FEBS Letters 583 (2009) 2727–2733

journal homepage: www.FEBSLetters .org

Binding of divalent cations is essential for the activity of the organellarpeptidasome in Arabidopsis thaliana, AtPreP

Hans G. Bäckman, João Pessoa, Therese Eneqvist, Elzbieta Glaser *

Department of Biochemistry and Biophysics, Stockholm University, SE-106 91 Stockholm, Sweden

a r t i c l e i n f o a b s t r a c t

Article history:Received 3 July 2009Revised 21 July 2009Accepted 22 July 2009Available online 28 July 2009

Edited by Miguel De la Rosa

Keywords:PrePProteasePeptidasomeMetal bindingMitochondriaChloroplast

0014-5793/$36.00 � 2009 Federation of European Biodoi:10.1016/j.febslet.2009.07.040

* Corresponding author. Fax: +46 8153679.E-mail address: e_glaser@dbb.su.se (E. Glaser).

The dual-targeted mitochondrial and chloroplastic zinc metallooligopeptidase from Arabidopsis,AtPreP, functions as a peptidasome that degrades targeting peptides and other small unstructuredpeptides. In addition to Zn located in the catalytic site, AtPreP also contains two Mg-binding sites.We have investigated the role of Mg-binding using AtPreP variants, in which one or both sites wererendered unable to bind Mg2+. Our results show that metal binding besides that of the active site iscrucial for AtPreP proteolysis, particularly the inner site appears essential for normal proteolyticfunction. This is also supported by its evolutionary conservation among all plant species of PreP.

Structured summary:MINT-7231937, MINT-7232017, MINT-7232035, MINT-7232051, MINT-7232070, MINT-7232090:AtPre-P1 (uniprotkb:Q9LJL3) enzymaticly reacts (MI:0414) pF1 beta (uniprotkb:P17614) by protease assay(MI:0435)MINT-7232132:AtPreP1 (uniprotkb:Q9LJL3) enzymaticly reacts (MI:0414) galanin (uniprotkb:P22466) byprotease assay (MI:0435)MINT-7232175:AtPreP1 (uniprotkb:Q9LJL3) enzymaticly reacts (MI:0414) Cecropin A (uniprotkb:P14954)by protease assay (MI:0435)MINT-7232163:AtPreP1 (uniprotkb:Q9LJL3) enzymaticly reacts (MI:0414) hPrPss (uniprotkb:P04156) byprotease assay (MI:0435)

� 2009 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.

1. Introduction

Proteolysis plays a vital role in the cell. Proteins and peptidesare continuously degraded to sustain the cell homeostasis. In addi-tion, the majority of mitochondrial and chloroplastic proteins areencoded in the nucleus and then directed to their correct intracel-lular destination by an N-terminal peptide extension called prese-quence for mitochondrial proteins and transit peptide forchloroplastic proteins (for reviews see [1–5]). After import to therespective organelle the targeting peptide is cleaved off in mito-chondria by the mitochondrial processing peptidase (MPP) [6,7],and in chloroplasts by the stromal processing peptidase (SPP) [8].Targeting peptides can be harmful to cells due to their membranepenetrating properties, which can lead to uncoupling of the mem-brane potential and disruption of the cellular compartmentaliza-tion [9,10].

Presequence protease from Arabidopsis thaliana, AtPreP, is a pro-tease that degrades targeting sequences and other unstructured

chemical Societies. Published by E

peptides of 10–65 amino acids in length in both mitochondriaand chloroplasts [11–14]. AtPreP is a 110 kDa zinc metallooligo-peptidase that belongs to the pitrilysin M16C family of peptidases(MEROPS database) [15]. The M16 peptidases are often referred toas inverzincins, since they comprise an inverted zinc-binding mo-tif, HXXEH, at the active site [16]. In A. thaliana there are two iso-forms of PreP, referred to as AtPreP1 and AtPreP2, with highsequence conservation (89% identity). AtPreP1 and AtPreP2 are du-ally targeted to both the mitochondrial matrix and chloroplaststroma [14]. The yeast homolog of AtPreP, called MOP112 orCYM1p, appears to be located in the mitochondrial intermembranespace [17] whereas the human homolog (hPreP or MP1) is localizedto the mitochondrial matrix and is also known to degrade amyloid-beta peptide (Ab) [18], the toxic agent that accumulates in mito-chondria in Alzheimer‘s disease [19].

The crystal structure of AtPreP1 with a bound peptide in the ac-tive site was determined at 2.1 Å [20]. The structure showed thatthe 995-residue polypeptide folds into four topologically similardomains, which form two bowl-shaped halves connected via a V-shaped hinge region that protrudes from the protease (Fig. 1).The inverted zinc-binding motif of the active site is situated inthe N-terminal portion of the enzyme. However, Arg-848 and

lsevier B.V. All rights reserved.

Fig. 1. (A) Crystallographic structure of AtPreP1 colored from blue (N-terminal) to yellow (C-terminal). The substrate peptide bound to the crystallized protein is shown asorange sticks, the active site zinc as a purple sphere, the two Mg2+ ions as green spheres and waters coordinated to Mg2+ as red spheres. (B) The inner Mg2+-binding site. (C)The outer Mg2+-binding site. (D) The active site and inner Mg2+-binding site located within the proteolytic chamber. (E) SDS–PAGE of the overexpressed and purified wildtypeAtPreP1 and AtPreP1 variants, D1, D2, and D1,2.

2728 H.G. Bäckman et al. / FEBS Letters 583 (2009) 2727–2733

Tyr-854 located at in the C-terminal portion, at a distance of about800 residues in sequence from the zinc-binding motif, contributeto the catalytic site. The structure of AtPreP1 revealed a uniqueproteolytic chamber of more than 10 000 Å3, in which the activesite resides and to which the substrate peptide binds. A novel pro-teolytic mechanism was proposed, involving hinge bending toopen and close the two enzyme halves, allowing the C-terminalresidues Arg-848 and Tyr-854 to interact with N-terminal residuesto form a functional active site [20].

Magnesium ions were needed to facilitate the crystallization ofAtPreP1. Interestingly, the structure also revealed two hydratedmagnesium ions coordinated by acidic residues [20]. One magne-sium-binding site is located inside the proteolytic chamber (the in-ner site), whereas the second site is located on the surface of theenzyme (the outer site). In the present work we have investigatedthe importance of these non-catalytic metal binding sites with re-spect to the enzymatic activity of AtPreP1. We produced AtPreP1variants, in which the inner site (D1), outer site (D2) or both metalbinding sites (D1,2) were replaced with neutral residues unable tobind metal ions. We have investigated the effect of magnesium,calcium and zinc ions on peptide degradation using severalunstructured peptides of different lengths and amino acid compo-sitions as substrates. We also studied the kinetics of degradationby means of a fluorogenic peptide substrate (Substrate V). Our re-sults point to the conclusion that non-catalytic metal binding,especially to the inner metal binding site, is essential for AtPreP1activity.

2. Materials and methods

2.1. Multiple sequence alignment

Sequences were found using BLAST [21] and conveniently man-aged with the Biology Workbench available from San Diego Super-computer Center at http://workbench.sdsc.edu. In most cases

default parameters were applied. The amino acid sequence of At-PreP1 was used to scan protein sequences (blastp), using the BLO-SUM62 substitution matrix with a gap-opening penalty of 11 and agap extension cost of 1. The multiple sequence alignment was con-structed with ClustalW [22], using the Gonnet weight matrix withgap-opening and gap extension penalties of 10.0 and 0.10,respectively.

2.2. Generation of AtPreP1 variants

The mutations D705N, D859N, D861N for the inner metal bind-ing site and D828N, T947V, D952N for the outer metal binding sitewere introduced to a pGEX-6P-2 plasmid (Amersham Biosciences)containing wildtype AtPreP1 using the Quick-Change Site DirectedMutagenesis Kit (Stratagene). The constructs were verified by DNAsequencing utilizing the DYEnamic Sequencing Kit (Amersham Bio-sciences). All primers are listed in Supplementary Table 1.

2.3. Expression and purification of AtPreP1 and its variants

The pGEX-6P-2 vector containing wildtype and the three At-PreP1 variants N-terminally fused to Glutathione S-Transferase(GST) were used to transform E. coli BL21 (DE3). Colonies weregrown at 37 �C in 500 ml LB medium containing 100 lg/ml ampi-cillin, induced with 1 mM Isopropyl b-D-1-thiogalactopyranoside(IPTG) at OD600 of 0.5 and incubated at 30 �C to OD600 of 2. Thecells were harvested by centrifugation and stored at �80 �C. Cellswere resuspended in phosphate-buffered saline (PBS) buffer(140 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, and 1.8 mMKH2PO4, pH 7.4) and cell lysis was induced by the addition of0.5 mg/ml lysozyme followed by sonication. The cell lysate wascentrifuged for 30 min at 20 000�g to remove cell debris andthe supernatant was filtered through a 0.20 lm membrane. Thefiltered lysate was then incubated with Glutathione Sepharose�

beads (Amersham Biosciences) equilibrated with PBS buffer and

H.G. Bäckman et al. / FEBS Letters 583 (2009) 2727–2733 2729

incubated at 4 �C for 4 h with light agitation. Unbound proteinswere washed off with PBS buffer and the washed beads wereequilibrated with PreScission cleavage buffer (50 mM Tris–HCl,150 mM NaCl, 1 mM DL-dithiolthreitol). Bound AtPreP1 wascleaved off using PreScissionTM Protease in PreScission cleavagebuffer at 4 �C for 12 h and collected from the supernatant. Thepurity was analyzed on 12% SDS–PAGE in the presence of 4% ureastained with Coomassie Brilliant Blue and the protein concentra-tion was determined using BioRad Protein Assay (BioRad). Thepurified protein was stored in PreScission cleavage buffer and40% glycerol at �20 �C to preserve proteolytic activity.

2.4. Peptides used as substrates for degradation

The presequence of the F1b subunit of ATP synthase from Nico-tiana plumbaginifolia (pF1b2–54) was constructed, overexpressedand purified using the Pharmacia LKB.UV-M II FPLC system (Phar-macia Biotech) according to a procedure described previously [23].The pF1b derived peptides, pF1b1–20 and pF1b2–16, Penetratin,human prion protein peptide hPrPss1–28, Galanin, and CecropinA1–33, were kind gifts from Prof. U. Langel, Prof. A. Gräslund andProf. H. Steiner, Stockholm University. The fluorescent peptide C1was purchased from Promega, and the fluorogenic peptide derivedfrom bradykinin called Substrate V, coupled to fluorescent 7-meth-oxycoumarin (Mca) quenched by 2,4-dinitrophenol (Dnp), waspurchased from R&D Systems.

2.5. Proteolytic activity assays

Proteolytic assays with Penetratin, hPrPss1–28, Cecropin A1–33,Galanin and the F1b presequence peptides were performed in deg-radation buffer (20 mM HEPES-KOH pH 8.0); 1 lg of purified wild-type AtPreP1 or AtPreP1 variants and 1 lg of a substrate wereincubated at 30 �C for 1 h with soft agitation. MgCl2, CaCl2 or ZnCl2

were added at different concentrations as indicated in the figurelegends. Degradation was analyzed by SDS–PAGE using 15% poly-acrylamide gels and stained with Coomassie Brilliant Blue proteindye after electrophoresis. Degradation of the fluorescent C1 pep-tide was performed as above. 10 mM MgCl2 was added when indi-cated. Degradation was analyzed by electrophoresis in 1% agarosegels, after addition of 15% glycerol. Bands were visualized in UVlight.

Kinetics of proteolysis by wildtype AtPreP and AtPreP variantswere monitored using the fluorogenic peptide Substrate V (7-Methoxycoumatin-4-yl-acetyl-RPPGFSAFK-2,4-dinitrophenyl, R&DSystems). 1 lg AtPreP and AtPreP variants in 20 mM HEPES, pH8.0, with or without MgCl2 was mixed with 1 lg of Substrate Vin a final volume of 1 ml. The hydrolysis of Substrate V was mea-sured for 90 s, using the fluorometer (Fluorolog Jobin Yvon, Horibagroup) with excitation and emission wavelength set at 320 nm and405 nm, respectively.

3. Results and discussion

3.1. Divalent cations are integral to the AtPreP1 structure and appearconserved among plant species

The X-ray crystallographic structure of AtPreP1 revealed thatbeside the zinc found at the catalytic site there are two additionalmetal binding sites. Hydrated magnesium ions occupy these sites,coordinated by acidic residues located in the C-terminal portionof the enzyme. The structure of AtPreP1 forms two halves con-nected by a hinge region (Fig. 1A). Each half of the enzyme isformed by two topologically similar domains comprising a 5–7stranded b-sheet and surrounded by 7–8 helices, domain 1 and

2 in the N-terminal half and domain 3 and 4 in the C-terminalhalf. The inner magnesium-binding site, is located within the pro-teolytic chamber, less than 20 Å from the active site, and forms aconnecting bridge between domains 3 and 4 (Fig. 1B). The Mg2+

ion is coordinated by direct binding to the Asp-705 side chain lo-cated in helix 5 of domain 3, and via water molecules to residuesAsp-859 and Asp-861 from b-strand D in domain 4 (the samestrand as Tyr-854). The outer magnesium-binding site is locatedon the surface of domain 4 and is coordinated by Asp-828 fromhelix 1, Thr-947 from helix 6 and Asp-952 from helix 7 (Fig. 1C).

Sequence similarity searches using AtPreP1 against multipledatabases identified numerous homologues in bacteria andeukaryotes. After initial screening we made a multiple sequencealignment of 19 sequences to give a representative view of the con-servation of the two magnesium-binding sites identified in thecrystallographic structure (Fig. 2). Not surprisingly, all active siteresidues were identical except Glu-94, which is situated in theS10 pocket and binds the P10 side chain of the substrate (Fig. 1D).The inner magnesium-binding site appears conserved in plantPrePs and could be found in both sequences from the eudicots Vitisvinifera (grape), Populus trichocarpa (poplar) and Arabidopsis thali-ana (thale cress), as well as the sequences from the eudicot Glycinemax (soybean), the monocot Oryza sativa (rice) and the moss Physc-omitrella patens, and in one of two sequences derived from thegreen algae Micromonas and Ostreococcus lucimarinus. The outermagnesium-binding site was only fully conserved in AtPreP1, At-PreP2 and Physcomitrella PreP. Interestingly, the only non-plantprotein in which any magnesium-binding site could be foundwas the outer site found in falcilysin, the PreP homologue foundin the malaria parasite Plasmodium falciparum [24].

In order to investigate the importance of these magnesium-binding sites for catalytic activity, we produced AtPreP1 variants,in which the inner site (D1), outer site (D2) and both sites (D1,2)were rendered unable to bind metal ions. We overexpressed thesevariants as GST-fusion constructs in E. coli and purified them tohomogeneity using affinity chromatography and PrescissionPro-tease in order to cleave off GST (Fig. 1E).

3.2. Mg2+-binding sites of AtPreP1 are essential for proteolytic activity

The effect of the removal of the AtPreP inner site (D1), outer site(D2) and both metal binding sites (D1,2) on the proteolytic activitywas investigated using a native substrate, the mitochondrial prese-quence from the F1b subunit of ATP synthase, pF1b2–54 at increasingconcentrations of Mg2+, Ca2+ and Zn2+ ions (Fig. 3). No degradationof pF1b2–54 was observed for wildtype or any of the AtPreP1 vari-ants in the absence of metals. Mg2+ and Ca2+ ions stimulated cata-lytic activity of wildtype AtPreP1 and the D2 variant atconcentrations P1 mM, whereas D1 and D1,2 were only partiallyactive at increasing concentrations of Mg2+ and completely inactiveeven at 10 mM concentrations of Ca2+ (Fig. 3A and B). These resultsclearly show that the inner Mg2+-binding site is essential for the At-PreP1 activity. Complete stimulation of PreP activity was achievedbelow the physiological concentrations of free Mg2+ in both mito-chondria and chloroplats as it has been reported that free [Mg2+] inchloroplasts is between 2 and 3 mM and in mitochondria between1 and 2 mM [25]. Concentration of free Ca2+ in the organelles ismuch lower; it has been estimated to 2.4–6.3 lM in chloroplasts[26], whereas little is known about free [Ca2+] in plant mitochon-dria. In mammalian mitochondria free [Ca2+] lies between 0.05and 5 lM. However, it has been shown for plant mitochondria thatat 100 lM [Ca2+] there was no PTP (permeability transition pore)opening, swelling or cytochrome c release that was observed at0.5–2.5 mM [Ca2+] [27]. Thus, we conclude that completeactivation of PreP by Ca2+ occurs below concentrations that induceapoptosis. Addition of Zn2+ at concentrations in the range of

Fig. 2. Extract from the multiple sequence alignment of 19 PreP sequences, which displays the conservation of active site residues (top), inner Mg2+-binding site (bottom left)and outer Mg2+-binding site (bottom right). Sequence accession numbers are included within brackets following the species’ names and residue numbering is based onmature AtPreP1 (marked with an asterisk, �). Completely conserved residues are shown with black background and amino acids that match those of AtPreP1 are highlighted ingrey for comparison.

2730 H.G. Bäckman et al. / FEBS Letters 583 (2009) 2727–2733

0.01–10 mM did not stimulate degradation of the pF1b2–54 peptide.In contrast to Mg2+ and Ca2+, when Zn2+ was added to activatedAtPreP1 in the presence of 10 mM Mg2+, complete inhibition wasobserved already at 1 mM Zn2+ (Fig. 3C). Zn2+ has previously beenshown to be a potent inhibitor of varying mitochondrial enzymessuch as e.g. the respiratory cytochrome bc1 complex [28,29],matrix multienzyme complexes involved in energy productionand antioxidant defense [30] or the outer mitochondrial mem-brane pore-forming proteins [31]. Furthermore, Zn2+ also inhibitsat lM concentrations mitochondrial processing peptidase, MPP(Eriksson et al. unpublished data) that exhibits similar fold as At-PreP despite no sequence conservation [20]. On the other handboth Ca2+ and Mg2+ are important regulators of the organellar func-tions, Ca2+ is an important regulator of the mitochondrial signalingand homeostasis [32], whereas Mg2+ is an essential cofactor for theoperation of the oxygenic photosynthetic electron transfer appara-tus in chloroplasts [33]. Direct effects of Ca2+ and Mg2+ on manyenzymes have been reported. Thermolysin and the thermolysin-like proteases (TLPs) contain calcium binding sites that are impor-tant for the protein stability [34,35]. Calcium was also shown tosuppress cobalt-dependent inactivation of thermolysin [36] andprotected cultured seedlings from cadmium stress alleviating lipidperoxidation and promoting activity of antioxidative enzymes [37].

3.3. Mg2+ stimulates the activity of all AtPreP1 variants, suggesting anadditional, yet unknown role for cations during proteolysis

The proteolytic activity of the wildtype and AtPreP1 variantswas further analyzed using 9 unstructured peptides of differentlengths and amino acid compositions: the 16-residue membrane-penetrating peptide Penetratin, the 28-residue human prion-de-rived peptide hPrPss1–28, the 27-residue neuropeptide Galanin,the 33-residue antibacterial peptide Cecropin A1–33, the 53-residuepresequence from the F1b subunit of ATP synthase pF1b2–54 andshorter peptides derived from the pF1b presequence, pF1b2–16

and pF1b1–20R5G (Fig. 4) as well as a fluorescent C1 and SubstrateV peptides (Figs. 5 and 6). Sequences of all peptides used are listedin Table 1. Analysis of proteolytic activity by SDS–PAGE revealedoverall lower or complete lack of the activity in the absence ofMg2+ (Fig. 4). In the cases where partial activity in the absenceof Mg2+ was observed, such as with hPrPss1–28, Galanin and Cecr-opin A1–33, the D1 and D1,2 AtPreP variants displayed the highestdegree of inhibition. The activity of D2 did not vary from wildtypeAtPreP1 as far as we could determine by this method, and bothregained full activity by the addition of Mg2+. D1 and D1,2responded less to the addition of this ion and in the case ofpF1b2–54 and pF1b1–20 R5G the double variant D1,2 remained

wt

Δ1,2

Δ2

Δ1

MgCl2 (mM)

AtPreP

pF1β2-54

+ + + + + +

+ + + + + ++

0 0,1 1 2,5 50 10

-

wt

Δ1,2

Δ2

Δ1

wt

Δ1,2

Δ2

Δ1

CaCl2 (mM)

AtPreP

pF1β2-54

+ + + + + +

+ + + + + ++

0 0,1 1 2,5 50 10

-

ZnCl2 (mM)

AtPreP

pF1β2-54

+ + + + + +

+ + + + + ++

0 0,1 1 102,5 50

MgCl2 (mM) 0 10 10 10 10 100

-

Fig. 3. The effect of increasing concentrations of Mg2+ (A), Ca2+ (B) and Zn2+ (C) ions on the activity of wildtype AtPreP1 and AtPreP1 variants, D1, D2, and D1,2. Thedegradation was investigated using the mitochondrial presequence of the F1b subunit of ATP synthase, pF1b2–54. MgCl2, CaCl2 or ZnCl2 were added to the reaction mixture atconcentrations as indicated in the figures. ZnCl2 was added to activated AtPreP1 in the presence of 10 mM MgCl2.

MgCl2 (mM)

AtPreP Wt

0 0 100 10

-

0 1010

Wt Δ1,2

Penetratin

hPrPss

Galanin

pF1β 2-16

pF1β 1-20 R5G

pF1β 2-54

Δ2Δ1Δ1,2Δ2Δ1

0

Cecropin A

Fig. 4. Degradation of Penetratin, human prion-derived peptide hPrPss1–28, Galanin,Cecropin A1–33, the presequence of the F1b subunit of ATP synthase, pF1b2–54 andpeptides derived from the pF1b presequence, pF1b2–16 and pF1b1–20R5G by wildtypeAtPreP1 and AtPreP1 variants, D1, D2 and D1,2 in the absence and presence of Mg2+.SDS–PAGE and coomassie blue staining were used for detection of the peptides.

MgCl2 (mM)

AtPreP Wt

0 00

-

0

Δ1,2Δ2Δ1

0

MgCl2 (mM)

AtPreP Wt

10 100

-

10

Δ1,2Δ2Δ1

10

-

+

-

+

C1-peptide + + ++ +

C1-peptide + + ++ +

Fig. 5. Degradation of a fluorescent peptide C1 by wildtype AtPreP1 and AtPreP1variants, D1, D2, and D1,2 in the absence (A) and presence (B) of Mg2+. Thefluorescent peptide was detected on agarose gels under UV light.

H.G. Bäckman et al. / FEBS Letters 583 (2009) 2727–2733 2731

Fig. 6. Kinetics of degradation of the fluorogenic Substrate V by wildtype AtPreP1 (A) and AtPreP1 variants, D1 (B), D2 (C) and D1,2 (D) in the absence and presence of Mg2+.The change in fluorescence quenching was measured with excitation and emission wavelength set at 320 nm and 405 nm, respectively.

Table 1Peptides used in the experiments.

Peptide aa pI Sequence

Substrate V 9 11,0 Mca-RPPGFSAFK-Dnp-OHC1 11 11,0 Dye-PLSRTLSVAAKPenetratin 16 12,3 RQIKIWFQNRRMKWKKGalanin 27 6,74 TLNSAGYLLGPHAIDNHRSFSDKHGLThPrPss 28 9,38 MANLGCWMLVLFVATWSDLGLCKKRPKPCecropin1–33 33 9,82 MNFYNIFVFVALILAITIGQSEAGWLKKIGKKIPre-F1b2–54 54 12,5 ASRRLLASLLRQSAQRGGGLISRSLGNSPKSASRASSRASPKGFLLNRAVQYMPre- F1b2–16 15 12,3 ASRRLLASLLRQSAQPre-F1b1–20R5G 20 12,3 MASRRLLASLLRQSAQRGGG

2732 H.G. Bäckman et al. / FEBS Letters 583 (2009) 2727–2733

completely inactive even at 10 mM MgCl2 (Fig. 4). AtPreP was pre-viously shown to cleave unstructured peptides in the range of10–65 amino acids. Moreover, it exhibits cleavage preference fora basic amino acid in position P10 and small, uncharged aminoacids or serine residue in the P1 position [14]. Also structural ele-ments in the substrate peptide were crucial for degradation asintroduction of proline, a helix breaker, instead of leucine in thepF1b1–20 peptide resulted in reduction of the activity. Here, wehave used peptides from 9 to 54 amino acids long. All the peptidestested were partially or completely degraded by wildtype AtPrePand its D1 and D2 variants in the presence of Mg2+, however short-er peptides derived from the presequence of the pF1b subunit ofthe ATP synthase, were more prone to proteolysis than the full-length pF1b presequence and other longer peptides. This suggeststhat positioning of a longer peptide in the catalytic chamber of At-PreP affects the enzyme catalysis. All the peptides tested containedpositively charged amino acids, and even if the isoelectric pointsfor the peptides differ from 6.74 to 12.48, no differences in thedegradation measured by SDS–PAGE could be related to the overallpeptide charge.

We also studied the degradation of the 11 amino acid residuefluorescent C1 peptide (Fig. 5) and the kinetics of proteolysisusing a fluorogenic peptide derived from bradykinin, called Sub-strate V (Fig. 6). The fluorescent C1 peptide is positively chargedand upon cleavage fragments of different sizes and charges aregenerated that correspond to different cleavage sites. Incubationof wildtype AtPreP1 and AtPreP1 variants, D1 and D2, with theC1 peptide both in the absence and presence of 10 mM Mg2+

produced negatively charged fragments moving towards thepositively charged electrode, indicating cleavage of the C1 peptideafter Pro, Leu or Ser (Fig. 5A). In contrast, the double D1,2 variantwas only partially active in the absence of Mg2+, producing onlyabout 10% of negatively charged fragments while the rest of theC1 peptide remained intact and moved towards the negativelycharged electrode. Interestingly, the activity of D1,2 AtPreP1was completely restored in the presence of Mg2+ (Fig. 5B). Finally,we performed kinetic studies using Substrate V, a fluorogenicpeptide containing 9 amino acids residues. This peptide is modi-fied at its N-terminus with fluorescent 7-methoxycoumarin (Mca)and at its C-terminus by 2,4-dinitrophenol (Dnp). Fluorescence is

H.G. Bäckman et al. / FEBS Letters 583 (2009) 2727–2733 2733

quenched in the native state and increases as the peptide iscleaved. The initial phase of degradation was followed during90 seconds in the absence and presence of Mg2+ (Fig. 6). For allAtPreP1 variants there was an increase of the activity in the pres-ence of Mg2+. The reason for the Mg2+ activation is not known butbeside the importance of the inner magnesium-binding site wecan speculate that since AtPreP1 peptidasome is very acidic andthe catalytic chamber consisting of the enzyme halves is nega-tively charged, Mg2+ may stabilize the protein by neutralizingthe surface of the chamber and enable the two halves to closeand complete the active site. As the requirement of Mg2+ is stron-ger for longer peptides, it is possible that the Mg2+-binding sitesand especially the inner site between the domain 3 and 4 stabi-lizes the enzyme and contributes to the conformational changesrequired for catalysis. The other possibility is that Mg2+ promotessubstrate unfolding making its conformation more easily degrad-able by the protease. We have recently showed that despite thefact that metal binding sites are not conserved in human PreP(hPreP), the enzyme is slightly stimulated by divalent cations.Most interestingly, single nucleotide polymorphism (SNP) hPrePvariants that show reduced enzymatic activity can be completelyreactivated by addition of Mg2+. This finding may indicate thatMg2+ binding to PreP functions as a ‘‘rescue” mechanism underconditions when the PreP activity is reduced to ascertain effectiveremoval of toxic peptides from the organelles.

Acknowledgments

We would like to thank Dr. Shashi Bhushan, Dr. Stefan Nilsson,Nyosha Alikhani and Catarina Pinho for technical assistance in theinitial phase of the project. This work was supported by grantsfrom the Swedish Research Council to E.G. and T.E.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.febslet.2009.07.040.

References

[1] Bolender, N., Sickmann, A., Wagner, R., Meisinger, C. and Pfanner, N. (2008)Multiple pathways for sorting mitochondrial precursor proteins. EMBO Rep. 9,42–49.

[2] Lister, R., Hulett, J.M., Lithgow, T. and Whelan, J. (2005) Protein import intomitochondria: origins and functions today (review). Mol. Membr. Biol. 22, 87–100.

[3] Neupert, W. and Herrmann, J.M. (2007) Translocation of proteins intomitochondria. Annu. Rev. Biochem. 76, 723–749.

[4] Benz, J.P., Soll, J. and Bolter, B. (2009) Protein transport in organelles: thecomposition, function and regulation of the Tic complex in chloroplast proteinimport. Febs J. 276, 1166–1176.

[5] Jarvis, P. (2008) Targeting of nucleus-encoded proteins to chloroplasts inplants. New Phytol. 179, 257–285.

[6] Hawlitschek, G., Schneider, H., Schmidt, B., Tropschug, M., Hartl, F.U. andNeupert, W. (1988) Mitochondrial protein import: identification of processingpeptidase and of PEP, a processing enhancing protein. Cell 53, 795–806.

[7] Glaser, E. and Dessi, P. (1999) Integration of the mitochondrial-processingpeptidase into the cytochrome bc1 complex in plants. J. Bioenerg. Biomembr.31, 259–274.

[8] Richter, S. and Lamppa, G.K. (1998) A chloroplast processing enzyme functionsas the general stromal processing peptidase. Proc. Natl. Acad. Sci. USA 95,7463–7468.

[9] Nicolay, K., Laterveer, F.D. and van Heerde, W.L. (1994) Effects of amphipathicpeptides, including presequences, on the functional integrity of rat livermitochondrial membranes. J. Bioenerg. Biomembr. 26, 327–334.

[10] Hugosson, M., Andreu, D., Boman, H.G. and Glaser, E. (1994) Antibacterialpeptides and mitochondrial presequences affect mitochondrial coupling,respiration and protein import. Eur. J. Biochem. 223, 1027–1033.

[11] Stahl, A., Moberg, P., Ytterberg, J., Panfilov, O., Brockenhuus Von Lowenhielm,H., Nilsson, F. and Glaser, E. (2002) Isolation and identification of a novelmitochondrial metalloprotease (PreP) that degrades targeting presequences inplants. J. Biol. Chem. 277, 41931–41939.

[12] Bhushan, S., Lefebvre, B., Stahl, A., Wright, S.J., Bruce, B.D., Boutry, M. andGlaser, E. (2003) Dual targeting and function of a protease in mitochondriaand chloroplasts. EMBO Rep. 4, 1073–1078.

[13] Moberg, P., Stahl, A., Bhushan, S., Wright, S.J., Eriksson, A., Bruce, B.D. andGlaser, E. (2003) Characterization of a novel zinc metalloprotease involved indegrading targeting peptides in mitochondria and chloroplasts. Plant J. 36,616–628.

[14] Stahl, A. et al. (2005) Two novel targeting peptide degrading proteases, PrePs,in mitochondria and chloroplasts, so similar and still different. J. Mol. Biol.349, 847–860.

[15] Rawlings, N.D., Tolle, D.P. and Barrett, A.J. (2004) MEROPS: the peptidasedatabase. Nucleic. Acids Res. 32, D160–D164.

[16] Becker, A.B. and Roth, R.A. (1992) An unusual active site identified in a familyof zinc metalloendopeptidases. Proc. Natl. Acad. Sci. USA 89, 3835–3839.

[17] Kambacheld, M., Augustin, S., Tatsuta, T., Muller, S. and Langer, T. (2005) Roleof the novel metallopeptidase Mop112 and saccharolysin for the completedegradation of proteins residing in different subcompartments ofmitochondria. J. Biol. Chem. 280, 20132–20139.

[18] Falkevall, A. et al. (2006) Degradation of the amyloid beta-protein by the novelmitochondrial peptidasome, PreP. J. Biol. Chem. 281, 29096–29104.

[19] Anandatheerthavarada, H.K. and Devi, L. (2007) Amyloid precursor proteinand mitochondrial dysfunction in Alzheimer’s disease. Neuroscientist 13,626–638.

[20] Johnson, K.A., Bhushan, S., Stahl, A., Hallberg, B.M., Frohn, A., Glaser, E. andEneqvist, T. (2006) The closed structure of presequence protease PreP forms aunique 10,000 Angstroms3 chamber for proteolysis. Embo J. 25, 1977–1986.

[21] Altschul, S.F., Madden, T.L., Schaffer, A.A., Zhang, J., Zhang, Z., Miller, W. andLipman, D.J. (1997) Gapped BLAST and PSI-BLAST: a new generation of proteindatabase search programs. Nucleic. Acids Res. 25, 3389–3402.

[22] Thompson, J.D., Higgins, D.G. and Gibson, T.J. (1994) CLUSTAL W: improvingthe sensitivity of progressive multiple sequence alignment through sequenceweighting, position-specific gap penalties and weight matrix choice. Nucleic.Acids Res. 22, 4673–4680.

[23] Pavlov, P.F., Moberg, P., Zhang, X.P. and Glaser, E. (1999) Chemical cleavage ofthe overexpressed mitochondrial F1beta precursor with CNBr: a new strategyto construct an import-competent preprotein. Biochem. J. 341 (Pt 1),95–103.

[24] Murata, C.E. and Goldberg, D.E. (2003) Plasmodium falciparum falcilysin: ametalloprotease with dual specificity. J. Biol. Chem. 278, 38022–38028.

[25] Igamberdiev, A.U. and Kleczkowski, L.A. (2001) Implications of adenylatekinase-governed equilibrium of adenylates on contents of free magnesium inplant cells and compartments. Biochem. J. 360, 225–231.

[26] Kreimer, G., Melkonian, M., Holtum, J.A. and Latzko, E. (1988) Stromal freecalcium concentration and light-mediated activation of chloroplast fructose-1,6-bisphosphatase. Plant Physiol. 86, 423–428.

[27] Virolainen, E., Blokhina, O. and Fagerstedt, K. (2002) Ca(2+)-induced highamplitude swelling and cytochrome c release from wheat (Triticum aestivumL.) mitochondria under anoxic stress. Ann. Bot. (Lond) 90, 509–516.

[28] Link, T.A. and von Jagow, G. (1995) Zinc ions inhibit the QP center of bovineheart mitochondrial bc1 complex by blocking a protonatable group. J. Biol.Chem. 270, 25001–25006.

[29] Lorusso, M., Cocco, T., Sardanelli, A.M., Minuto, M., Bonomi, F. and Papa, S.(1991) Interaction of Zn2+ with the bovine-heart mitochondrial bc1 complex.Eur. J. Biochem. 197, 555–561.

[30] Gazaryan, I.G., Krasinskaya, I.P., Kristal, B.S. and Brown, A.M. (2007) Zincirreversibly damages major enzymes of energy production and antioxidantdefense prior to mitochondrial permeability transition. J. Biol. Chem. 282,24373–24380.

[31] Bonanni, L. et al. (2006) Zinc-dependent multi-conductance channel activity inmitochondria isolated from ischemic brain. J. Neurosci. 26, 6851–6862.

[32] Bernardi, P. (1999) Mitochondrial transport of cations: channels, exchangers,and permeability transition. Physiol. Rev. 79, 1127–1155.

[33] Shcolnick, S. and Keren, N. (2006) Metal homeostasis in cyanobacteria andchloroplasts. Balancing benefits and risks to the photosynthetic apparatus.Plant Physiol. 141, 805–810.

[34] Veltman, O.R., Vriend, G., Berendsen, H.J., Van den Burg, B., Venema, G. andEijsink, V.G. (1998) A single calcium binding site is crucial for the calcium-dependent thermal stability of thermolysin-like proteases. Biochemistry 37,5312–5319.

[35] Takita, T., Aono, T., Sakurama, H., Itoh, T., Wada, T., Minoda, M., Yasukawa, K.and Inouye, K. (2008) Effects of introducing negative charges into themolecular surface of thermolysin by site-directed mutagenesis on itsactivity and stability. Biochim. Biophys. Acta 1784, 481–488.

[36] Hashida, Y. and Inouye, K. (2007) Kinetic analysis of the activation-and-inhibitiondual effects of cobalt ion on thermolysin activity. J. Biochem. 141, 843–853.

[37] Wang, C.Q. and Song, H. (2009) Calcium protects Trifolium repens L. seedlingsagainst cadmium stress. Plant Cell Rep.