Current Genetics

Curr Genet (1992) 21:37-41

9 Springer-Verlag 1992

NIT2, the nitrogen regulatory protein of Neurospora crassa, binds upstream of nia, the tomato nitrate reductase gene, in vitro Gabor Jarai 1, Hoai-Nam Truong 2, Francoise Daniel-Vedele 1, and George A. Marzluf 1 1 Department of Biochemistry,The Ohio State University, 484 West 12th Avenue, Columbus, OH 43210, USA 2 Laboratoire de Biologic Cellulaire, INRA/Vcrsailles, F-78026 Versailles Codex, France Received June 22/August 9, 1991

Summary. The nit-2 gene of Neurospora crassa encodes a trans-acting regulatory protein that activates the expression of a number of structural genes which code for nitrogen catabolic enzymes, including nitrate reductase. The NIT2 protein contains a Cys2/Cys2-type zinc-finger DNA-binding domain that recognizes promoter regions of the Neurospora nitrogen-related genes. The NIT2 zincfinger domain/fl-Gal fusion protein was shown to recognize and bind in a specific manner to two upstream fragments of the nia gene of Lycopersicon esculentum (tomato) in vitro, whereas two mutant NIT2 proteins failed to bind to the same fragments. The dissociation kinetics of the complexes formed between the NIT2 protein and the Neurospora nit-3 and the tomato nia gene promoters were examined; NIT2 binds more strongly to the nit-3 promoter DNA fragment than it does to fragments derived from the plant nitrate reductase gene itself. The observed specificity of the binding suggests the existence of a NIT2-1ike homolog which regulates the expression of the nitrate assimilation pathway of higher plants. Key words: Neurospora - Tomato - Nitrate reductase Nitrate regulation - GATA-binding factor

Introduction The nitrogen regulatory circuit of Neurospora crassa consists of a number of unlinked structural genes which code for enzymes necessary for the utilization of a variety of secondary nitrogen sources when primary sources (e.g., ammonia, glutamine) are unavailable (Marzluf 1981). The control of the circuit involves both metabolic repression and induction and requires the cooperative action of several regulatory genes. The nit-3 gene, which encodes nitrate reductase, is a well characterized structural gene in the nitrogen control circuit (Fu and Marzluf 1988; Okamoto et al. 1991). Its Offprint requests to: G. A. Marzluf

expression is highly regulated at the level of mRNA content and requires nitrogen de-repression and simultaneous induction by nitrate. The regulation of nit-3 is governed by the pathway-specific positive control gene, nit-4, and by nit-2, a major positive-control gene. The nit-2 gene encodes a trans-acting regulatory protein which turns on the expression of various structural genes in the nitrogen circuit. The nucleotide sequence of the nit-2 gene revealed that it encodes a protein that contains a DNA-binding element consisting of a single Cys2/Cys2 "zinc-finger" and an adjacent basic region (Fu and Marzluf 1990 a). The putative DNA-binding domain of the NIT2 protein was expressed and was shown to bind to specific regions upstream of the nitrate reductase (Fu and Marzluf 1990 b) and allantoicase genes (Lee et al. 1990) of N. crassa. The related filamentous fungus, Aspergillus nidulans, possesses a nitrogen regulatory gene, areA, which appears to be homologous to nit-2 (Caddick et al. 1986; Kudla et al. 1990). The AREA protein also contains a Cys2/Cys2-finger type of DNA-binding domain that has 98% amino acid identity with that of NIT2. It was demonstrated that the nit-2 gene of Neurospora can substitute for areA function in vivo in Aspergillus (Davisand Hynes 1987). Furthermore, in vitro gel-band mobilityshift and DNA footprinting studies showed that the NIT2 protein specifically binds to sequences which are closely related to recognition sequences identified in Neurospora nitrogen genes and in the 5' promoter regions of the Aspergillus nitrate and nitrite reductase genes (Fu and Marzluf 1990 b). These data suggest that the function of these nitrogen regulatory proteins is highly conserved among fungi. It appeared to us as an interesting possibility that nitrogen control mediated by a NIT2-1ike factor might be conserved more generally in other eukaryotes, particularly in plants. Nitrate assimilation in plants is also a highly regulated process. Although additional factors such as light, cytokinin, and circadian rhythm, affect nitrate reductase expression in plants (Craword and Campbell 1990; Caboche and Rouze 1990), there are significant

38 similarities to the regulation found in fungi. Thus, in plants, nitrate induces, whereas a m m o n i a or glutamine down-regulates, nitrate reductase expression (Caboche and Rouze 1990; Solomonson and Barber 1990). Furthermore, as found in fungi, the regulation of nitrate reductase expression appears to occur at the m R N A level (Galangau et al. 1988; Vincentz and Caboche 1991). Additionally, nitrate and nitrite reductases are co-regulated both in fungi and in plants (Faure et al. 1991). Very little is known about the trans-acting factors that are required for the regulation o f nitrate reductase in higher plants. With the possible exception of the ehl-2 locus ofArabidopsis thaliana (Cheng et al. 1988), analysis of plant mutants affected in nitrate assimilation has so far failed to uncover loci involved in nitrogen regulation by the nitrogen source. Only limited information exists concerning cis-acting elements that affect the regulation of nitrate reductase in higher plants. However, a 3 kb 5'-upstream region of the t o m a t o nia gene (Daniel-Vedele et al. 1989) was shown to be sufficient for nitrate induction and light regulation when introduced into a nitrate reductasedeficient m u t a n t of Nicotiana plumbaginifolia (M. F. Dorbe, M. Caboche and F. Daniel-Vedele unpublished results). The aim of the present work was to determine whether the N I T 2 nitrogen regulatory protein o f N . erassa is capable of recognizing specific sequences in the 5'-upstream region of the Lyeopersieon eseulentum (tomato) nia gene in vitro. The results obtained suggest the possible existence of an homologous trans-acting regulatory protein in tomato. Materials and methods

Plasmids and DNA fragments. The promoter region of the tomato nitrate reductase gene was cut with SspI and the individual fragments subcloned into the SmaI restriction site of pBluescript (Stratagene, La Jolla, Calif.). The 0.3 kb KpnI-XbaI and the 0.4 kb XbaI-XhoI Y-upstream fragments of the nit-3 gene of N. erassa were also subcloned into pBluescript. The DNA fragments isolated from these subclones were end-labelled by filling in with the Klenow fragment of DNA polymerase I (Sambrook et al. 1989). Expression and purification of the NIT2-fl-galfusion proteins. The construction of the expression vectors using pSKS106 (Casadaban et al. 1983) to express the wild-type and mutant NIT2-fl-Gal fusion proteins was described previously (Fu and Marzluf 1990 b, c). These plasmid constructs were transformed into the E. coliJM103 or XLI-1 Blue host strains, and the bacteria were grown in liquid medium until the absorbance reached approximately 0.5 at 600 nm. Expression of the fusion protein was then induced for 2 h with 0.5 mM IPTG. The fusion proteins were affinity purified from the supernatants of sonicated cells on an aminobenzyl l-thio-fl-D-galactopyranoside-agarose column essentially as described by Ullmann (1984). Gel band mobility-shift experiments. Mobility-shift experiments were carried out as described (Hope and Struhl 1985; Eisen et al. 1988). The reaction mixture contained 0.5-2.0 ~tg of the fusion protein and 1-5pM of the labelled DNA probe (1000030000 cpm) in a buffer consisting of 12 mM HEPES (pH 7.9), 50 mM KCI, 4 mM MgC12, 2 mM DTT, 15% glycerol, 0.01% gelatin, 10 I~gBSA and 1-2 Ixgof poly-(dIdC) in a volume of 25 ~tl.The mixture was incubated at 0~ or at room temperature for 60 or 30 min, respectively, before being loaded onto a 4% or a 5% native polyacrylamide gel. Gels were run in 0.25 • TBE buffer.

DNA-protein binding affinity (off-rate analysis). The method of Fried and Crothers (1981) was used to determine the dissociation rates for the protein-DNA complexes formed between the NIT2 fusion protein and 5' DNA fragments of the tomato or Neurospora nitrate reductase structural genes. The complexes were formed by incubating 1 pM of labelled probes with 1 ~tg of the affinity-purified NIT2 fusion protein in the same buffer as described for the mobility-shift experiments. After incubating the mixture for 1 h at 0 ~ or 30 min at 25 ~ a great excess of the unlabelled DNA fragment was added (200-600-fold molar excess). Equal portions of the mixture were then analyzed as a function of time with the band-shift assay. After drying the gels, the intensities of the individual bands were quantified with a Betascope Blot 603 analyzer (Betagen, Boston, MA). Results

Mobility-shift experiments with different upstream fragments of the tomato nia gene promoter region The N I T 2 regulatory protein of N. crassa binds to elements upstream of several structural genes in the nitrogen regulatory circuit, e.g., nitrate reductase and allantoicase. We used mobility-shift experiments to identify fragments upstream of the nitrate reductase gene of t o m a t o which might be recognized and bound by the Neurospora N I T 2 protein. Five different Sspl restriction fragments, which cover the p r o m o t e r region of the t o m a t o nia gene, were tested in these experiments. These D N A fragments represent a total of 833 bp, from - 8 3 7 to - 4 ( + 1 being the transcriptional start site), and were designated LE1, LE2, LE3, LE4 and LE5, respectively (Fig. 1). Binding sites for regulatory proteins m a y be located both near to, and far upstream of, structural genes. However, it has been demonstrated previously that at least one binding site for the N I T 2 protein is located very close to the transcriptional start site (within 0.2 kb) in all structural genes examined so far, including the nitrate reductase and allantoicase genes of N. crassa as well as the nitrate and nitrite reductase genes of A. nidulans (Fu and Marzluf 1990b; Lee et al. 1990). Consequently, we examined D N A fragments representing the proximal 0.8 kb segment of the p r o m o t e r of the t o m a t o nia gene for possible NIT2-binding. The individual SspI fragments were subcloned, purified, end-labelled and tested in gel retardation experiments with the N I T 2 fusion protein (Fig. 1). F o u r of the five fragments seemed to be retarded to some extent after being incubated with the N I T 2 fusion protein. However, addition of the nonspecific competitor, poly-(dIdC), completely abolished any band-shift of fragments LE2 and LE5 and somewhat weakened the binding to LE1, whereas LE3 appeared to have a higher affinity for the N I T 2 protein. A second retarded band, which is especially strong in the case of LE1, m a y indicate the binding of a second N I T 2 protein molecule to that fragment.

Specific binding of LE1 and LE3 by the NIT2 fusion protein To examine whether the observed binding of N I T 2 to D N A fragments L E I and LE3 is indeed specific, a series

39

1

SspI Ssp_TSspISspI

l

$I LE5

i

LE4 ~--~

LE5

SspI

SspI

$

$

L

i

i

,

LE2

,

LE1

III

/

i

I 0 0 bp,

Fig. I. Binding of different promoter fragments of the nia gene by the NIT2-/~-Gal fusion protein. Upper, mobility-shift experiments. The individual fragments, LE1 (pandA), LE2 (panel B), LE3 (panel C), LE4 (panel D) and LE5 (panel E) were end-labelled and used in mobility-shift experiments with the NIT2 fusion protein. Lanes: 1, free DNA (t-2 ng); 2, the same amount of DNA and 0.5 lag of fusion protein; 3, the same amount of DNA and protein and, in addition, 10 lag of the nonspecific competitor, poly-(dIdC). Lower, restriction map of the nia gene promoter region. The horizontal arrow indicates the start site and direction of transcription, the black bar represents the Y-end of the coding region of the nia gene. The SspI fragments used in this study are shown below the map

of competition experiments were carried out. The competitors used included two fragments of the upstream region o f the N. crassa nit-3 gene. One of these, an 0.4 kb X b a I - X h o I D N A fragment, contains a binding site for the NIT2 protein, whereas the other, an 0.3 kb K p n I - X b a I fragment, does not (Fu and Marzluf, 1990b). As shown in Fig. 2, panel A, unlabelled LE3, or the 0.4 kb XbaIX h o I fragment o f nit-3, competed very strongly with the binding of LE3 by the NIT2 fusion protein, whereas the addition o f the same amount o f the 0.3 kb K p n I - X b a I fragment o f nit-3 did not have a significant effect on the retarded band. These data indicate that binding of LE3 by the NIT2 fusion protein is in fact specific. Similar experiments, carried out with the LE1 fragment (Fig. 2, panel B), also indicated that the binding of LEI by the NIT2 protein is more specific and stronger than that of other unrelated D N A fragments, although LE1 is not bound as tightly as is LE3. The nit-3 promoter fragment, which contains a NIT2-binding site, competed strongly for NIT2 binding of LE 1, whereas the other nit-3 fragment, which lacks a binding site, failed to compete. We concluded, therefore, that both LE1 and LE3 contain

2

3

4

Fig. 2 A, B. Competition for binding the NIT2 fusion protein by LEI, LE3 and and N. crassa nit-3 promoter fragments. Panel A, competition experiments with LE3. Lanes I - 8 contain 1 ng of the labelled LE3 fragment; lanes 2 - 8 also contain 0.5 lag of the NIT2 fusion protein. The unlabelled competitors are: lanes i and 2, none; lane 3, 400 ng of LE3; lanes 4 and5, 200 and 600 ng, respectively, of the 0.4 kb XbaI-XhoI fragment of nit-3; lanes 6 and 7, 200 and 600 rig, respectively, of the 0.3 kb KpnI-XbaI fragment of the nit-3 gene; lane 8, 400 ng of the LE1 fragment. Panel B, competition experiments with LE1. Lanes t - 4 contain 1 ng of the labelled LEt fragment and lanes 2 - 4 also contain 0.5 I~g of the NIT2 fusion protein. The unlabelled competitors are: lanes 1 and 2, none; lane 3, 600 ng of the 0.3 kb KpnI-XbaI fragment of nit-3; lane 4, 600 ng of the 0.4 kb of the XbaI-XhoI fragment of nit-3

a sequence element which is recognized and bound by the NIT2 protein in vitro. Further evidence, which demonstrates the specificity of the D N A binding, was obtained from gel retardation experiments with mutated NIT2 proteins. Two mutant NIT2 proteins, previously designated Sl and $2 (Fu and Marzluf 1990c), were used. Both mutant proteins have substitutions of conserved amino acids of the DNA-binding domain of NIT2. S 1 has two amino acid replacements in the 17-amino acid loop o f the Cys2/Cys2 zinc-finger motif, whereas $2 contains two changes in the adjacent basic region. It was demonstrated previously in N. erassa that both of these mutants lack function in vivo and also failed to bind to the nit-3 gene in vitro. The results obtained revealed that each of these mutant N I T 2 proteins failed to show any detectable binding of LE1 or LE3 (data not shown). Thus, it appears that the observed retardation of these D N A fragments requires an active, functional N I T 2 protein, further confirming the specificity of the binding.

Dissociation kinetics o f the p r o t e i n - D N A complexes

In order to quantitatively compare the strength of binding between the NIT2 fusion protein and the L E I , LE3 and nit-3 D N A fragments, an off-rate analysis was car-

40 nit-3 (data not shown). This technique does not allow the measurement of half-lives shorter then approximately 60 s since the time required for the loading of the fragments onto the gel, or for complexes to enter the gel, are within this time range. Thus, we were not able to differentiate between the binding strengths of LE1 and LE3 for the NIT2 protein; however, they both appear to be bound somewhat more weakly by the NIT2 fusion protein than is the nit-3 fragment.

Discussion

A

100908070-

• °

6050-o C

40-

g .~

30-

20-

\ lO

I 1

B

I

I

2 3 4 time (min)

I

I

5

Fig. 3A, B. Dissociation kinetics of the NIT2-nit-3 protein-DNA complex. Panel A, to an equilibrated mixture of the 0.4 kb XbaIXhoI fragment of nit-3 and the NIT2 fusion protein, a 400-fold molar excess of the same non-radioactive fragment was added and samples were analysed by electrophoresis at several time-points after mixing. Lane 1, zero time-point, before adding the cold competitor; lane 2, 30 s; lane 3, 55 s; lane 4, 75 s; lane 5, 2 min; lane 6, 4 min; lane 7, 8 min; lane 8, 16 min. The small differences in the apparent mobilities of the DNA fragment and the complexes are the result of differences in the time of electrophoresis since each sample was loaded onto the gel as soon as it was taken. Panel B, the dried gels were scanned, the individual band-intensities measured, and the values for the first six time-points plotted ried out. We first examined the kinetics of DNA-protein association. In approximately 10-15 rain at room temperature, an equilibrium was reached by all these complexes (data not shown). Thereafter, we incubated the components for 2 5 - 3 0 min at room temperature to allow DNA-protein complexes to form, before adding a great excess of nonradioactive competitor, in order to prevent any reassociation of the labelled D N A fragment and NIT2 protein once they had dissociated. An aliquot of the equilibrated mixture was loaded onto a polyacrylamide gel (zero time point); then, the competitor was added and further aliquots were taken at several time points and loaded onto the same gel. The results obtained for the dissociation of NIT2 and the nit-3 D N A fragment are shown in Fig. 3. The dried gel was scanned in a beta scanner and the radioactivity within each band was measured and plotted (Fig. 3 B). The half-life of this complex appears to be approximately 80 s. Similar experiments were carried out with the LE1 and LE3 fragments; both seemed to dissociate from NIT2 more rapidly than did

The assimilation of nitrate is a highly regulated process both in fungi and higher plants. The first step in the assimilatory pathway is the reduction of nitrate to nitrite catalysed by nitrate reductase. While the regulatory factors of fungal nitrate reductase expression have been characterized in some detail, very little is known about cis- and trans-acting regulatory elements of higher plants. Here we demonstrate that the NIT2 regulatory protein of Neurospora binds to specific fragments of the tomato nia gene promoter region in vitro. The NIT2 protein was shown to bind to three different sites upstream of the nit-3 gene and to a single site upstream of the alc gene of N. crassa (Fu and Marzluf 1990b; Lee et al. 1990). Although the regions bound by the NIT2 protein were different in size and sequence, a core consensus sequence, TATCT (or on the other strand AGATA) was identified in all of these binding sites. Moreover, each of the four regions bound by NIT2 in the intergenic region of the heterologous A. nidulans niiA-niaD genes also possessed this consensus sequence. It is intriguing that the very same consensus sequence, (A/T)GATA(A/G), was identified as the recognition sequence for the major transcription factor, GATA-1 (previously G F I , EryFl), of the erythroid lineage in the promoters and enhancers of all globin genes (Evans and Felsenfeld 1989). These GATA-1 proteins have two zincfinger DNA-binding domains that are highly conserved among vertebrates, including man (Trainor et al. 1990), mouse (Tsai et al. 1989) and chicken (Evans and Felsenfeld 1989). The GATA-I proteins appear to be a subclass of a large multigene family that includes several transcriptional activators such as the GATA-2 proteins of human endothelial cells (Wilson et al. 1990) and the GATA-3 proteins of the T-lymphocyte cell lineage (Ko et al. 1991). The C-X2-C-X17-C-X2-C zinc-finger motif and an adjacent basic region of the GATA proteins also show significant homology with Neurospora N I T 2 and the Aspergillus A R E A regulatory proteins. This indicates a broad evolutionary conservation of this functional unit, which is utilized in individual organisms as the DNAbinding domain of different specialized transcription factors. In this study, two fragments of the tomato nia gene promoter region, LE1 and LE3, were found to be specifically recognized and bound by the NIT2 protein. Both fragments contain the core GATA recognition sequence; in LE1 the sequence AGATAG (perfect match) is located at - 52 and CGATAA is located at - 182 ( + 1 being the transcriptional start site). LE3 contains a single copy,

41 TGATAT, in the opposite orientation, situated at - 5 6 9 . Although it must be emphasized that all o f these protein~ D N A interactions were observed in vitro, we suggest that a h o m o l o g of N I T 2 m a y exist in t o m a t o and other higher plants and play a central role in controlling nitrate reductase expression and perhaps that of other nitrogen-regulated or light-regulated genes. This suggestion is based on the following facts: (1) Members o f the GATA-binding protein family appear to be conserved a m o n g eukaryotes. (2) The regulatory proteins o f nitrate assimilation are conserved between A. nidulans and N. crassa and belong to the G A T A family. (3) Several features of the regulation o f nitrate assimilation are c o m m o n between fungi and higher plants. (4) The N I T 2 protein recognizes specific D N A fragments o f the t o m a t o nia gene p r o m o t e r region. (5) Nuclear extracts o f tobacco, t o m a t o and pea have been demonstrated to possess multiple D N A - b i n d i n g factors, including factors which appear to recognize a GATA-type sequence (Schindler and Cashmore 1990; Lain and Chua 1989). The results reported here imply that additional experiments, particularly in vivo studies, are needed to demonstrate that LE1 and LE3 in fact possess cis-acting elements which are required for the regulation of nitrate reductase in tomato. It is especially noteworthy that the suggested h o m o l o g y between N I T 2 and a putative transacting factor in t o m a t o m a y facilitate the identification and analysis of such a plant regulatory protein and the isolation and characterization of the corresponding regulatory gene. Acknowledgements. We thank Ying-Hui Fu for providing some of the proteins used in this work. We acknowledge Michele Caboche for encouragement, advice and comments on the manuscript. This research was supported by U.S. Public Health Service Grant GM23367 from the National Institutes of Health to G. A. M. This work was in part supported by a grant of Elf-biorecherches (Labege, France) to H. N. T.

References Caboche M, Rouze P (1990) Trends Genet 6:187-192 Caddick MX, Arst HN Jr, Taylor LH, Johnson RI, Brownlee AG (1986) EMBO J 5:1087-1090

Casadaban MJ, Martinez-Arias A, Shapira SK, Chou J (1983) Methods Enzymol 100:293-308 Cheng C, Dewdney J, Nam H, den Boer BGW, Goodman HM (1988) EMBO J 7:3309-3314 Crawford NM, Campbell WH (1990) Plant Cell 2:829-835 Daniel-Vedele E Dorbe MF, Caboche M, Rouze P (1989) Gene 85:371-380 Davis MA, Hynes MJ (1987) Proc Natl Acad Sei USA 84: 37533757 Eisen A, Taylor WE, Blumberg H, Young ET (188) Mol Cell Biol 8:4552-4456 Evans T, Felsenfeld G (1989) Cell 58:877-885 Faure JD, Vincentz M, Kronenberger J, Caboche M (1991) Plant J (in press) Fried M, Crothers DM (1981) Nucleic Acids Res 9:6505-6524 Fu Y-H, Marzluf GA (1988) J Bacteriol 170:657-661 Fu Y-H, Marzluf GA (1990a) Mol Cell Biol 10:1056-1065 Fu Y-H, Marzluf GA (1990b) Proc Natl Acad Sei USA 87: 53315335 Fu Y-H, Marzluf GA (1990c) Mol Microbiol 4:1847-1852 Galangau F, Daniel-Vedele F, Moureaux T, Dorbe MF, Leydeeker MT, Caboche M (1988) Plant Physiol 88:383-388 Hope IA, Struhl K (1985) Cell 43:177-188 Ko LJ, Yamamoto M, Leonard MW, George KM, Ting P, Engel JD (1991) Mol Cell Biol 1l:2778-2784 Kudla B, Caddick MX, Langdon T, Martinez-Rossi NM, Bennett CF, Sibley S, Davies RW, Arst HN Jr (1990) EMBO J 9: 13551364 Lain E, Chua NH (1989) Plant Cell 1:1147-1156 Lee H, Fu Y-H, Marzluf GA (1990) Biochemistry 29:8779-8787 Marzluf GA (1981) Microbiol Rev 45:437-461 Okamoto P, Fu H-Y, Marzluf GA (1991) Mol Gen Genet 227:213223 Sambrook J, Fritsch EF, Maniatis T (1989) Molecular cloning: a laboratory manual, 2nd edn. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York Schindler U, Cashmore AR (1990) EMBO J 9:3415-3427 Solomonson LP, Barber MJ (1990) Annu Rev Plant Physiol Plant Mol Biol 41:225-253 Tsai SF, Martin DI, Zon LI, E'Andrea AD, Wong GG, Orkin WH (1989) Nature 339:446-451 Trainor CD, Evans T, Felsenfeld G, Boguski MS (1990) Nature 343: 92- 96 Ullmann A (1984) Gene 29:27-31 Vincentz M, Caboche M (1991) EMBO J 10:1027-1035 Wilson DB, Dorfman DM, Orkin Ski (1990) Mol Cell Biol I0:4854-4862 C o m m u n i c a t e d by K. Esser