A group of chromosomal proteins is specifically released by spermine and loses DNA-binding activity...

Plant Physiol. (1994) 106: 559-566

A Group of Chromosomal Proteins I s Specifically Released by Spermine and loses DNA-Binding Activity upon

Phosphorylation'

Dirk Van den Broeck, Dominique Van Der Straeten, Marc Van Montagu*, and Allan Caplan' Laboratorium voor Genetica, Universiteit Gent, B-9000 Gent, Belgium

Biologically relevant concentrations as low as 500 PM spermine led to the specific release of chromatin-associated proteins from nuclei of rice (Oryza sativa) seedlings. Using a southwestern technique, it was shown that several of these proteins bind DNA. This affinity was lost upon in organello phosphorylation by an endoge- nous kinase. The effect of spermine was very specific. Spermidine was far less effective and putrescine was essentially ineffective in releasing these proteins. The most abundant spermine-released protein was shown to be homologous to the maize HMC1 protein. Our results suggest that spermine induces the release of spermine- released proteins by changing DNA conformation. Binding of these proteins might be sensitive to long-range changes in chromosome structure caused by torsional stress.

Polyamines are important modulators of biological processes such as cell division, stress responses, and development (Smith, 1985; Galston and Kaur-Sawhney, 1990; Koetje et al., 1993), but how they are involved in such diverse phe- nomena is not clear. Most reports correlate the level of indwidual polyamines with different physiological and developmental states but have been unable to clarify whether these molecules are determinants. Controlled manipulation of polyamines is dfficult because their direct uptake is slow and often depends on specific camers, whereas their concentration is highly regulated (for a review on regulation, see Heby and Persson, 1990). The concentration of a given polyamine can also be altered by the combined use of inhibitors of putrescine-synthesizing and spermidine-synthesizing enzymes. Both types of inhibitors have been developed as anti-cancer drugs (Pegg et al., 1982). These agents have also been used to establish the involvement of polyamines in mediating auxin- and ethylene-induced effects on plant growth (Friedman et al., 1982; Lee and Chu, 1992).

Causal relationships were further established by mutating genes involved in polyamine synthesis and/or regulation.

' This work was supported by grants from the Rockefeller Foun- dation (RF 86058 No. 59), the Belgian Programme on Interuniversity Poles of Attraction (Prime Minister's Office, Science Policy Program- ming No. 38), and the Commission of the European Communities TS2-0053-8 (GDF). D.V.D.S. is a Senior Research Associate of the National Fund for Scientific Research (Belgium).

Present address: Department of Bacteriology and Biochemistry, University of Idaho, Moscow, ID 83843.

* Corresponding author; fax 32-9-264-5349.

Tobacco mutants resistant to the inhibitor methylglyoxal bis- (guanylhydrazone), which blocks synthesis of spermidine and spermine, all showed changed polyamine levels as well as a multitude of floral abnormalities (Malmberg and Rose, 1987), whereas a floral developmental mutant in Petunia exhibits changed polyamine levels (Gerats et al., 1988). In yeast, it has been shown that when the Om decarboxylase gene is inactivated, cell division stops but cell expansion continues (Whitney and Moms, 1978; Cohn et al., 1980). On the other hand, a mutant unable to synthesize spermidine still grows, albeit very slowly, but can no longer sporulate (Cohn et al., 1978).

Because interpretation of these in vivo effects is difficult, we decided to look for clear in vitro effects at a molecular level. One can expect that altered chromatin organization and gene expression in general are involved in such processes as cell division and developmental programming. Therefore, we focused on possible effects within the nucleus, where two main actions of polyamines can be anticipated. Polyamines, being cations, show a high affinity for DNA, and approximately 50% of the cell's total content is thought to reside in the nucleus (McCormick, 1978; Snyder, 1989). Moreover, in plants at least one isoform of Om decarboxylase is located in the nucleus, tightly bound to chromatin (Foudouli and Kyr- iakidis, 1989). Spermine and spermidine but not putrescine readily change the B-DNA to Z-DNA equilibrium (Behe and Felsenfeld, 1981) and can also cause DNA bending (Feuer- stein et al., 1990). The interactions between Z-DNA and spermine have been resolved by x-ray crystallography (Egli et al., 1991). It is tempting to speculate that spermine can lead either to binding or detachment of nuclear factors by competition or by conformational changes. In fact, Pfeifer and Riggs (1991) showed that the genomic footprint of the PGK-1 promoter can be affected by the loss of particular factors resulting from the isolation of nuclei in buffers con- taining spermine and spermidine.

Spermine also exerts an effect on the activity of nucleus- associated CKII (Hathaway and Traugh, 1984; Datta et al., 1987). When factors like Myb (Liischer et al., 1990) or the plant factor AT-1 (Datta and Cashmore, 1989) are phosphorylated by CKII, they lose their DNA-binding ability. Other proteins, such as the plant factor GBF-1, gain DNA binding upon phosphorylation (Klimczak et al., 1992).

Abbreviations: CKII, casein kinase 11; HMG, high-mobility group; SRP, spermine-released protein.

559

www.plant.org on October 6, 2015 - Published by www.plantphysiol.orgDownloaded from Copyright © 1994 American Society of Plant Biologists. All rights reserved.

http://www.plantphysiol.org/

http://www.plant.org

5 60 Van den Broeck et aI. Plant Physiol. L'ol. 106, 1994

Our results show that changes in polyamine levels and phosphorylation can coordinately affect the DNA-binding ability of the same set of proteins. We describe some characteristics of the affected proteins including specific release from chromatin by 500 p~ spermine but not putrescine, subsequent phosphorylation by a nuclear kinase, and loss of DNA-binding ability upon phosphorylation. One protein was microsequenced and turned out to be similar to the HMGl homolog of maize.

MATERIALS AND METHODS

Preparation of Nuclei and Chromatin

Nuclei were isolated from leaf tissue of 10-d-old rice seedlings (Oryza sativa var Taichung native 1) essentially as described by Hamilton et al. (1972). Leaves were cut into 5- mm pieces with a razor blade and subsequently submerged in diethyl ether for 1 min. The material was then filtered. To remove traces of ether, the residue was washed once in Hamilton buffer (H buffer: 1.14 M SUC, 5 mM MgCI2, 5 mM /3-mercaptoethanol, 10 mM Tris, pH 7.6). All subsequent manipulations were done at 4OC. Nuclei were squeezed out serially by grinding small aliquots (about 5 g of wet material) with mortar (30 cm diameter) and pestle. The homogenate was collected and stirred gently in H buffer for 1 h to further increase separation between nuclei and cell debris. Upon gentle filtration through two layers of Miracloth (Calbiochem, San Diego, CA), the residue was resuspended in H buffer. The collected filtrates were centrifuged at low speed (1OOOg) to prevent disruption of nuclear structures. The pellet (starch grains, chloroplasts, and nuclei) was resuspended slowly in H buffer, enriched with 0.15% Triton, and incubated on ice for 30 min. Centrifugation and resuspension of the pellet in the same buffer for three more times was usually enough to obtain an essentially chloroplast-free pellet. For all further manipulations, this pellet was washed once in H buffer without Triton to remove the detergent. Starting material (50 g) normally yielded about 7 X lo9 rice nuclei. These were taken up in H buffer to a final volume of 3 mL. This nuclear preparation could be stored for months at -2OOC without any detectable alteration of the characteristics tested.

Chromatin was obtained upon lysis of this nuclear preparation by osmotic shock. Nuclei (5 X lo'), obtained as a pellet by centrifugation, were rapidly resuspended in a hypotonic lysis buffer (10 mL: 20 m~ EDTA, 80 m~ NaCI, 10 m Na2S205, 10 m~ DTT, pH 6.3). Upon incubation for 30 min on ice, chromatin was collected as the fraction that pelleted by a low-speed centrifugation (2500g). This preparation was always used immediately.

Differential Salt Wash and Phosphorylation

To perform an in organello phosphorylation assay, 5 X lo7 nuclei were collected by centrifugation (1OOOg) and resuspended in 50 pL of kination buffer (10 m~ MgC12, 1.14 M SUC, 10 m Tris-HCI, pH 7.5, 50 p~ ATP, and 10 pCi [T-~~PIATP) with different salt concentrations (50, 100, or 150 m~ NaCl) and with or without polyamines. Suc was added to prevent osmotic shock. After incubation at room temperature for 20 min to allow phosphorylation and release/

redishibution of proteins at the respective salt concentration, the nuclei were put on ice and centrifuged at low speed (1 OOOg).

Characterization of Released Proteins

All proteins that remained in solution after centrifugation at the respective salt concentrations, in the presence or absence of polyamines or after being phosphcrylated, were collected by TCA precipitation. Pellets were taken up in solubilization buffer (0.06 M Tris-HC1, pH 6.8, 2% SDS, 100 m DTT, 10% glycerol, 0.002% bromphenol blue), and after adjustment of pH by Tris-base, loaded on a 1-nim-thick SDS- PAGE gel. After electrophoresis, proteins were visualized by silver staining (Ansorge, 1985). Phosphorylated. proteins were detected by autoradiography (XR-Kodak) upon drying the gel on Whatman 3MM paper.

DNA-Binding Assays

The DNA-binding activity of released proteins was assayed using a southwestern technique essentially a!; described by Bowen et al. (1980). After renaturation within the polyacryl- amide gel, the proteins were relocated to a p,iir of nitrocellulose membranes by diffusion for 48 h. Once located on the membrane, DNA-binding polypeptides were detected by binding to a labeled DNA probe. Genomic rice DNA, prepared according to Dellaporta et al. (1983), was multiprime labeled using a kit from Boehringer (Mannheh, Germany).

Preparation of Histones and HMG

HMG proteins were isolated as described t)y Goodwin et al. (1973). The chromatin pellet obtained from 5 X lo7 nuclei was resuspended in 100 pL of 5% perchloric acid with 50 m~ DTT by shaking for 30 min. To diminish contamination of the extracted HMG proteins with histone H2, which is partially solubilized by 5 % perchloric acid, 1 he suspension was kept on ice for another 30 min without agitation. Chro- matin was then pelleted at 1500g for 15 min. Soluble HMG proteins were subsequently precipitated in 20 % TCA.

For the isolation of the histone fraction, the same amount of chromatin was taken up in 100 pL of 0 9 % NaCl and incubated for 15 min on ice. After centrifugation (1500g for 15 min), the pellet was resuspended in 100 pL of 0.9% NaCl, and 11 pL of 2.5 N HCI was added in a drop-wise fashion. The solution was then incubated for 4 h on ice and subsequently centrifuged at 1500g for 15 min. The resulting chromatin pellet was washed once again with 0.215 N HCl. Both supernatants were collected and the histone fraction was precipitated by adding 5 volumes of acetone.

Microsequencing

Starting from 60 g of 10-d-old rice seedlini;~, 7 X lo9 rice nuclei were obtained, providing some 8 pg of the 23-kD protein. This was loaded in two lanes of a :!-"-thick gel and separated by a one-dimensional SDS-PAGE. Further manipulations such as blotting, proteolytic cleavage, recovery, and separation of oligopeptides were performed as described by Bauw et al. (1990).




Spermine-Specific Release of DNA-Binding Proteins 561

RESULTS

Previously, it was shown that the addition of spermine toa preparation of nuclei strongly enhanced the level of endog-enous phosphorylation. The phosphorylation of at least threeprotein species (23, 35, and 55 kD) was entirely sperminedependent (Van den Broeck et al., 1991). We developed asimple assay to analyze the consequences of changes in thephosphorylated state of these proteins on their affinity forchromatin and/or DNA by a differential solubilization in thepresence of increasing amounts of salt.

Differential Salt Wash on NucleiNuclei were prepared as described in "Materials and Meth-

ods' in the presence of high Sue concentrations. In organellophosphorylation was performed in 50, 80, or 150 mM NaClbuffers in the presence and absence of 2 mM spermine. TheSue concentration of all buffers was kept constant to preventlysis of nuclei by osmotic shock. The integrity of the nucleiwas verified by microscopy after staining with 4',6-diami-dino-2-phenylindole (Huang et al., 1986). Low-speed cen-trifugation allowed fractionation into sedimentable and sol-uble components. The supernatant was carefully removedand soluble proteins were subsequently precipitated in 20%TCA and separated by SDS-PAGE. Few proteins precipitatedfrom the supernatant when nuclei had been treated with 150mM salt without any spermine (Fig. 1, lanes 1 and 2). On theother hand, when phosphorylation was performed in the

21-

U-

Figure 1. SDS-PAC E analysis showing the set of proteins specificallyreleased by spermine. Proteins were separated in the presence of1 % SDS on a 12% acrylamide gel. The protein pattern was visualizedby silver staining. Lanes 1 and 2, Proteins released from nuclei with150 mM NaCl without spermine (control). Lanes 3 and 4, Proteinsreleased from nuclei with 150 mM NaCl and 2 mM spermine. Lanes5 and 6, Proteins released from chromatin with 50 HIM NaCl and 2mM spermine (lane 5) or 10 mM spermine (lane 6). Lanes 1 and 3represent proteins phosphorylated during the extraction by addingATP (20 MM final concentration). Lanes 2 and 4 contain proteinsprepared in the absence of ATP. The five polypeptides that arespecifically released by spermine are marked on the right by theirapparent molecular mass. Migration distances of molecular massmarkers are shown on the left.

presence of 2 HIM spermine, a discrete number of proteinscould be precipitated from the supernatant (Fig. 1, lanes 3and 4). Three main bands of 23, 26, and 35 kD were repro-ducibly detected by silver staining. Lanes 3 and 4 also show55- and 100-kD polypeptides that were less reproduciblydetected.

To determine whether the appearance of these proteins inthe supernatant was phosphorylation dependent or sperminedependent, a control reaction was included in which phos-phorylation was restrained by leaving out exogenous ATP.Both the amount and size of released proteins remained thesame (Fig. 1, lane 4): the presence of 2 mM spermine appearedto be sufficient for protein release. Moreover, the endogenouskinase activity itself could be recovered quantitatively fromthe nuclei by a triple 150-mM NaCl wash. No phosphoryla-tion was detectable in the residual nuclei (data not shown).Nevertheless, the same proteins were readily released fromthese nuclei by 2 mM spermine (data not shown). Decreasingthe initial NaCl concentration of 150 to 50 HIM did not changethe pattern of proteins released (Fig. 1, lanes 5 and 6). Thus,the protein release from nuclei appears to depend primar-ily on the addition of spermine and not on their state ofphosphorylation.

Differential Salt Wash on Chromatin

Nuclei were lysed by pelleting and quick resuspension ina hypotonic lysis buffer (see "Materials and Methods"). Whenchromatin obtained as a low-speed centrifugation fractionwas treated as nuclei, the same set of polypeptides wasreleased upon addition of 2 HIM spermine (Fig. 1, lanes 5 and6). The main difference was that all detectable kinase activitywas lost during chromatin preparation (data not shown),confirming that the proteins are directly released by sperminewithout first being phosphorylated. We will refer to theseproteins as spermine-released proteins (SRPs).

The Concentration of Spermine Influences the Release ofSRPs Qualitatively and Quantitatively

All previous extractions were done in 2 or 10 mM spermine.Figure 2 shows the amount of proteins released from chro-matin with concentrations ranging from 100 MM to 20 mM. Atransition occurred between 100 and 500 HM spermine. With100 fiM spermine, no SRPs were visibly released, whereas thereleased amounts of 26- and 35-kD proteins increased onlyslightly when spermine concentrations were raised from 500HM to 20 mM. Simultaneous with this transition, some lowmolecular mass proteins (10-20 kD) as well as one 30-kDprotein that were readily extracted at low spermine concen-trations were no longer detectable at higher concentrations.This was also seen when whole nuclei were extracted (Fig. 2,lanes 1 and 2). The decreased recovery of these proteins inthe presence of 500 ^M spermine or higher concentrationsmight reflect the mechanism through which spermine exertsits effect. It seems as if the conformational state of thechromatin is changed dramatically between concentrationsof 100 and 500 ^M spermine.




562 Van den Broeck et al. Plant Physiol. Vol. 106, 1994

Figure 2. The effect of spermine concentration on the release ofSRPs. Proteins were separated in the presence of 1% SDS on a 12%acrylamide gel. The protein patterns were visualized by silver stain-ing. Lanes 1 and 2 are control lanes and represent a 150-mM NaCIextraction on total nuclei in the absence (lane 1) and the presence(lane 2) of 2 HIM spermine. Lanes 3 to 8 show the proteins releasedfrom chromatin with 50 mM NaCI (lane 3) enriched with 100 MM(lane 4), 500 MM (lane 5), 2 mM (lane 6), 10 mM (lane 7), and 20 mMspermine (lane 8). The 23-, 26-, and 35-kD SRPs are indicated bytheir apparent molecular mass.

The Release of SRPs Is Spermine Specific

To determine whether the effect of spermine depends onits cationic nature or on more specific structural properties,its capacity to release proteins was compared to that of twoother polyamines, spermidine and putrescine, and to mag-nesium. The released amounts of the 23-, 26-, and 35-kDSRPs were measured indirectly by performing a southwesternblot on each extract (see below). After a 2-h exposure, proteinbands were cut out, regardless of whether they were visibleor not, and radioactivity of the bound DNA was measuredin a scintillation counter. The results are summarized graph-ically in Figure 3. The amounts of SRPs released by purrescineand 20 mM MgCl2 fell within background levels. Spermidinedid release SRPs, but far less effectively than spermine. Twomillimolar spermine released more of each SRP than 10 HIMspermidine. This comparison also revealed a differenceamong the SRPs. When the amounts of the 23-, 26-, and 35-kD proteins released by spermine were compared with theamounts released by spermidine, the release of the 26-kDprotein seemed the most specific. About 30-fold more of thisprotein was released by 10 mM spermine than by 10 HIMspermidine, compared with a 5- and a 15-fold difference forthe 23- and the 35-kD proteins, respectively. Similarly, 2 mMspermine released about 20-fold more 26-kD protein than 2mM spermidine, whereas the 23- and 35-kD proteins differedonly 5- and 9-fold, respectively.

At Least Five SRPs Show DNA-Binding Ability

We considered it possible that the SRPs would be chro-mosomal proteins capable of binding DNA. The ability to

bind DNA was tested using a southwestern technique, essen-tially as described by Bowen et al. (1980). Total genomic riceDNA was used as a probe for DNA binding. This techniqueenabled us to visualize DNA-binding proteins after a 20-minexposure (Fig. 3B). The binding and subsequent washingprocedures appeared to be stringent enough to be specific.Lysozyme, which binds DNA nonspecifically (Cattan andBourgoin, 1968), was included as a molecular mass markerin 10-fold higher amounts and did not bind DNA at all inour procedure. The proteins released from chromatin with150 mM NaCI did not bind any genomic DNA (Fig. 4B, firstthree lanes). Only proteins extracted with 2 mM spermineexhibited DNA-binding activity. All five proteins visualizedin this assay (23, 26, 35, 55, and 100 kD) coincided withthose that were visualized by silver staining. DNA bindingof the 55-kD polypeptide could be visualized only after a 60-min exposure. Although southwestern blots tend to swellupon hybridization, all bands could still be correlated bysuperimposing four different patterns: the silver-stained gel,the autoradiogram of the dried gel, and two autoradiograms

80

Vx^>

S

H2 1

Spm

035kDa

Q26kDa

• 23kDa

LfL_0 it 2 10 ,. 2 10 .. 20 50mf

Spd Put MgCI2

kDaI

-26-23

B

Figure 3. Graphic representation of the protein-releasing capacityof spermine, spermidine, putrescine, and magnesium. A, Theamounts of the 23-, 26-, and 35-kD SRPs that were released by 2and 10 mM spermine (Spm) were compared to the amounts thatwere released by 2 and 10 mM spermidine (Spd) and putrescine(Put) and by 20 and 50 mM MgCI2 using the standard extractionprocedure (150 mM NaCI on nuclei). The protein samples obtainedwith each extraction were separated by SDS-PACE and subjectedto a southwestern procedure as shown in B. Membrane piecescontaining the 23-, 26-, and 35-kD polypeptides were cut out andcounted separately in a scintillation counter. The counting resultsare shown as the percentage relative to the highest counting result(amount of the 26-kD SRP released by 10 mM spermine). B, South-western blot demonstrating DNA binding of SRPs. Total genomicrice DNA was used as a probe (see "Materials and Methods").Binding with four proteins (23, 26, 35, and 100 kD) is detected aftera 20-min exposure.




Spermine-Specific Release of DMA-Binding Proteins 563

NUCLEI CHR. NUCLEI

Figure 4. Loss of DNA-binding activity of SRPs upon phosphoryla-tion. A and B result from the same SDS-PAGE gel. Proteins separatedin the different lanes were extracted from nuclei (lanes 1-6) with150 HIM NaCl (lanes 1 -3) or with 150 rriM NaCl and 2 FTIM spermine(lanes 4-6), or from chromatin with 50 mM NaCl and 2 HIM spermine(lane 7). Proteins were phosphorylated by incubation with 20 HMATP (K) or with 20 MM ATP and [-y-32P]ATP (K*). Upon separation,proteins were renatured and then blotted onto two membranes. Ashows the autoradiogram of one membrane after a 24-h exposureand represents the pattern of phosphorylated proteins. B representsthe pattern of DNA-binding proteins. For this, the second mem-brane was hybridized with 32P-labeled total genomic rice DNAfollowing the southwestern procedure as described in "Materialsand Methods." An exposure time of 20 min was sufficient to resultin an autoradiogram as shown in this panel. The 23-, 26-, 35-, and100-kD SRPs are indicated on the right by their apparent molecularmasses.

of nitrocellulose-bound proteins, one showing the labeledphosphoproteins (24-h exposure) and one showing the DNA-binding proteins after a 20-min exposure (for a schematicrepresentation, see Fig. 5).

Addition of 50 HIM spermine during the hybridization andwashes of the southwestern blot did not decrease the amountof DNA bound to these proteins (data not shown), althougha 100-fold lower spermine concentration (500 /UM) is sufficientto release the SRPs from the chromatin (Fig. 2). These datasuggest that the in vivo association of SRPs is different, orthat spermine can exert an effect only on chromatin and noton free linear DNA as used in the southwestern blot.

Phosphorylation Decreases the DNA-BindingAbility of SRPs

We examined whether phosphorylation could decrease theDNA-binding ability of these proteins. Phosphorylation wasallowed to take place during the salt wash by adding coldATP and was visualized by incorporating [7-32P]ATP (see'Materials and Methods'). DNA binding was assessed by asouthwestern blot with total genomic rice DNA as a probe.The results are presented in Figure 4: panel A shows thephosphorylated proteins, panel B displays the proteins bind-ing DNA. The first three lanes in each panel contain proteins

that were released from nuclei by 150 mM NaCl withoutspermine. None of them showed any DNA binding (Fig. 4B).Proteins released from nuclei by 2 mM spermine were sepa-rated in the next lanes. DNA binding was detected for the23-, 26-, 35-, and 100-kD polypeprides (Fig. 4B, fourth lane).Upon incubation with ATP to allow phosphorylation, theylost their DNA-binding ability (Fig. 4B, fifth and sixth lanes).As such, the assay could not distinguish between thoseproteins that were labeled heavily during phosphorylationand those binding the labeled DNA. Note, however, that thesouthwestern blot was exposed for 20 min to visualize theDNA-binding pattern, whereas an overnight autoradiogra-phy was necessary to detect the phosphorylated proteins.

Abundance of the SRPs

The relative abundance of the released SRPs was evaluatedby comparison with the abundance of the HMG proteins andof histones that could be semi-quantitatively recovered bychemical means. HMG proteins were specifically extractedwith 5% perchloric acid, whereas histones were isolated witha 0.25-N HC1 extraction (see "Materials and Methods'). Theabundance of each protein group was estimated by compar-ing the intensities of the silver-stained bands after SDS-PAGE(Fig. 6).

As expected, 10-fold more material of HMG had to beloaded onto a gel to reach the same abundance that wasreached for the histone fraction. In turn, 10- to 20-fold morenuclei were needed to extract an amount of SRPs comparableto that of HMG proteins.

cut gel in two

Silver-staining Diffusion blotting

Drying of gel +autoradlography

Protein-pattern

run

Autoradiography1st P32-pattern

of phosphorylatedproteins

2nd P32-patternof phosphorylated

proteins

Hybridisationwith labe|ed

DNA

pattern of

Figure 5. Procedure for correlation of the silver-stained proteinpattern with the corresponding phosphorylation and DNA-bindingpattern.





1-31

m -21

-U

Figure 6. An estimation of the abundance of the SRPs. SRP, HMC,and histone extractions were performed on chromatin as describedin "Materials and Methods," starting from 5 X 107 rice nuclei.Different amounts of each extract were loaded onto the gel toobtain comparable abundances: 100% of SRP (lane 1), 10% ofHMC (lanes 2 and 3), and 1% of histone extraction (lane 4). HMCwas extracted immediately after shaking for 30 min (lane 3) or afteran additional 30-min incubation without shaking, to decrease H3and H1 contamination (lane 2). Lane 5 shows a total nuclear extractobtained by adding 5 /iL of solubilization buffer to 10s rice nuclei.

The 23-kD SRP Is the HMG1 Homolog of Rice

The 23-kD SRP was sufficiently abundant for microse-quencing protease-derived oligopeptides after scaling up theextraction (see 'Materials and Methods"). Two separate di-gests were performed: a trypsin digest yielded three and achymotrypsin digest yielded two peptides that could be se-quenced. When combined, the obtained amino acid se-quences showed 93% similarity with the HMG1 cDNA frommaize (Fig. 7) (Crasser and Feix, 1991). This similarity can beextended by taking into account the recognition sites of theused proteases: trypsin cleaves after Arg or Lys and chymo-trypsin cleaves after Phe, Trp, or Tyr. Of the 29 sequencedamino acids, only two differed from the maize protein. Onewas a conservative Ser-to-Thr substitution. Besides severalCKH target sites located in the acidic C-terminal region, thereare also two perfect target sites located in the sequenceKSLTEAD, which is bordered by two o-helices and comprisesthe DNA-binding domain.

As shown above, the amount of the 23-kD SRP releasedby spermine was about 10-fold lower than a semi-quantita-tive extraction of HMG1. From this we infer that only afraction (about 10%) of the 23-kD HMG homolog is releasedby spermine in the concentration range from 0.5 to 20 mM.

DISCUSSION

Low levels of spermine (500 ^M) specifically release a setof DNA-binding proteins (SRPs) from chromatin (Figs. 1 and2). The intracellular concentration of spermine in bovinelymphocytes and rat liver has been estimated at around 1mM (Watanabe et al., 1991). Because one can expect polya-mine amounts to vary depending on physiological conditions

such as stress (Flores and Galston, 1984), development(Koetje et al., 1993), or progress through the cell cycle (Tor-rigiani et al., 1987), this releasing concentration of 500 JIMpresumably has biological significance. Different polyaminescan exhibit different physiological effects (Galston and Kaur-Sawhney, 1990). Recently, it was shown (Snyder, 1989) thatsome characteristics of chromatin, isolated from polyamine-depleted HeLa cells (slower sedimentation, higher DNasesensitivity, and Mg solubility), could be restored by additionof spermine but not of spermidine or putrescine. Our resultsalso show a differential effect of spermine when comparedto the other polyamines. Spermidine was much less effective,whereas putrescine was ineffective. We assume that thespecific release of SRPs from chromatin is not an artifact ofour in vitro manipulation but actually reflects an in vivoeffect that occurs whenever the local spermine concentrationreaches a specific threshold level.

Five of the SRPs were shown to bind DNA. One of theSRPs is now proven to be the rice HMG1 homolog. Aconsensus DNA-binding motif was recognized in severaltranscription factors (Jantzen et al., 1990; Sinclair et al., 1990),which was named the HMG-box motif because it was similarto the DNA-binding domain of HMG1 and HMG2. All ofthese transcription factors contact DNA via the minor groovethat leads to a bent DNA duplex (Bianchi et al., 1992). Onemight expect that within this group of factors more than justone (HMG1) will be released by spermine. Currently, we arescaling up our extraction procedure to microsequence theother SRPs.

Extracted SRPs were phosphorylated by an endogenousnuclear kinase(s) resulting in a slightly increased recovery inthe supernatant, as if they had lost their DNA-binding activ-ity. Indeed, the DNA-binding activity of all five SRPs was

M K G A K S K G A A K A D A K L A V K S K G A E K P A K G

R K G K A G K D P N K P K F V F M E E F R K E

-Trl-

F K E K N F K N D R U K

K

S

sL

L

S E S

RflD

D

K

K

A

A

P

P

Y

Y

V

V

A

A

K

K

A N K L

-Tr2-

K L E Y C E S T A A K K A P A K E E E E

E D E E E K S D K S K S E V N D E D D E E G S E E D E D D D E

Figure 7. Alignment of the microsequenced part of the 23-kD SRPwith the aminoacid sequence of HMC1 of Zea mays. The peptidesTrl, Tr2, and Tr3 were isolated after a trypsin digest; Chl and Ch2were isolated after a chymotrypsin digest. Identical amino acids areboxed.




Spermine-Specific Release of DNA-Binding Proteins 565

decreased upon phosphorylation. It could be speculated that when a transient rise in spermine concentration occurs within the nucleus, the resulting dissociation of SRPs from chromatin will be prolonged by this phosphorylation. In a previous report it was shown that at least one SRP was phosphorylated by a CKII-like activity (Van den Broeck et al., 1991). This 23-kD protein has now been shown to be the HMGl homolog of rice. The HMGl protein of maize (Grasser and Feix, 1991) also contains consensus sequences phosphorylated by CKII. It was shown that in vitro phosphorylation of HMGl by CKII (Grasser et al., 1991) does not affect any of the tested DNA-binding activities. Therefore, it remains to be seen which kinase phosphorylates the rice SRPs in organello, such that their DNA-binding activity is influenced.

How spermine can exert such a specific effect remains to be elucidated. One particular effect of spermine is that it can induce the formation of left-handed Z-DNA of altemating purine pyrimidine tracts (Behe and Felsenfeld, 1981). The fact that spermidine is less effective at inducing Z-DNA and putrescine is ineffective correlates with our results. A potential site of action could be one of the repetitive sequences [d(TG)n] that is found in intergenic regions and that readily forms Z-DNA in vitro under appropriate conditions (Single- ton et al., 1982), and possibly also in vivo (Lancillotti et al., 1987). Such a transition of right-handed B-DNA to left- handed Z-DNA will create positive supercoiling along native or chromatinized DNA. This, in tum, could lead to the release of proteins that are inherently storing negative super-helicity by binding cruciform structures or by trapping them in met- astable interconversion intermediates (Sheflin and Spaulding, 1989; Pontiggia et al., 1993). Since HMGl specifically rec- ognizes cruciform structures (Bianchi et al., 1989), it could be expected that spermine would release the rice HMGl homolog, as found. Analogous effects are seen with various DNA intercalators. They too cause a specific release of HMG proteins from chromatin (Schroter et al., 1985) by creating positive supercoiling upon intercalation. The DNA probe, once bound to the SRPs on the southwestem blot, could not be washed away by spermine (data not shown). If spermine works indirectly by increasing positive supercoiling, such a negative result could be anticipated. Since the DNA probe is linear with free ends, these molecules do not have any topological constraint and the formation of left-handed Z- DNA will not lead to positive supercoiling. Spermine will induce stress only near DNA sequences that are sensitive to a B- to Z-DNA transition. As a result, only proteins bound in the vicinity of such sequences are expected to be released by spermine. Indeed, by comparing the amount of the 23-kD SRP released by spermine to the total amount of HMGl protein in the nucleus, we estimated that approximately 10% of all HMGl is susceptible to spermine treatment.

We propose that spermine causes the release of SRPs by changing torsional stress via a B- to Z-DNA transition. First, the spermine-induced release resembles the specific release seen with DNA intercalators that cause positive supercoiling. Second, spermine does not decrease the amount of DNA bound to SRPs on a southwestem blot: this DNA is linear and therefore does not have a topological constraint. Third, the ability of spermine, spermidine, and putrescine to induce

a B- to Z-DNA transition (Behe and Felsenfeld, 1981) coin- cides exactly with their respective capacity to release SRPs.

ACKNO WLEDCMENTS

The authors thank Guy Bauw and José Van Damme for microsequencing, Rudy Dekeyser for advice and concem, Tom Gerats for critical reading of the manuscript, Karel Spruyt and Vera Vermaercke for figures and photographs, and Martine De Cock for help with the manuscript.

Received March 4, 1994; accepted June 28, 1994. Copyright Clearance Center: 0032-0889/94/l06/0559/08.

LITERATURE CITED

Ansorge W (1985) Fast and sensitive detection of protein and DNA bands by treatment with potassium permanganate. J Biochem Biophys Methods 11: 13-20

Bauw G, Holm Rasmussen H, Van den Bulcke M, Van Damme J, Puype M, Gesser B, Celis JE, Vandekerckhove J (1990) Two- dimensional gel electrophoresis, protein electroblotting and microsequencing: a direct link between proteins and genes. Electropho- resis 11: 528-536

Behe M, Felsenfeld G (1981) Effects of methylation on a synthetic polynucleotide: the B-Z transition in poly(dG-mdC).poly(dG- mdC). Proc Natl Acad Sci USA 78: 1619-1623

Bianchi ME, Beltrame M, Paonessa G (1989) Specific recognition of cruciform DNA by nuclear protein HMGl. Science 243

Bianchi ME, Falciola L, Ferrari S, Lilley DMJ (1992) The DNA binding site of HMGl protein is composed of two similar segments (HMG boxes), both of which have counterparts in other eukaryotic regulatory proteins. EMBO J 11: 1055-1063

Bowen B, Steinberg J, Laemmli UK, Weintraub H (1980) The detection of DNA-binding proteins by protein blotting. Nucleic Acids Res 8: 1-20

Cattan D, Bourgoin D (1968) Interactions between DNA and hen’s egg white lysozyme: effects of composition and structural changes of DNA, ionic strength and temperature. Biochim Biophys Acta

Cohn MS, Tabor CW, Tabor H (1978) Isolation and characterization of Saccharomyces cerevisiae mutants deficient in S-adenosylmethi- onine decarboxylase, spermidine, and spermine. J Bacteriol 134:

Cohn MS, Tabor CW, Tabor H (1980) Regulatory mutations affect- ing ornithine decarboxylase activity in Saccharomyces cerevisiae. J Bacteriol142 791-799

Datta N, Cashmore AR (1989) Binding of a pea nuclear protein to promoters of certain photoregulated genes is modulated by phosphorylation. Plant Cell 1: 1069-1077

Datta N, Schell MB, Roux SJ (1987) Spermine stimulation of a nuclear NII kinase from pea plumules and its role in the phosphorylation of a nuclear polypeptide. Plant Physiol84 1379-1401

Dellaporta SL, Wood J, Hicks JB (1983) A plant DNA mini-preparation: version 11. Plant Mol Biol Rep 1: 19-21

Egli M, Williams LD, Gao Q, Rich A (1991) Structurqof the pure- spermine form of Z-DNA (magnesium free) at I-A resolution. Biochemistry 3 0 11388-11402

Feuerstein BG, Pattabiraman N, Marton LJ (1990) Molecular me- chanics of the interactions of spermine with DNA: DNA bending as a result of ligand binding. Nucleic Acids Res 18: 1271-1282

Flores HE, Galston AW (1984) Osmotic stress-induced polyamine accumulation in cereal leaves. Plant Physiol 7 5 102-109

Foudouli AC, Kyriakidis DA (1989) Chromatin-associated omi- thine decarboxylase at early stages of growth and differentiation of etiolated mono- and dicotyledonous plants. Plant Growth Regul

Friedman R, Altman A, Bachrach U (1982) Polyamines and root formation in mung bean hypocotyl cuttings. Plant Physiol 7 0

1056-1059

161: 56-67

208-213

8: 233-242

844-848





Galston AW, Kaur-Sawhney R (1990) Polyamines in plant physi- ology. Plant Physiol94 406-410

Gerats AGM, Kaye C, Collins C, Malmberg RL (1988) Polyamine levels in Petunia genotypes with normal and abnormal floral development. Plant Physiol86 390-393

Goodwin GH, Sanders C, Johns EW (1973) A new group of chromatin-associated proteins with a high content of acidic and basic amino acids. Eur J Biochem 38: 14-19

Grasser KD, Feix G (1991) Isolation and characterisation of maize cDNAs encoding a high mobility group protein displaying a HMG- box. Nucleic Acids Res 1 9 2573-2577

Grasser KD, Maier UG, Haass MM, Feix G (1991) Maize high mobility group proteins bind to CCAAT and TATA boxes of a zein gene promoter. J Biol Chem 265 4185-4188

Hamilton RH, Kunsch U, Temperli A (1972) Simple rapid proce- dures for isolation of tobacco leaf nuclei. Anal Biochem 4 9 48-57

Hathaway GM, Traugh JA (1984) Kinetics of activation of casein kinase I1 by polyamines and reversal of 2,3-bisphosphoglycerate inhibition. J Biol Chem 259: 7011-7015

Heby O, Persson L (1990) Molecular genetics of polyamine synthesis in eukaryotic cells. Trends Biochem Sci 1 5 153-158

Huang C-N, Cornejo MJ, Bush BS, Jones RL (1986) Estimating viability of plant protoplasts using double and single staining. Protoplasma 135 80-87

Jantzen H-M, Admon A, Bell SP, Tjian R (1990) Nucleolar transcription factor hUBF contains a DNA-binding motif with homology to HMG-proteins. Nature 344 830-836

Klimczak LJ, Schindler U, Cashmore AR (1992) DNA binding activity of the Arabidopsis G-box binding factor GBFl is stimulated by phosphorylation by casein kinase I1 from broccoli. Plant Cell 4:

Koetje DS, Kononowicz H, Hodges TK (1993) Polyamine metabolism associated with growth and embryogenic potential of rice. J Plant Physioll41: 215-221

Lancillotti F, Lopez MC, Arias P, Alonso C (1987) Z-DNA in transcriptionally active chromosomes. Proc Natl Acad Sci USA 84:

Lee T-M, Chu C (1992) Ethylene-induced polyamine accumulation in rice (Oryza sativa L.) coleoptiles. Plant Physiol 100 238-245

Liischer B, Christenson E, Litchfield DW, Krebs EG, Eisenman RE (1990) Myb DNA binding inhibited by phosphorylation at a site deleted during oncogenic activation. Nature 344: 517-522

87-98

1560-1564

Malmberg RL, Rose DJ (1987) Biochemical genetics of resistance to MGBG in tobacco: mutants that alter SAM decarboxylase or polyamine ratios, and floral morphology. Mol Gen Genet 207: 9-14

McCormick F (1978) Use of cell enucleation in the study of polyamine distribution and metabolism. Adv Polyamine Res 1: 173-179

Pegg AE, Tang KC, Coward JK (1982) Effects of S adenosyl-1,8- diamino-3-thiooctane on polyamine metabolism. Bilxhemistry 21:

Pfeifer GP, Riggs AD (1991) Chromatin differences between active and inactive X chromosomes revealed by genomic i'ootprinting of permeabilized cells using DNase I and ligation-mediated PCR. Genes Dev 5 1102-1113

Pontiggia A, Negri A, Bettrame M, Bianchi M,(1993) Protein HU binds specifically to kinked DNA. Mol Microbiol7: 343-350

Schriiter H, Maier G, Ponsting H, Nordheim A (19fL5) DNA intercalators induce specific release of HMG 14, HMC, 17 and other DNA-binding proteins from chicken erythrocyte chromatin. EMBO

Sheflin LG, Spaulding SW (1989) High mobility group protein 1 preferentially conserves torsion in negatively supercoiled DNA. Biochemistry 28: 5658-5664

Sinclair AH, Berta P, Palmer MS, Hawkins JR, Grifliths BL, Smith MJ, Foster JW, Frischauf A-M, Lovell-Badge R, C:oodfellow PN (1990) A gene from the human sex-determining region encodes a protein with homology to a conserved DNA-binding motif. Nature

Singleton CK, Klysik J, Stirdivant SM, Wells R D (1982) Left- handed Z-DNA is induced by supercoiling in physiological ionic conditions. Nature 299 312-316

Smith TA (1985) Polyamines. Annu Rev Plant Physiol36 117-143 Snyder RD (1989) Polyamine depletion is associatcd with altered

chromatin structure in HeLa cells. Biochem J 260 fs97-704 Torrigiani P, Serafini-Fracassini D, Bagni N (19 87) Polyamine

biosynthesis and effect of dicyclohexylamine during the cell cycle of Helianfhus fuberosus tuber. Plant Physiol84 14l3-152

Van den Broeck D, Van Montagu M, Caplan A (1091) Polyamine induced phosphorylation in rice nuclei. Rice Genetics Newsletter

Watanabe S, Kusama-Eguchi K, Kobayashi H, Iguashi K (1991) Estimation of polyamine binding to macromolecules and ATP in bovine lymphocytes and rat liver. J Biol Chem 266: 20803-20809

Whitney PA, Morris DR (1978) Polyamine auxotmphs of Saccha- romyces cerevisiae. J Bacterioll34 214-220

5082-5089

J 4 3867-3872

346: 240-244

8: 145-149




A group of chromosomal proteins is specifically released by spermine and loses DNA-binding activity...

Documents

Transcript of A group of chromosomal proteins is specifically released by spermine and loses DNA-binding activity...