisted antiparallel β-sheet, which packs against an N-terminal α-helix on one side (α1) and one shorter helix α2 on the other side (Fig 1). The remaining part of the structure comprises loops between the main secondary structure elements, with only sporadic and very short stretches of secondary structure, for example a short β-strand (β″), which packs in a parallel fashion to the first β-strand in the antiparallel β-sheet, and a helical turn (α[triple prime]) preceding α2. The extreme N-terminal portion, which contains three extra amino acids encoded by the expression vector, is in an extended conformation. Much of this N-terminal tail is poorly defined in the crystal structure, except where it is involved in crystal contacts. The content of observed secondary structure elements is in good agreement (Fig 2) with previous predictions based on sequence and with circular dichroism spectra (Greve et al, 2003).

Figure 1
ANAC NAC domain fold. (A) Stereo diagram showing the Cα trace of monomer B in the dicyanoaurate derivative. The secondary structure elements are colour coded as in (B) for clarity, and disordered regions are drawn in dashed lines. The figure was ...
Figure 2
Sequence alignment of ANAC and nine other NAC proteins from Arabidopsis. Red residues are common to at least half of the sequences, whereas blue residues are chemically similar in more than half of the sequences or similar to the dominating residue. Predicted ...

Searches using DALI (Holm & Sander, 1993) and TOPS-based topology searches (http://www.tops.leeds.ac.uk; Gilbert et al, 2001) failed to identify close structural homologues. For all the hits found by DALI, structural similarities were limited to the central β-sheet, and the Z-score for the best hit was rather low (3.6). None of the TOPS-based topology search hits had a very significant score, and none of the most significant hits was a transcription factor. The architecture of the NAC domain can thus be considered a novel fold for a DNA-binding transcription factor domain.

Nac Dimer

Dimerization of DNA-binding domains is common and can function in modulating the DNA-binding specificity (Müller, 2001). Gel filtration studies on the ANAC NAC domain have shown that it forms mostly dimers in solution (Olsen et al, 2004). The diffraction data showed evidence of almost perfect noncrystallographic two-fold symmetry axes parallel to b and c in the crystal, which could both in principle relate the monomers in the biological dimer (Olsen et al, 2004). The extensive contacts involving conserved residues in the dimer formed by the NCS axis parallel to b strongly suggest that this is the functional dimer (Fig 3A). The interactions between the monomers are summarized in Table 1, and the residues involved are marked in Fig 2. Among the interactions are two prominent salt bridges formed by the conserved Arg 19 and Glu 26 (Fig 3A). A short antiparallel β-sheet is also formed at the dimer interface, with hydrogen bonds between Arg 19(N–H)…Arg 19(C=O) and Tyr 21(N–H)…Gly17(C=O). Many hydrophobic interactions occur at the monomer–monomer interface, most of which involve residues in a conserved N-terminal block (Fig 2).

Figure 3
Functional NAC dimer in the 1.9 Å dicyanoaurate derivative structure. (A) Cartoon representation of the dimer (excluding the first 11 residues in monomer A and nine in monomer B in the disordered N-terminal tail region). The salt bridges between ...
Table 1
Analysis of the dimer interface by the Protein–Protein Interaction Server (http://www.biochem.ucl.ac.uk/bsm/PP/server)

There is overall similarity between the two monomers in the dimer (r.m.s.d. 1.27 and 1.25 Å for 137 and 132 Cα atoms in the dicyanoaurate derivative and the native structure, respectively). However, in the highest resolution dicyanoaurate derivative structure, the heavy atom compound binds asymmetrically to the dimer. Three minor heavy atom sites are located near Cys 33 (two in monomer A, one in B), whereas the main dicyanoaurate site is found near residues Phe 41–Ile 46 in monomer A only. The corresponding region in monomer B forms a short helix (α″), which is absent in monomer A. The structural asymmetry in this region is preserved in the native structure. The two monomers also differ in the N-terminal tail region, which interacts with the region near helix α″ within each monomer and might affect its conformation. As packing effects cannot be excluded, it is difficult at this stage to ascertain whether the asymmetry has biological relevance or not.

Dna Binding

NAC domains from the Arabidopsis proteins NAC1 and AtNAM have been shown to bind specifically to a fragment of the CaMV 35S promoter (Xie et al, 2000; Duval et al, 2002). The ability of the ANAC NAC domain to bind DNA was examined using in vitro gel mobility shift assays. As NAC proteins do not bind DNA randomly (Duval et al, 2002), and no biological target gene has been identified for NAC proteins, the −90 to +9 CaMV promoter fragment was used for the assays (Fig 4). The gel mobility of the labelled −90 to +9 CaMV fragment decreased by the addition of purified NAC domain, showing that the NAC domain of ANAC is also able to bind DNA. The promoter fragment contains several palindromic motifs, which could potentially be the binding site for a NAC domain dimer.

Figure 4
Interactions of recombinant ANAC(1–168) with the −90 to +9 fragment of the CaMV 35S promoter. Lane 1, no ANAC(1–168); lane 2, 10 ng ANAC(1–168); lane 3, 50 ng ANAC(1–168); lane 4, 100 ng ANAC(1–168); ...

However, no well-known DNA-binding motif (Müller, 2001) could be identified in the domain. Duval et al (2002) analysed AtNAM deletion mutants in yeast for their ability to bind the CaMV promoter fragment, and reported that the region corresponding to ANAC residues Val 111–Lys 168 constitutes the DNA-binding domain of the AtNAM protein. On the basis of secondary structure predictions, they hypothesized that the region would fold into a helix–turn–helix structure. The region from Val 111 to Gln 163 (five residues could not be traced at the C-terminus) is shown in yellow in Fig 3A. It consists of βstrands 4, 5 and 6 and the connecting turns, evidently not a helix–turn–helix structure. The only NAC domain region that might be reminiscent of the motif is helices α1a/b and α″ in monomer B. However, the α1–α″ region of the ANAC sequence has four acidic residues and only two basic residues, and is thus not likely to form a DNA-binding motif. Furthermore, the sequence forming α″ is not well conserved in NAC proteins. Surface representations showing the charge distribution on the NAC dimer are presented in Fig 3B–D. It is apparent that one face is rich in positive charges, whereas the other is not. The extremities of β-strands 4 and 5 and the connecting turn protrude from the face of the NAC dimer that is rich in positively charged residues. As this area of the structure is part of the region designated as DNA binding by Duval et al (2002), it is tempting to speculate that it is involved in the interaction with DNA. From alignments of NAC primary sequences, it is apparent that the area between Arg 76 and Arg 138 (ANAC numbering) is particularly rich in conserved, basic residues (see labelling in Fig 2). This corresponds to β1, β2, β3a/b, β4a/b and β5a/b in the NAC domain structure, and thus encompasses the protruding β-‘arm' and a major part of the twisted β-sheet. A number of proteins use β-sheet structures for DNA binding (Luscombe et al, 2000). There are several examples of two-stranded β-sheets that bind in either the major or the minor groove (Tateno et al, 1997). Other modes of DNA recognition by β-sheets include the three-stranded β-sheet of the GCC-box binding domain (Allen et al, 1998) and the five-stranded β-sheet of the GCM domain, in which the edge of the β-sheet protrudes into the major groove (Cohen et al, 2003). Considering the extent of the NAC domain area that is rich in conserved, basic residues, a number of modes of DNA interaction can be envisioned. It will be interesting to determine whether the NAC domain adds to the diversity of ways in which β-sheets mediate DNA binding.

Nuclear Localization Signals

The control of nuclear localization is pertinent to the proper function of many transcription factors. Two NAC proteins, NAC1 and ANAC, have been shown to locate to the nucleus (Xie et al, 2000; Greve et al, 2003), but no nuclear localization signal (NLS) has been experimentally identified.

On the basis of sequence analysis, NAC1 has been predicted to contain a basic bipartite NLS, and the corresponding region in the ANAC sequence has been suggested to constitute a degenerate bipartite NLS (Xie et al, 2000; Greve et al, 2003). The first two residues of the predicted NLS region (marked in Fig 2), Lys 114 and Lys 115, are located on β-strand 4 and are thus partly buried in the structure. The basic residues constituting the C-terminal part of the predicted NLS are, conversely, exposed at the surface. The role of the predicted NLS region in nuclear import of NAC proteins needs further investigation by mutagenesis and localization studies. The structure presented here can aid the design and interpretation of such studies.


In recent years, the importance of NAC proteins in plant development, transcription regulation and regulatory pathways involving protein–protein interactions has been increasingly recognized. The structure determination of the N-terminal domain of ANAC provides the first structural template for a NAC domain. We show that the NAC domain adopts a novel fold for a transcription factor, consisting of a twisted antiparallel β-sheet sandwiched between two helices. The structure suggests that this domain mediates dimerization of the NAC proteins through conserved interactions including a salt bridge, and DNA binding through the NAC dimer face rich in positive charges.


Crystallization and structure determination. The NAC domain of ANAC was recombinantly produced in Escherichia coli and crystallized as described in Olsen et al (2004). Native and derivative data of form III crystals (as described in Olsen et al, 2004) were collected at 100 K at beamlines I711 at MAXLAB, Lund, Sweden (λ=1.0350 Å) and ID29, ESRF, Grenoble, France (λ=0.9999 Å). Form III crystals belong to space group P212121 with cell parameters a=62.4 Å, b=75.3 Å and c=80.5 Å in the dicyanoaurate derivative, from which the highest resolution data were obtained. Extensive data collection statistics have been shown previously (Olsen et al, 2004), and a short summary of the collected data is given in Table 2. The structure was solved by MIRAS after the identification of heavy atom sites in SOLVE (version 2.03; Terwilliger & Berendzen, 1999). Phasing in SOLVE with the number of sites specified in Table 2 gave a score of 14.8 and a figure of merit of 0.55. Further density modification and autotracing were carried out in RESOLVE (version 2.03; Terwilliger, 2000). The NCS symmetry was not identified automatically by RESOLVE, due to asymmetry in the binding of heavy atoms to the NCS-related monomers. The NCS matrix could, however, be identified manually using the information contained in the native Patterson map and the initial autotraced model. This matrix was used in subsequent density modification runs. To exploit the higher resolution of the dicyanoaurate derivative, density modification was carried out using amplitudes from this data set to the full resolution limit (1.9 Å). Approximately half of the backbone structure was autotraced by RESOLVE and then built manually in O (version 7.0.1; Jones et al, 1991) and by further autotracing with ARP/wARP (version 6.0; Perrakis et al, 1999). Crystallographic refinement was carried out in REFMAC5 (version 5.1.24; Murshudov et al, 1997). The final model includes 297 residues where regions A5–A8, A79–A85, B78–B85, 144–151 (A and B) and 164–168 (A and B) could not be traced. The final dicyanoaurate derivative model contains four heavy atom sites of which two had been identified by SOLVE. The four sites include two near Cys 33 in A, one near Cys 33 in B and one near region Phe 41–Ile 46 in A. Due to the low occupancies (0.13, 0.07, 0.10 and 0.21, respectively), only the gold atom of dicyanoaurate has been modelled. Refinement statistics are summarized in Table 3. Towards the end of the refinement, the coordinates were used to refine the structure against the native data set. To ascertain whether dimer asymmetry was also present in the native structure, a simulated annealing omit map was calculated in CNS (version 1.1; Brünger et al, 1998). A round of simulated annealing refinement in CNS was carried out before further refinement in REFMAC5. The native model consists of 296 residues, and the final refinement statistics are also shown in Table 3.

Table 2
Data collection statistics
Table 3
Refinement and model statistics

Gel mobility shift assay. A 100 bp fragment (−90 to +9) of the CaMV 35S promoter was amplified by PCR from pCAMBIA3300 using the following primers: 5′-TTTCAGCGTGTCCTCTCCA-3′ and 5′-ATCTCCACTGACGTAAGGG-3′. The promoter fragment was 3′ end labelled with digoxigenin (DIG) using a DIG Gel Shift Kit (Roche). DNA binding reactions contained 4 μl 5 × binding buffer (100 mM Hepes, pH 7.6, 5 mM EDTA, 50 mM (NH4)2SO4, 5 mM DTT, 1% (w/v) Tween-20, 150 mM KCl), 1 μg poly[d(I-C)], 0.8 ng DIG-labelled DNA and the specified amounts of recombinant ANAC NAC domain in a final volume of 20 μl. In competition experiments, 0.1 μg of unlabelled promoter fragment was added. The binding reactions were incubated for 15 min at room temperature and resolved on a 10% polyacrylamide gel in TBE buffer (PAGEr Gold Precast Gels, Cambrex). Following electrophoresis, electroblotting was performed using a Hybond-N nylon membrane (Amersham Pharmacia Biotech), and chemiluminescent detection was carried out according to the instructions of the manufacturer of the DIG Gel Shift Kit (Roche).


We thank Eva Johansson, Flemming Hansen (University of Copenhagen) and Andrew McCarthy (ESRF) for help with data collection, and MAX-LAB (Lund, Sweden) and ESRF (Grenoble, France) for synchrotron time. Travel to synchrotrons was supported through the EU ‘Access to Research Infrastructures' programme and a Danish National Science Council grant to DANSYNC. H.A.E., L.L.L. and S.L. acknowledge financial support from the Danish National Research Foundation. K.S. was supported by the Danish National Research Foundation, and A.N.O. by a Ph.D. stipend from the University of Copenhagen.


  • Aida M, Ishida T, Fukaki H, Fujisawa H, Tasaka M ( 1997) Genes involved in organ separation in Arabidopsis: an analysis of the cup-shaped cotyledon mutant. Plant Cell 9: 841–857 [PMC free article] [PubMed]
  • Allen MD, Yamasaki K, Ohme-Takagi M, Tateno M, Suzuki M ( 1998) A novel mode of DNA recognition by a β-sheet revealed by the solution structure of the GCC-box binding domain in complex with DNA. EMBO J 17: 5484–5496 [PMC free article] [PubMed]
  • Brünger AT et al. ( 1998) Crystallography & NMR System: a new software suite for macromolecular structure determination. Acta Crystallogr D 54: 905–921 [PubMed]
  • Cohen SX, Moulin M, Hashemolhosseini S, Kilian K, Wegner M, Müller CW ( 2003) Structure of the GCM domain–DNA complex: a DNA-binding domain with a novel fold and mode of target site recognition. EMBO J 22: 1835–1845 [PMC free article] [PubMed]
  • Collinge M, Boller T ( 2001) Differential induction of two potato genes, Stprx2 and StNAC, in response to infection by Phytophthora infestans and to wounding. Plant Mol Biol 46: 521–529 [PubMed]
  • Duval M, Hsieh T-F, Kim SY, Thomas TL ( 2002) Molecular characterization of AtNAM: a member of the Arabidopsis NAC domain superfamily. Plant Mol Biol 50: 237–248 [PubMed]
  • Gilbert D, Westhead D, Viksna J, Thornton J ( 2001) A computer system to perform structure comparison using TOPS representations of protein structure. Comput Chem 26: 23–30 [PubMed]
  • Greve K, La Cour T, Jensen MK, Poulsen FM, Skriver K ( 2003) Interactions between plant RING-H2 and plant-specific NAC (NAM/ATAF1/1/CUC2) proteins: RING-H2 molecular specificity and cellular localization. Biochem J 371: 97–108 [PMC free article] [PubMed]
  • Holm L, Sander C ( 1993) Protein structure comparison by alignment of distance matrices. J Mol Biol 233: 123–138 [PubMed]
  • Hoth S, Morgante M, Sanchez J-P, Hanafey MK, Tingey SV, Chua N-H ( 2002) Genome-wide gene expression profiling in Arabidopsis thaliana reveals new targets of abscisic acid and largely impaired gene regulation in the abi1-1 mutant. J Cell Sci 115: 4891–4900 [PubMed]
  • Jones TA, Zou J-Y, Cowan SW, Kjeldgaard M ( 1991) Improved methods for building protein models in electron density maps and the location of errors in these models. Acta Crystallogr A 47: 110–119 [PubMed]
  • Kraulis PJ ( 1991) MOLSCRIPT: a program to produce both detailed and schematic plots of protein structures. J Appl Crystallogr 24: 946–950
  • Luscombe NM, Austin SE, Berman HM, Thornton JM ( 2000) An overview of the structures of protein–DNA complexes. Genome Biol 1, reviews001 [PMC free article] [PubMed]
  • Merrit EA, Bacon DJ ( 1997) Raster3D photorealistic molecular graphics. Methods Enzymol 277: 505–524 [PubMed]
  • Müller CW ( 2001) Transcription factors: global and detailed views. Curr Opin Struct Biol 11: 26–32 [PubMed]
  • Murshudov GN, Vagin AA, Dodson EJ ( 1997) Refinement of macromolecular structures by the maximum-likelihood method. Acta Crystallogr D 53: 240–255 [PubMed]
  • Nicholls A, Sharp KA, Honig B ( 1991) Protein folding and association: insights from the interfacial and thermodynamic properties of hydrocarbons. Proteins 11: 281–296 [PubMed]
  • Olsen AN, Ernst HA, Lo Leggio L, Johansson E, Larsen S, Skriver K ( 2004) Preliminary crystallographic analysis of the NAC domain of ANAC, a member of the plant-specific NAC transcription factor family. Acta Crystallogr D 60: 112–115 [PubMed]
  • Perrakis A, Morris R, Lamzin VS ( 1999) Automated protein model building combined with iterative structure refinement. Nat Struct Biol 6: 458–463 [PubMed]
  • Ren T, Qu F, Morris TJ ( 2000) HRT gene function requires interaction between a NAC protein and viral capsid protein to confer resistance to turnip crinkle virus. Plant Cell 12: 1917–1925 [PMC free article] [PubMed]
  • Riechmann JL et al. ( 2000) Arabidopsis transcription factors: genome-wide comparative analysis among eukaryotes. Science 290: 2105–2110 [PubMed]
  • Souer E, van Houwelingen A, Kloos D, Mol J, Koes R ( 1996) The No Apical Meristem gene of petunia is required for pattern formation in embryos and flowers and is expressed at meristem and primordia boundaries. Cell 85: 159–170 [PubMed]
  • Tateno M, Yamasaki K, Amano N, Kakinuma J, Koike H, Allen MD, Suzuki M ( 1997) DNA recognition by β-sheets. Biopolymers 44: 335–359 [PubMed]
  • Terwilliger TC ( 2000) Maximum-likelihood density modification. Acta Crystallogr D 56: 965–972 [PMC free article] [PubMed]
  • Terwilliger TC, Berendzen J ( 1999) Automated MAD and MIR structure solution. Acta Crystallogr D 55: 849–861 [PMC free article] [PubMed]
  • The Arabidopsis Genome Initiative ( 2000) Analysis of the genome sequence of the flowering plant Arabidopsis thaliana. Nature 408: 796–815 [PubMed]
  • Thompson JD, Higgins DG, Gibson TJ ( 1994) CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, positions-specific gap penalties and weight matrix choice. Nucleic Acids Res 22: 4673–4680 [PMC free article] [PubMed]
  • Xie Q, Sanz-Burgos AP, Guo H, García JA, Gutiérrez C ( 1999) GRAB proteins, novel members of the NAC domain family, isolated by their interaction with a geminivirus protein. Plant Mol Biol 39: 647–656 [PubMed]
  • Xie Q, Frugis G, Colgan D, Chua N-H ( 2000) Arabidopsis NAC1 transduces auxin signal downstream of TIR1 to promote lateral root development. Genes Dev 14: 3024–3036 [PMC free article] [PubMed]
  • Xie Q, Guo H-S, Dallman G, Fang S, Weissman AM, Chua N-H ( 2002) SINAT5 promotes ubiquitin-related degradation of NAC1 to attenuate auxin signals. Nature 419: 167–170 [PubMed]

Articles from EMBO Reports are provided here courtesy of The European Molecular Biology Organization
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...