Gamma carbonic anhydrase like complex interact with plant mitochondrial complex I

Plant Molecular Biology 56: 947–957, 2004. Ó 2005 Springer. Printed in the Netherlands.

947

Gamma carbonic anhydrase like complex interact with plant mitochondrial complex I Mariano Perales1;y , Gustavo Parisi2;y , Mar
a Silvina Fornasari2 , Alejandro Colaneri1 , Fernando Villarreal1 , Nahuel Gonza´lez-Schain1 , Julia´n Echave2 , Diego Go´mez-Casati1 , Hans-Peter Braun3 , Alejandro Araya4 and Eduardo Zabaleta1;5; * 1

Instituto de Investigaciones Biotecnolo´gicas, IIB-INTECH (CONICET/UNSAM), C.C. 164, 7130 Chascom
us, Argentina; 2 Centro de Estudios de Investigaciones, Universidad Nacional de Quilmes, Roque Sa´enz Pen˜a 180, B1876BXD Bernal., Argentina; 3 Institut fu¨r Angewandte Genetik, Universita¨t Hannover, Herrenha¨user Str. 2, D-30419 Hannover, Germany; 4 UMR 5097 R.E.G.E.R., CNRS and Universite´ Victor Segalen Bordeaux 2, 146 rue Le´o Saignat, 33076 Bordeaux, France; 5 Instituto de Investigaciones Biolo´gicas. Universidad Nacional de Mar del Plata, Funes 3250, 7600 Mar del Plata, Argentina (*author for correspondence; e-mail [email protected]); y These authors have contributed equally to the article.

Received 31 July 2004; accepted in revised form 16 November 2004

Key words: Complex I, Gamma carbonic anhydrase, plant mitochondria, respiratory chain

Abstract We report the identification by two hybrid screens of two novel similar proteins, called Arabidopsis thaliana gamma carbonic anhydrase like1 and 2 (AtcCAL1 and AtcCAL2), that interact specifically with putative Arabidopsis thaliana gamma Carbonic Anhydrase (AtcCA) proteins in plant mitochondria. The interaction region that was located in the N-terminal 150 amino acids of mature AtcCA and AtcCA like proteins represents a new interaction domain. In vitro experiments indicate that these proteins are imported into mitochondria and are associated with mitochondrial complex I as AtcCAs. All plant species analyzed contain both AtcCA and AtcCAL sequences indicating that these genes were conserved throughout plant evolution. Structural modeling of AtcCAL sequences show a deviation of functionally important active site residues with respect to cCAs but could form active interfaces in the interaction with AtcCAs. We postulate a CA complex tightly associated to plant mitochondrial complex. Introduction Enzymes containing the left-handed parallel beta helix (LbH) domain display imperfect tandem-repeated copies of a hexapeptide sequence characterized as [LIV]-[GAED]-X2-[STAV]-X and are termed ‘hexapeptide repeat’ enzymes (Vaara, 1992). It is a very diverse group of proteins involved in many biological processes, including coenzymes, many types of transferases, carrier proteins, and gamma carbonic anhydrase/ferripyochelin binding protein family (Parisi et al., 2000). All carbonic anhydrases are divided into three classes (a, b, and c) that evolved independently (Hewett-Emmett and Tashian 1996, Hewett-Emmett,

2000). To date, there is only one example fully characterized of cCA from Methanosarcina thermophila, named CAM (Alber and Ferry, 1994, Kisker et al., 1996). A putative cCA homologue, CcmM, from Synechococcus PCC7942 without CA activity has been described (Price et al., 1993). The protein is a subunit of the carboxysomal complex, a key compartment in the CO2 concentrating mechanism (Badger and Price, 2003). Recently, very similar proteins from E. coli (i.e. CaiE, etc) were described, which are postulated to work in coordination with multienzymatic complexes as coenzymes (Merlin et al., 2003). A sequence pattern characteristic of cCA was recently defined, which is useful to identify this kind of enzyme

948 among the large LbH protein family (Parisi et al., 2004a). We previously have identified three putative mitochondrial cCA protein homologues in Arabidopsis thaliana, the first example for this type of enzyme in eukaryotes. These putative AtcCAs are imported into mitochondria and presumably associated to respiratory chain complex I (Parisi et al., 2004a). Here, we report on the identification of two novel nuclear genes from Arabidopsis thaliana whose products interact in vivo with AtcCAs. These two novel proteins, named Arabidopsis thaliana gamma Carbonic Anhydrase Like (AtcCALs) also show high similarity and sequence-structure compatibility with prokaryotic and plant c class carbonic anhydrases (cCAs). Phylogenetic trees indicate that cCALs derive from cCAs very early in the evolution since both types of proteins are present in green algae and primitive plants. Furthermore, experimental evidence is presented indicating that these gene products possess their own mitochondrial transit peptides. The AtcCAL and AtcCA proteins share the same suborganelar location and interact specifically between themselves suggesting that cCA holoenzyme are more complex in photosynthetic eukaryotes than in other organisms.

Materials and methods

TACG for N2, ACCTCGAGTGGT-TGC TGCTGGT for N3, ACTCTCGAGTCTC-TCA ACCTCG for C1, ACTCTCGAGACCTCT CCAGAAG for C2, ACTCTCGAGTGTTACAT TGTCC for C3 and GACTCGAGTT-AGAAG TACTGAG as reverse primer. Restriction sites (underlined) were included to facilitate the cloning of fragments into the appropriate vectors (pGBKT7 and pGADT7). The yeast strain used for these assays was PJ69-4A (Clontech), genotype: MATa; trp1-901; leu 2-3,112; ade2-101; ura 3-52; his 3-200; gal44512; gal804538; GAL2-2ADE; LYS2::GAL1-HIS3; met2::GAL7lacZ. Yeast transformation was performed by using the lithium acetate method (Gietz et al., 1992). Analysis of b-galactosidase activity were performed as described (Busi et al., 2003). Similarity searches Using the sequence At3g48680 as input we used the program BLASTP to search the protein nonredundant database at NCBI and the program TBLASTN to search the six reading frames of the EST database (Viridiplantae). Sequences with an E-value below 1
104 from both databases were retrieved and combined in one dataset. Sequences were aligned using the program CLUSTALW (Thompson et al., 1997). In the case of the ESTs, previous to being aligned, the retrieved sequences were filtered and processed as described (Cuff et al., 2000).

Yeast two hybrid analyses Phylogenetic analysis Yeast two-hybrid screens and analyses (Gietz et al., 1992) were performed with cHybriZAP 2.1 system (Stratagene, CA). The bait construct was the complete AtcCA2 sequence in the vector pBDGal4Cam (Stratagene). Yeast strain YM190::pBDcCA2 was transformed with an Arabidopsis flower cDNAs library (1
106 pfu/ lg). Yeast colonies were selected on auxotrophy media (-leu, -trp, -his dropout, Clontech, CA, Palo Alto) with 40 mM 3 Aminotrioazole (3-AT). Y190 genotype: MATa; trp1-901; leu 2-3,112;ade2-101; ura 3-52; his 3-200; gal44512; gal804538; URA3::GAL=lacZ; LYS2::GAL(UAS)=HIS2. Positive clones were tested for LacZ activity on filters. For deletions of AtcCA2 the following primers were used: TCCTCGAGTGATGAATG TGTTT for N1, TACTCGAGTCCAAGATAA-

A maximum parsimony topology was obtained using the program PROTPARS from the package PHYLIP (Felsestein, 1993). Bootstrapping support was assessed using 500 replicates obtained with SEQBOOT and analyzed with CONSENSE. Both programs are also from the PHYLIP package. Protein analysis and immunoblots Protein extracts from mitochondria of Arabidopis thaliana cell culture were performed following standard techniques (Eubel et al., 2003). Recombinant proteins were isolated following pGEX (AmershamPharmaciaBiotech, Piscataway, NJ) or pET24a (Novagen, Madison, WI) protocols. The

949 GST moiety was cleaved off and the cCA moiety further purified. Samples were subjected to standard SDS-PAGE or two dimension BN (Blue Native) followed by a SDS-PAGE (Scha¨gger and Pfeiffer, 2000) and blotted onto Nitrocellulose membranes (BioRad, Hercules, CA). Blots were incubated with anti-cCA2 polyclonal antiserum and revealed following standard techniques (Sambrook and Russell, 2001). Mitochondria isolation and Import assays Mitochondria were prepared from potato tubers (Solanum tuberosum) as described (Parisi et al., 2004a). 35 [S] Methionine-35 [S] Cysteine labeled proteins were synthesized using TnT Coupled Reticulocyte Lysate System (Promega, Madison, WI). Protein imports into mitochondria were performed as described (Parisi et al., 2004a). Model building and structural analysis

corresponds to a gene that encodes an unknown small protein we named CA Interacting Protein 14 kDa (CIP14) (At4g35320). Two other fulllength cDNA clones were selected corresponding to genes At5g63510 and At3g48680 with unknown function, but annotated as transferases belonging to the hexapeptide repeat protein family. The predicted amino acid sequences from both genes have a potential N-terminal signal peptide for targeting to mitochondria. Interestingly, both proteins show 87% similarity and 61–64% identity to bona fide AtcCAs. They contain the PaaY domain (COG0663) as found in carbonic anhydrase/ acetyltransferase/isoleucine patch superfamily. Proteins belonging to this family fold into a lefthanded parallel b-helix (Raetz and Roderick, 1995) (LbH) as do CAM (Alber and Ferry, 1994) and AtcCAs (Parisi et al., 2004a), however, both proteins lack some catalytically important residues proposed to be essential for CA activity. These two proteins were accordingly re-named AtcCAL1 and AtcCAL2 (for gamma Carbonic Anhydrase Like).

Tridimensional structures of At3g48680 were built using the program MODELLER 6 (Sali and Blundell, 1993) on the TITO server (http : == bioserv:cbs:cnrs:fr=)(Database size: 9864 structures). The structural template used was the structure of the Ferripyochelin Binding Protein from Pyrococcus horikoshii (PDB 1v67) which produced the best score in threading analysis with the server 3D-PSSM for fold recognition (Kelley et al., 2000). The secondary structure of At3g48680 and At5g63510 was estimated using SOPM (Combet et al., 2000) and PSIPRED (McGuffin et al., 2000) programs.

Results Two LbH proteins interact in vivo with AtcCAs To investigate the metabolic pathway in which AtcCAs could be embedded, a two-hybrid screen was performed using the entire coding region of AtcCA2 (Parisi et al., 2004a) fused to the Gal4BD as a bait. More than five millions independent clones were screened for his auxotrophy and LacZ activity. After several rounds of screening, three different positive clones that interact stably with AtcCA2 were isolated (Figure 1). One of them

Figure 1. AtcCA2 interacts with three novel Arabidopsis proteins. The Y190 yeast strain containing the plasmid pBDGal4cCA2 fusion was grown on liquid culture lacking tryptophan and leucine. It was transformed with an Arabidopsis library in khybridZAP2.1 (Stratagene). Transformants were plated on selective medium lacking tryptophan, leucine and histidine, containing 40 mM 3-aminotriozole (-HLT 40 mM 3-AT) and b-galactosidase activity was measured on his positive clones. As negative control yeast cells harbouring the pGal4BD were used for both assays. Different positive clones of yeast containing cCA2 and interacting proteins were expressed in yeast cells. The effect of the different constructs on growth was monitored. b-galactosidase activity in yeast cells harbouring the same constructs as above was assayed by a simple filter method.

950 The cCAs and these novel cCALs were previously identified to co-migrate with mitochondrial complex I on Blue Native gels by Heazlewood et al. (2003) as plant-specific component of this complex. AtcCALs are imported into mitochondria and co-migrate mainly with complex I Database searches revealed the presence of AtcCALs in several other plants and green algae. All these proteins are predicted to have mitochondrial localization when analyzed with the PSORT program. A presequence 40 amino acids in length was predicted for AtcCALs. This and the fact that they interact with AtcCA2, which is considered a mitochondrial protein (Eubel et al., 2003; Heazlewood et al., 2003; Parisi et al., 2004a) prompted us to investigate the subcellular localization of these proteins in more detail. In vitro labeled AtcCAL proteins were incubated with isolated mitochondria. As shown in the Figure 2, AtcCAL1 and 2 are efficiently imported into potato mitochondria resulting in a protein with a reduced size of about 5 kDa. This indicates that AtcCALs are imported into mitochondria by a classical mechanism that includes the removal of a signal peptide. The suborganellar localization of these novel LbH proteins was studied by an immunoblot of an 2D Blue Native/SDS gel (Kruft et al., 2001; Sch€ agger, 2001; Eubel et al., 2003, Heazlewood et al., 2003) using a polyclonal antibody directed against AtcCA2 which recognizes all AtcCA and AtcCAL proteins and an unknown protein of around 70 kDa (Figure 3A). All cCA and cCAL proteins are localized in the same vertical row as the subunits forming part of singular complex I of the respiratory chain and the complex I containing

supercomplexes I + III2 and I2 + III4 . However, these proteins are also detected with weaker intensity in the lower molecular weight region of the 2D gels, (Figure 3b and data not shown). These results confirm previous data found by subunit identification of complex I using mass spectrometry (Heazlewood et al., 2003). The hexapeptide repeat domain is essential for interaction Yeast strain PJ69-4A was tested for survival on selective medium and Lac Z activity in vitro after co-transformation with plasmids carrying different deletions of AtcCA2 fused to Gal4 binding domain (BD) and AtcCAL2 containing the activation domain (AD) of Gal4 (see M and M). Deletions spanning amino acid residues 40 to 180 of AtcCA2 completely abolished interaction with AtcCALs indicating that the N-terminal half of the predicted mature protein containing the hexapeptide repeats is required for stable interaction (Figure 4A). To evaluate the interaction between cCA2 and cCAL proteins, different combination of AD and BD constructs were tested. cCAL proteins were not able to form homo- or cCAL heterodimers (cCAL1/cCAL2 heterodimers). In contrast, cCA proteins do form homodimers (data not shown). Interestingly, interaction between cCAL2 and cCA2 is stronger than between cCA2 alone or with cCAL1, suggesting certain specificity in the interaction (Figure 4B). Divergence of AtcCAs and AtcCALs is an ancient event Since AtcCALs showed some homology to the bait protein (AtcCA2), we decided to study the

Figure 2. AtcCALs are imported into plant mitochondria by a classical mechanism. Import of AtcCAL1 and AtycCAL2 into mitochondria. In vitro translated AtcCAL1 and AtcCAL2 were incubated with purified potato mitochondria. Lanes 1 and 5 shows the precursor proteins without incubation. After incubation, samples were loaded onto gel without treatment (lanes 2 and 6) or pretreated with proteinase K (lanes 3 and 7) or triton X-100 plus proteinase K (lanes 4 and 8).

951

Figure 3. AtcCALs, together with AtcCAs co-migrate mainly with mitochondrial complex I. Isolated mitochondria were treated with digitonin as described in (Eubel et al., 2003) and protein complexes were subsequently resolved by one or two-dimensional Blue Native/ SDS PAGE. These gels were transferred to nitrocellulose membrane and incubated with a polyclonal antibody anti-cCA2. The membrane was subsequently incubated with a secondary anti-antibody conjugated with alkaline phosphatase and revealed with NBT and BCiP. (A) 12.5% SDS gel of a mitochondrial extract (left) probed with anti-cCA2 (right). Arrows on the left indicate reacting bands. (B) 1D Blue-Native (top) and 2D Blue-Native/SDS gels (middle) probed with anti-cCA2 (bottom). Designations of the protein complexes are given above the gels. (C) A blot of a 15% SDS gel separating total protein extract of E. coli BL21 without IPTG (lane1) or with 10 mM IPTG and 3 h induction of cCA2 (lane2), cCAL1 (lane 3), cCAL2 (lane 4) and cCA3 (lane 5) subcloned in pET vector. These result confirm that only proteins with expected molecular weight are recognized by the antiserum. Molecular weight markers are shown. CB: Coomasie Blue and WB: western blot.

952

Figure 4. N-terminal domain of AtcCAs is essential for interaction with AtcCALs. The PJ69-4A yeast strain containing the plasmids listed were grown on liquid culture lacking tryptophan and leucine. Serial dilutions of 105 –102 cells were plated on non-selective medium lacking tryptophan and leucine (LT) or on selective medium lacking tryptophan, leucine and histidine containing 10 mM 3-aminotriozole (HLT 10 mM 3-AT) and b-galactosidase activity was measured. As negative control yeast cells harbouring the pGal4BD were used for both assays. (A) Left panel: Different N and C-terminal deletions of cCA2 and cCAL2 were expressed in yeast cells. The effect of the different constructs on growth was monitored. Right panel: b-galactosidase activity in yeast cells harbouring the same constructs as above. (B) Left panel: Different combination of AtcCAs and AtcCALs were expressed in yeast cells. The effect of the different constructs on growth was monitored. Right panel: b-galactosidase activity in yeast cells harbouring the same constructs as above was calculated according to Miller (1992). The graphic was generated with Excel software (Microsoft).

evolutionary origin of these proteins to gain insight about their probable physiological roles. Using At3g48680 as starting sequence, similarity searches were performed on EST (Viridiplantae) databases. We found 1249 EST sequences, whose deduced amino acid sequences show the hexapeptide-repeat motif with different degrees of conservation. The non-redundant coding sequences retrieved from EST searches were supplemented with the protein sequences from protein–protein searches in such a way as to produce an overall non-redundant dataset. Seventy-three sequences were found, all of them belong to the LbH fold family. They were aligned with ClustalW and maximum parsimony analysis was performed. In the Figure 5, the phylogenetic tree obtained shows clearly that relatively similar genes cluster in two main clades (bootstrap 96%), one containing AtcCAs (A in the figure) and other containing AtcCALs (B in the figure). Sequences in the Clade A contain all catalytically important residues predicted for a cCA in plants (Parisi et al., 2004a). This includes three His residues, essential to coordinate a Zn ion, and residues Arg 86, Asp 88, Gln 101, Asp 102 and Tyr 207 following AtcCA1

annotation, (Parisi et al., 2004a). Sequences in the clade B only one His is fully conserved, other is replaced by an Ala residue and the last one is replaced by an Arg residue. The latter could represent a conservative replacement in binding a Zn ion (see below) (Ferraroni et al., 2002). Interestingly, other important residues are properly arranged into the putative active site that includes Arg 103, Asp 105, Gln 118, Glu 119 (conservative replacement) and Tyr 231 (following cCAL2 annotation). Whereas cCAs are present in all prokaryotes and photosynthetic eukaryotes (Smith et al., 1999, Parisi et al., 2004a), cCALs are not present in prokaryotes. However, in all plant and green algae analyzed, there is at least one representative of each clade (Figure 5), indicating that duplication and divergence of these genes might be occurred before or when green algae arose. AtcCALs could form a CA active site only interacting with cCAs Deduced amino acid sequences of cCALs from Arabidopsis were subjected to protein modeling

953

Figure 5. cCAs and cCALs are present in all plants and green algae analyzed. Maximum parsimony tree inferred for cCAs (clade A) and cCALs (clade B) from plants and green algae using all the proteins found EST database. All the proteins in clade A contain the His residues important to bind the Zn atom. The proteins in clade B contain only one of the His mentioned above. Cluster A and B contains paralogous sequences probably originated early in the evolution of plants. All photosynthetic organisms contain representative of both gene types. Arabidopsis sequences are highlighted. The support for the node separating both clusters is 96 %. Boostrap values are indicated by black triangles (60–80%) or black circles (80–100%).

using the Modeller 6a program (Sali and Blundell, 1993). A structural template, the ferripyochelin binding protein of Pyrococcus horikoshii Ot3 (PhFBP, synonym carbonic anhydrase) was found with an E-value of 1.09e05, using threading searches with 3DPSSM server. The 1v67 structure co-crystallized with Zn, -HCO3 and Ca (Iverson et al., 2000; Tripp and Ferry, 2000; Tu et al., 2002) represents a homotrimeric protein where each subunit adopts the left-handed b helix fold as CAM (Iverson et al., 2000) and AtcCAs (Parisi et al., 2004a). The deduced structure of the cCALs fits perfectly to PhFBP (as do CAM and AtcCAs, data not shown), with an a-helix protruding from the left-handed centre to the C-terminus predicted by using the SOPM (Combet et al., 2000) and PSIRED (McGuffin et al., 2000) servers. As with

CAM, PhFBP and AtcCAs, the AtcCALs sequences contain six coils separated by loops or turns. The calculated root means square deviation (r.m.s.d.) value between the models and the template structure is 2.3 A indicating a significant similarity. It has been previously suggested that LbH proteins could form stable heterooligomers (Parisi et al., 2004b). Models as heterotrimers with cCAs were built. Three different interfaces could be formed using different combinations of the two distinct subunits (Figure 6A and B). The first interface has been previously described (Parisi et al., 2004a). The second interface comprises the His 130 of cCAs (red in the figure) and the His 124 and Arg 152 (conservative replacement) to bind a zinc ion. The putative third interface (Figure 6B)

954

Figure 6. Putative CA active interfaces in heterotrimers cCA–cCAL. Left panel. Diagram of a heterotrimer. Arrows indicate the interface presented in the corresponding right panel. Two different interfaces and the main residues involved in the enzyme mechanism are shown. The ball in the center of the figure represents the Zn atom. Residues Arg 152 and Gln 164 coordinating Zn atom are shown. cCA subunit is shown in red whereas cCAL subunit is shown in blue. Residues Asp 105, Arg 103, Gln 118 and Glu 119.participating in the mechanism are also shown.

comprises two His residues of cCAs (His 107 and 135) and the Gln 164 of cCALs. The equivalent position to His 130 in cCAs is 100substituted by Ala (Ala 147) in cCALs. The Gln 164 is conserved in all proteins belonging to B cluster (Figure 5). As in case of Arg 152 mentioned above, Gln was also described as a conservative replacement of His in various carbonic anhydrases to coordinate a putative Zn ligand (Lesburg et al., 1997). Following the predicted reaction mechanism of CAM (Iverson et al., 2000; Tripp and Ferry, 2000; Tu et al., 2002; Tripp et al., 2004) and PhFBP, Arg 59, Asp 61, Gln 75 and Asp 76 (CAM residues) are important residues. Asp 76 is replaced by Glu 119 in cCALs which is a conservative replacement. Thus, all these residues are conserved and properly arranged in the active site cavity of

the models. Alternative amino acids that might serve as proton transfer residues, Tyr 207 (AtcCAs residues) and Tyr 159 (PhFBP residues) have been postulated (Iverson et al. 2000a; Tripp and Ferry, 2000; Tu et al., 2002 Parisi et al., 2004a;). It is interesting to note that these residues are also conserved in AtcCALs (Tyr 231). These results, together with two hybrid analyses, suggest that heteromers cCA/cCAL could be present possibly forming a CA complex mainly associated to complex I in plants.

Discussion In this report, two genes coding for LbH proteins imported into mitochondria have been identified

955 by two hybrid analyses. These proteins interact with AtcCAs, were found to be tightly associated mainly to mitochondrial complex I and show high similarity in sequence and structure with AtcCAs. Although AtcCALs lack two essential His residues to act as carbonic anhydrase, they could form putative active interfaces only interacting with AtcCAs. One of these putative active interfaces requires that Arg 152 that replaces a His residue could coordinate a Zn ion, a situation that has been described for variants of human carbonic anhydrase I (Ferraroni et al., 2002). Indeed, one exception is found in the sequence from Vitis aestivalis where the VacCAL has a His residue replacing Arg 152, suggesting that His/Arg replacement is conservative. A second putative active interface contains the Gln 164, a residue that has been postulated to coordinate Zn ions (Lesburg et al., 1997; McCall and Fierke, 2004). In monocotyledoneous plants some cCAs present a Gln residue at the position of a conserved His (Parisi et al., 2004a and this work]. Whereas cCA proteins are able to form homodimers, interaction between AtcCAL and AtcCA proteins is stronger than between cCA proteins themselves. Two-hybrid screens using the cCA2 protein neither identified complex I subunits nor cCAs. This could be interpreted in favour of a cCA/cCAL complex, which could exist independently and peripherally associated with complex I, which is supported by some Western Blotting results (Figure 3 and data not shown). Alternatively, interaction of cCA/cCAL could be required prior to association with complex I and thus, none of complex I subunit was detected in the two hybrid screen. The association to other respiratory complexes remains unclear. Recently, Millar et al. (2004) using mass spectrometry failed to identify the proteins described in this report in complex IV. An alternative interpretation is that presence of cCA/cCAL proteins in regions corresponding to smaller protein complexes rather reflects artificial aggregation due to digitonin solubilization and/or partial degradation of complex I into subcomplexes. Further work is required to solve this point. Further evidence, that AtcCA/cCAL do not only co-migrate with complex I on 2D BN/SDS gels, but indeed could be considered to physically interact with this complex, came from older investigations, which reported chromatographical preparations of complex I from other plants and

subunit identification by N-terminal protein sequencing (Leterme and Boutry, 1993; Herz et al., 1994). Thus, using completely different procedures to purify complex I, several proteins, having significant sequence similarity to Arabidopsis cCA and cCAL, have been identified. Preliminary experiments using sucrose gradient separation of digitonin treated mitochondrial complexes and western blotting with anti-cCA2, show very similar results to 2D BN/SDS gels (Perales, M., Eubel, H. and Braun H-P., unpublished results). Physiological function of these novel proteins associated to complex I is, so far, unclear. Recombinant AtcCAL proteins produced in E. coli showed no carbonic anhydrase activity. This situation was also found for AtcCAs (Parisi et al., 2004a). Thus, it is tempting to postulate that a proper CA activity of these proteins may require heteromeric organization and/or association with respiratory complex I. The presence of homologous ESTs of cCA and cCAL mitochondrial proteins in all plants and green algae analyzed so far, indicate that both protein types have been conserved throughout the evolution. In consequence, both might be required for proper mitochondrial function during plant growth and development. Thus, cCALs may represent an ancient divergence event from the unique prokaryotic cCA after eukaryotic green algae arose. In contemporary plants, both kinds of proteins show strong interactions associated with the same intramitochondrial complex. The results obtained in yeast suggest that at least two cCAs and two novel cCALs may interact to form heterodimers, or even more complex structures. This is in good agreement with the results obtained in planta that clearly indicate that these proteins have the same suborganellar localization, mainly in the respiratory complex I. Our results constitute a strong argument to support the idea that interaction between these proteins could effectively occur in vivo. The fact that cCALs interact with cCAs and all are associated to complex I raised some interesting questions about their physiological role. Association of CAs with chloroplastic NAD(P)H dehydrogenases has been postulated to function as local pH regulators (Maeda et al., 2002) necessary for proper electron flux between PSI and quinones. Based on this observation, it is tempting to postulate that a complex cCA/cCAL functions, in

956 association with complex I and probably supercomplexes, to regulate the differential pH caused by proton efflux. As suggested recently, the actual metal ligand of cCAs would be Fe2þ instead Zn2þ (Tripp et al., 2004), thus these carbonic anhydrase like proteins associated with mitochondrial complex I could link the respiratory chain function with a CA activity. Based on two hybrid results, we postulate that the holoenzyme could be formed in vivo with two or three different monomers. If this is the case, it could imply that plant cCAs might be more complex than prokaryotic ones. It is interesting to note that since AtcCA and AtcCAL proteins are able to interact themselves and they are detected at the migration positions of lower molecular weight complexes in BN 2D gel, the existence of minor amounts of free CA complexes, besides a high percentage associated to complex I, can not be ruled out. The free forms may represent the CA multimers precursor to their association with complex I or may play other physiological roles in mitochondria. cCA1 and cCA2 were found associated with complex I (Haezlewood et al., 2003, Braun, H-P. unpublished results). The strong signal detected at 28–30 kDa region corresponds to more than one protein, which could be interpreted as several isoforms of putative cCAs (Figure 3A, B and data not shown). Whereas cCALs are constitutively expressed in all tissue analyzed, cCAs show a 10fold induction in the flowers (Perales et al., unpublished results). This suggests that subunit composition of this putative CA complex associated with complex I may vary between different tissues. Differences in potato complex IV composition depending on tissue analyzed have been observed previously (Eubel et al., 2004). In conclusion, two novel LbH proteins have been identified in Arabidopsis (AtcCALs) that interact with AtcCA proteins forming putative active interfaces. All of them mainly occur associated with respiratory complex I. We postulate a CA complex consisting of AtcCA and AtcCAL proteins that might be important for complex I activity in plants.

Acknowledgments We would like to thank Dr. Jo¨rg Kudla for helps in performing two hybrid screens and

Mr. Jose´ Luis Burgos (CIC, Argentina) for excellent technical assistance. This work was supported in part by ANPCyT (05008, 09538 and 0111265) (Argentina), Fundacio´n Antorchas-DAAD (Germany-Argentina), ECOS-Sud, (A00B01) (FranceArgentina), Universidad Nacional de Quilmes, Third World Academy of Sciences, Fundacion Antorchas and the DFG (grant BR 1829-7/1). EJ, ZE and GCD are members of the National Research Council (CONICET). PM (ANPCyT, Argentina) and CA (CONICET, Argentina) are doctoral fellows and this work is part of their doctoral theses. Gene bank accession numbers: AB007649 (At5g63510) and AL133315 (At3g48680) References Alber, B.E. and Ferry, J.G. 1994. A carbonic anhydrase from the archaeon Methanosarcina thermophila. Proc. Natl. Acad. Sci. USA 91: 6909–6913. Badger, M.R. and Price, G.D. 2003. CO2 concentrating mechanisms in cyanobacteria: molecular components, their diversity and evolution. J. Exp. Bot. 54: 609–622. Busi, M.V., Bustamante, C., D’Angelo, M.C., Hidalgo-Cuevas, M., Boggio, M., Valle, E. and Zabaleta E. 2003. MADS box expressed during tomato seed and fruit development. Plant Mol. Biol. 54: 801–815. Combet, C., Blanchet, C., Geourjon, C. and Deleage, G. 2000. NPS@: network protein sequence analysis.Trends Biochem. Sci. 25: 147–150. Cuff, J.A., Birney, E., Clamp, M. and Barton, G. 2000. ProtEST: protein multiple sequence alignments from expressed sequence tags. Bioinformatics 16: 111–116. Eubel, H., Ja¨nsch, L. and Braun, H.-P. 2003. New Insights into the respiratory chain of plant mitochondria. Supercomplexes and a unique composition of complex II. Plant Physiol. 133: 274–286. Eubel, H., Heinemeyer, J. and Braun, H.-P. 2004. Identification and characterization of respirosome in potato mitochondria. Plant Physiol. 134: 1450–1459. Felsenstein, J. 1993. PHYLIP (Phylogeny Inference Package) version 3.5c. Distributed by the author. Department of Genetics, Univ. of Washington, Seatle. Ferraroni, M., Tilli, S., Briganti, F., Chegwidden, W., Supuran, C., Wiebauer, K., Tashian, R. and Scozzafava, A. 2002. Crystal structure of a zinc-activated variant of human carbonic anhydrase I, CA I Michigan 1: evidence for a second zinc binding site involving arginine coordination. Biochemistry 41: 6237–6244. Gietz, D., St Jean, A., Woods, A. and Schiesti, R. 1992. Improved method for high efficiency transformation of intact yeast cells. Nucl. Acids Res. 20: 1425. Heazlewood, J.L., Howell, K. and Millar, A.H. 2003. Mitochondrial complex I from Arabidopsis and rice: orthologs of mammalian and fungal components coupled with plantspecific subunits. Biochim. Biophys. Acta 1604: 159–169. Herz, U., Schroder, W., Liddell, A., Leaver, C.J., Brennicke, A. and Grohmann, L. 1994. Purification of the NADH:-

957 ubiquinone oxidoreductase (complex I) of the respiratory chain from the inner mitochondrial membrane of Solanum tuberosum. J. Biol. Chem. 269: 2263–2269. Hewett-Emmett, D. and Tashian, R.E. 1996. Functional diversity, conservation, and convergence in the evolution of the alpha-, beta-, and gamma-carbonic anhydrase gene families. Mol. Phylogenet. Evol. 5: 50–77. Hewett-Emmett, D. 2000. Evolution and distribution of the carbonic anhydrase gene families. In: W.R. Chegwidden, N.D. Carter and Y.H. Edwards (Eds.), The Carbonic Anhdyrases. New Horizons, Birkh€auser Verlag, Basel, pp. 29–76. Iverson, T.M., Alber, B.E., Kisker, C., Ferry, J.G. and Rees, D.C. 2000. A closer look at the active site of gamma-class carbonic anhydrases: high-resolution crystallographic studies of the carbonic anhydrase from methanosarcina thermophila. Biochemistry 39: 9222–9231. Kelley, L.A., MacCallum, R.M. and Sternberg, M.J. 2000. Enhanced genome annotation using structural profiles in the program 3D-PSSM. J. Mol. Biol. 299: 499–520. Kisker, C., Schindelin, H., Alber, B.E., Ferry, J.G. and Rees, D.C. 1996. A left-hand beta-helix revealed by the crystal structure of a carbonic anhydrase from the archaeon Methanosarcina thermophila. EMBO J. 15: 2323–2330. Kruft, V., Eubel, H., Ja¨nsch, L., Werhahn, W. and Braun, H-P. 2001. Proteomic approach to identify novel mitochondrial proteins in Arabidopsis. Plant Physiol. 127: 1694–1710. Lesburg, C.A., Huang, C.C., Christianson, D. and Fierke, C.A. 1997. Histidine carboxamide ligand substitutions in the zinc binding site of carbonic anhydrase Ii alter metal coordination geometry but retain catalytic activity. Biochemistry 36: 15780–15791. Leterme, S. and Boutry, M. 1993. Purification and preliminary characterization of mitochondrial complex I (NADH: ubiquinone reductase) from broad bean (Vicia faba L.). Plant Physiol.102: 435–443. Maeda, S., Badger, M.R. and Price, G.D. 2002. Novel gene products associated with NdhD3/D4-containing NDH-1 complexes are involved in photosynthetic CO2 hydration in the cyanobacterium, Synechococcus sp. PCC7942. Mol. Microbiol. 43: 425–435. Merlin, C., Masters, M., McAteer, S. and Coulson, A. 2003 Why is carbonic anhydrase essential to Escherichia coli ? J. Bacteriol. 185: 6415–6424. Millar, A.H., Eubel, H., Jansch, L., Kruft, V., Heazlewood, J.L. and Braun, H.P. 2004. Mitochondrial cytochrome c oxidase and succinate dehydrogenase complexes contain plant specific subunits. Plant Mol. Biol. (in press) McCall, K. and Fierke, C, 2004. Probing determinants of the metal ion selectivity in carbonic anhydrase using mutagenesis. Biochemistry 43: 3979–3986. McGuffin, L.J., Bryson, K. and Jones, D.T. 2000. The PSIPRED protein structure prediction server. Bioinformatics 16: 404–405.

Parisi, G., Fornasari, M. and Echave, J. 2000. Evolutionary analysis of gamma-carbonic anhydrase structurally related proteins. Mol. Phylogene. Evol. 14: 323–334. Parisi, G., Perales, M., Fornasari, M.S., Colaneri, A., Gonz
alez-Shain, N., G
omez-Casati, D., Zimmermann, S., Brennicke, A., Araya, A., Ferry, J.G., Echave, J. and Zabaleta, E. 2004a. Gamma Carbonic Anhydrases in Plant Mitochondria. Plant Mol. Biol. (In press). Parisi, G. Fornasari, M.S. and Echave, J. 2004b. Dynactins p25 and p27 are predicted to adopt the LbH fold FEBS lett. (In press). Price, G.D., Hewit, S.M., Harrison, K. and Badger, M.R. 1993. Analysis of genomic DNA region from the cyanobacteria Synechococcus sp strain PCC7942 involved in carboxysome assembly and function. J. Bacteriol. 175: 2871–2879. Raetz, C.R. and Roderick, S.L. 1995. A left-handed parallel beta helix in the structure of UDP-N-acetylglucosamine acyltransferase. Science 270: 997–1000. Sali, A. and Blundell, T.L. 1993. Comparative protein modeling by satisfaction of spatial restraints. J. Mol. Biol. 234: 779– 815. Sambrook, J. and Russell, D. 2001. Molecular Cloning, a Laboratory Manual. 3rd edn. Cold Spring Harbor Laboratory Press. New York. Scha¨gger H 2001. Blue-native gels to isolate protein complexes from mitochondria. Meth. Cell Biol. 65: 231–244. Scha¨gger H, Pfeiffer K 2000. The ratio of oxidative phosphorylation complexes I-V in bovine heart mitochondria and the composition of respiratory chain supercomplexes EMBO J. 19: 1777–1783. Smith, K., Jakubzick, S.C., Whittam, T.S. and Ferry, J.G. 1999. Carbonic anhydrase is an ancient enzyme widespread in prokaryotes. Proc. Natl. Acad. Sci. USA 96: 15184–15189. Thompson, J.D., Gibson, T.D., Plewniak, F. Jeanmougin, F., and Higgins, D.G. 1997. The CLUSTALX windows interface: flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucl. Acids Res. 25: 4876–4882. Tripp, B. C. and Ferry, J.G. 2000. A Structure-Function Study of a proton transport pathway in the gamma-class carbonic anhydrase from methanosarcina thermophila. Biochemistry 39: 9232–9240. Tripp, B., Bell, C., Cruz, F., Krebs, C. and Ferry, J.G. 2004. A role for iron in an ancient carbonic anhydrase. J. Biol. Chem. 279: 6683–6687. Tu, C. K., Rowlett, R., Tripp, B.C., Ferry, J.G. and Silverman, D.N. 2002. Chemical rescue of proton transfer in catalysis by carbonic anhydrases in the beta- and gamma-class. Biochemistry 41: 15429–15435. Vaara, M. 1992. Eight bacterial proteins, including UDP-Nacetylglucosamine acyltransferase (LpxA) and three other transferases of Escherichia coli, consist of a six-residue periodicity theme. FEMS Microbiol. Lett. 76: 249–254.