Calmodulin methyltransferase is an evolutionarily conserved enzyme that trimethylates Lys-115 in calmodulin

Abstract

Calmodulin (CaM) is a key mediator of calcium-dependent signalling and is subject to regulatory post-translational modifications, including trimethylation of Lys-115. In this paper, we identify a class I, non-SET domain protein methyltransferase, calmodulin-lysine N-methyltransferase (EC 2.1.1.60). A polypeptide chosen from a fraction enriched in calmodulin methyltransferase activity was trypsinized and analysed by tandem mass spectrometry. The amino-acid sequence obtained identified conserved, homologous proteins of unknown function across a wide range of species, thus implicating a broad role for lysine methylation in calcium-dependent signalling. Encoded by c2orf34, the human homologue is a component of two related multigene deletion syndromes in humans. Human, rat, frog, insect and plant homologues were cloned and Escherichia coli-recombinant proteins catalysed the formation of a trimethyllysyl residue at position 115 in CaM, as verified by product analyses and mass spectrometry.

ARTICLE

Received 1 Mar 2010 | Accepted 29 Jun 2010 | Published 27 Jul 2010 DOI: 10.1038/ncomms1044

Calmodulin methyltransferase is an evolutionarily conserved enzyme that trimethylates Lys-115 in calmodulin

Roberta Magnani1, Lynnette M.A. Dirk1, Raymond C. Trievel2 & Robert L. Houtz1

Calmodulin (CaM) is a key mediator of calcium-dependent signalling and is subject to regulatory post-translational modications, including trimethylation of Lys-115. In this paper, we identify a class I, non-SET domain protein methyltransferase, calmodulin-lysine N-methyltransferase (EC 2.1.1.60). A polypeptide chosen from a fraction enriched in calmodulin methyltransferase activity was trypsinized and analysed by tandem mass spectrometry. The amino-acid sequence obtained identied conserved, homologous proteins of unknown function across a wide range of species, thus implicating a broad role for lysine methylation in calcium-dependent signalling. Encoded by c2orf34, the human homologue is a component of two related multigene deletion syndromes in humans. Human, rat, frog, insect and plant homologues were cloned and Escherichia coli-recombinant proteins catalysed the formation of a trimethyllysyl residue at position 115 in CaM, as veried by product analyses and mass spectrometry.

1 Department of Horticulture, Plant Physiology / Biochemistry / Molecular Biology Program, University of Kentucky , Lexington, Kentucky 40546, USA.

2 Department of Biological Chemistry, University of Michigan Medical School , Ann Arbor , Michigan 48109-5606 , USA . Correspondence and requests for materials should be addressed to R.L.H. (email: [email protected] ) .

NATURE COMMUNICATIONS | 1:43 | DOI: 10.1038/ncomms1044 | www.nature.com/naturecommunications

2010 Macmillan Publishers Limited. All rights reserved.

1

ARTICLE

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms1044

Post-translational methylation of protein lysyl residues has emerged as an important determinant of protein protein interactions. This modication is catalysed by protein lysine

methyltransferases (PKMTs), which, in conjunction with proteins with binding domains that recognize methylated lysyl residues 1 and enzymes that reverse lysyl methylation 24 , have important roles in regulating several cellular and developmental processes. Calmodulin (CaM) is a ubiquitous, calcium-dependent, eukaryotic signalling protein with a large number of interactors that is frequently trimethy lated at Lys-115, a solvent-accessible residue (see PDB le 1UP5 5 ). A limited number of studies have shown that the methylation state of CaM can change in developmental and tissue-dependent manners 68, inuence the activator properties of CaM with target enzymes 9 and cause phenotypic changes in growth and developmental processes at the level of a whole organism 10. These observations suggest that CaM methylation could be a dynamic mechanism attenuating the interaction of CaM with target proteins inuencing a plethora of eukaryotic cellular and developmental processes.

The identication of the PKMTs responsible for histone methylation 11 led to our current understanding of the nuclear mechanisms that regulate gene expression and other genomic processes. The functional identication of genes for CaM lysine methyltransferase (KMT) could lead to a transformative understanding of new mechanisms regulating CaM signalling-dependent pathways in all eukaryotic organisms and highlights new roles for lysine methylation in cytoplasmic signalling pathways. Th e wide array of species that have protein / nucleotide sequences for CaM KMT suggests a potentially broad role for the methylation status of CaM as an important determinant of protein protein interactions.

Because of the potential importance of trimethyllysine-115 in CaM as a regulator of CaM function, we sought to provide a genetic identication of the enzyme catalysing formation of trimethyllysine-115 in CaM. Aer partial purication for enrichment of enzymatic activity, we identied a candidate polypeptide by photo-labelling with [ 3H-methyl]

S-adenosylmethionine (AdoMet) and obtained a polypeptide sequence using tandem mass spectro metry. Homologous conserved sequences were found across a wide range of eukaryotic species, all annotated as unknown function. Using recombinant bacterial expression, we showed that these sequences encode a class I protein methyltransferase responsible for the formation of trimethyllysine in CaM.

ResultsCaM KMT identication and characterization. Using a previously published protocol for purication of CaM KMT 12 , we obtained a protein fraction from lamb testicles enriched ~ 7,500-fold in CaM KMT activity. A single Sypro Ruby-stained polypeptide was identied as a

potential CaM KMT candidate based on molecular mass ( ~ 38 kDa) and comparison with a duplicate sample photolabelled by [ 3H-methyl] AdoMet ( Fig. 1a ). Th e polypeptide was proteolysed with trypsin, and the amino-acid sequence of tryptic fragments was identied using tandem mass spectrometry (MS / MS). The deduced full-length polypeptide sequence corresponded to a protein of unknown function with a number of homologues found throughout the plant and animal kingdoms. cDNA clones from Homo sapiens (Hs; Isoform 1), Rattus norvegicus (Rn), Xenopus laevis (Xl),

Tribolium castaneum (Tc) and Arabidopsis thaliana (At ) were obtained, and aer subcloning the coding regions into bacterial expression vectors, proteins were produced in Escherichia coli and the enzymes puried using bacterially expressed bovine ( Bt) CaM-Sepharose affi nity chromatography. Incu bation of the recombinant enzymes with bacterially expressed forms of CaM (which lack methylation at Lys-115) in the presence of [ 3 H-methyl] AdoMet resulted in the incorporation of radiolabel into CaM ( Fig. 1b ). In vitro enzymatic assays showed that all ve enzymes had robust methyltransferase activity against

CaM; for example, RnCaM KMT exhibited a kcat of 0.025s1, which is comparable to the results published for CaM KMT puried to homogeneity from sheep brain (0.03 s 1; Fig. 1c and Table 1)12.

SYPROR Ruby-stained SDS-PAGEkDa 1 2

97.466.2

45.0

31.0

21.5

14.4

Phosphorimage

Coomassie blue-stained SDS-PAGE1 2 3 4 104.3

kDa

94.651.6

36.8

28.5

19.5

0.020

0.015

0.010

0.005

0.000

0 2 4 6 8 10 12 14 Calmodulin (M)

Turnover (s-1)

100

90 80 70 60 50 40 30 20 10

0 799.0 1441.8 2084.6 2727.4 3370.2

100

90 80 70 60 50 40 30 20 10

0 799.0 1441.8 2084.6 2727.4 3370.2

Phosphorimage

Ninhydrin-stained Phosphorimage

TLC plate

% Intensity % Intensity

1 2 1 2

Mass (m/z)

Me2K

K

Me3K

MeK

Mass (m/z)

Figure 1 | Identication of CaM KMT. ( a) Sypro Ruby-stained SDSPAGE(12.5 % ) of a lamb CaM KMT-enriched protein fraction (left panel, lane 2) and [ 3H-methyl] AdoMet photolabelling of an ~38kDa polypeptide similar in molecular mass to that previously reported for the native formof CaM KMT puried to homogeneity from sheep brains 12 (right panel, only lane). Lane 1, molecular mass standards; ( b) Coomassie blue-stained SDS PAGE (12.5 % ) (top panel) and phosphorimage (bottom panel) of RnCaMT PKMT-dependent methylation of BtCaM. Lane 1, molecular mass standards; Lane 2, BtCaM and [ 3H-methyl] AdoMet, 9,950 disintegrations per min (d.p.m.) loaded; Lane 3, RnCaM KMT and [ 3H-methyl] AdoMet, 2,850 d.p.m. loaded; Lane 4, BtCaM, RnCaM KMT and [ 3H-methyl] AdoMet, 1,897,070d.p.m. loaded; (c) Michaelis Menten plot of turnover (s 1) versus BtCaM concentration with RnCaM KMT; ( d) Thin-layer chromatographic product analyses of radiolabelled methylated lysyl residues in BtCaM after incubation with [ 3H-methyl] AdoMet and RnCaM KMT (Lane 2), followed by total hydrolysis in 6 N HCl. In the right panel, vertical lines dene the lanes. Lane 1, standards of Lys and methylated Lys; ( e) Mass spectraof trypsinized BtCaM, unmethylated (upper panel) or after incubationwith RnCaM KMT and AdoMet (lower panel). Arrows indicate masses corresponding to peptides His-107 to Lys-115 (HVMTNLGEK; m/z 1,028)and Leu-116 to Arg-126 (LTDEEVDEMIR; m/z 1,350) in unmethylated BtCaM and peptide His-107 to Arg-126 (HVMTNLGE(Me3K)LT; m/z 2,401), containing three methyl groups in BtCaM incubated with RnCaM KMT.

NATURE COMMUNICATIONS | 1:43 | DOI: 10.1038/ncomms1044 | www.nature.com/naturecommunications

2010 Macmillan Publishers Limited. All rights reserved.

2

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms1044

ARTICLE

Table 1 | Kinetic parameters for native and CaM KMT clones.

Recombinant clone source

Specic activity (nmolmin1

mgprotein1)

Homo sapiens* ND 0.029 48 Rattus norvegicus 4.6 0.025 43 Xenopus laevis 2.1 0.030 50 Arabidopsisthaliana

Apparent

Km ( [H9262]M CaM)

Turnover (s1 )

ND 0.12 205

Tribolium castaneum

ND 0.019 31

0.1 0.028 44

Reported native protein (sheep brain 12)

Abbreviations: CaM KMT, calmodulin lysine methyltransferase; Hs, Homo sapiens; ND, not determined.

*Hs CaM KMT after Phe-209 to Val alteration.

Figure 2 | Modelling of

HsCaM KMT as a class I MT. Molecular model of the catalytic core of HsCaM KMT (Isoform 1) built using ESyPred3D 14 and

based on the homology with ribosomal L11 protein lysine methyltransferase (PrmA; 2ZBQ.pdb chain A). The loops highlighted in red denote the putative AdoMet-binding site. This gure was rendered using PyMOL( http://www.pymol.org/ ). In the ribbon representation of the catalytic domain of HsCaM KMT, Motif I (residues 149 158) is shown in blue, Post-Motif I (residues 174 180) in green, Motif II (residues 225 233) in tealand Motif III (residues 234 262) in purple, and are dened according to Petrossian and Clarke 13. For clarity, only residues 97 291 are depicted.

CaM KMT substrate specicity.Both,thedegreeof methylation and site specicity towards CaM, were investigated. Formation of a trimethyllysyl residue in BtCaM was documented by product analysis a er HCl hydrolysis using thin-layer chromatography and phosphorimagery aer

in vitro methylation with [ 3H-methyl]AdoMet ( Fig. 1d ). Site specicity was veried using in vitro assays, followed by SDS polyacrylamide gel electrophoresis (PAGE) and digestion of Bt CaM with trypsin and identication of peptides by MS / MS (Fig. 1e). Th e spectra included a peak at 2,401 Da, corresponding to the peptide containing Lys-115 modied by the addition of three methyl groups (42 Da) and resistant to tryptic digestion. Peptides with masses of 1,028 and 1,350 Da, products obtained from digestion in the absence of methylation, were not present in the spectra. To exclude the possibility that other lysyl residues in BtCaM were methylated, an additional digestion of Bt CaM with Asp-N protease, followed by MS / MS peptide analyses, was performed, as the trypsin-digested methylated CaM sample had <100% coverage of lysyl-containing peptides. Th e spectra from both digestions conrmed that only Lys-115 was methylated. Th ese results show that the nucleotide and associated polypeptide sequences identied here encode PKMTs specic for Lys-115 in CaM (CaM KMT; EC 2.1.1.60).

CaM KMT sequence analyses . All CaM KMTs contain an annotated AdoMet-binding motif found in a large number of class I methyltransferases, 13 including protein arginine methyltransferase, ribosomal protein L11 methyltransferase (PrmA) and protein isoaspartyl methyltransferase. Class I methyltransferase is known to be conserved on a dened secondary structure level, with several conserved primary sequences. Using the ESyPred3D server 14 and L11 PrmA from Thermus thermophilus as a template (2ZBQ.pdb; chain A), a molecular model for Hs CaM KMT corresponding to residues 69 292 was generated, corresponding to the catalytic domain of class I methyltransferases that bind AdoMet ( Fig. 2 ). The template, specically chosen as a class I PKMT, shares 17.1 % identity and 22.8 % homology with the sequence of Hs CaM KMT ( Fig. 3 ). The N- and C-terminal regions that ank the catalytic core are predicted to possess an ordered secondary structure according to structure prediction programs, but display no sequence homology with PrmA, protein arginine methyltransferases and protein isoaspartyl methyltransferase.

Th e human gene for Hs CaM KMT, c2orf34, is at locus 2p21, a region that is subject to deletions that are linked to cystinuria type I, hypotonia-cystinuria syndrome (HCS), an atypical HCS and the 2p21 deletion syndrome 15,16 ( Fig. 4 ). The deletions dier in size and the severity of the associated diseases correlates with the number

of genes aected. Patients with atypical HCS have a deletion that spans the two genes associated with HCS (SLC3A1 and PREPL) plus c2orf34; they manifest numerous biochemical symptoms and pheno-types that are intermediate in severity between HCS and the 2p21 deletion syndrome.

Discussion

Similar to many post-translational processing enzymes, CaM KMT is a relatively scarce enzyme. Purication of CaM KMT to homogeneity has only been accomplished once and required a purication scheme resulting in 20,000-fold enrichment 12 . A modied version of this protocol yielded a protein fraction ( Fig. 1a ) with high activity as a protein methyltransferase using AdoMet and BtCaM as substrates, and enabled the identication of the protein sequences associated with CaM KMT activity across many species ( Fig. 3 ). Kinetic parameters were generally similar among CaM KMTs, with two exceptions; a higher Km for CaM for the recombinant enzymes described here compared with that reported for the native sheep brain CaM KMT and a high kcat associated with AtCaM KMT (Table 1). These dierences could be a consequence of the dierent CaMs used as substrates. Previous studies relied on either naturally unmethylated forms of CaM from other species 17 or a genetically tailored and cloned form of CaM containing amino-acid sequences from both plant and vertebrate CaMs 18 . Here, we used a cloned form of bovine CaM ( BtCaM)for Hs, Rn and Xl CaM KMT assays (because CaM is identical in mammals and amphibians), and measured At and Tc CaM KMT activity with recombinant AtCaM2and TcCaM1,respectively.Th e activity of Tc CaM KMT is particularly interesting because insect CaM has been reported as unmethylated at Lys-115. Perhaps there are developmental and tissue-specic patterns to TcCaM KMT expression that result in methylated forms of Tc CaM that have not yet been detected. CaM has been reported to be monomethylated at Lys-94 in Drosophila eyes 8, but

TcCaM KMT only catalysed methylation of Lys-115 in our in vitro

studies, as

veried by MS analyses.

Class I methyltransferases have signature motifs associated with the catalytic core and an AdoMet-binding site. CaM KMTs contain all of these motifs ( Figs 2 and 3 ) and also have unique C- and N-terminal anking regions. These anking regions may be responsible for grasping CaM and docking Lys-115 into the active site of CaM KMT for trimethylation, as is observed in the ribosomal L11

NATURE COMMUNICATIONS | 1:43 | DOI: 10.1038/ncomms1044 | www.nature.com/naturecommunications

2010 Macmillan Publishers Limited. All rights reserved.

3

ARTICLE

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms1044

Figure 3 | Polypeptide sequence alignment of CaM KMTs and PrmA from

T. thermophilus. Using Jalview ( www.jalview.org 24), CaM KMTs and PrmA (corresponding protein accession no. AAH53733 Isoform 1, NP_001127935, NP_001099111, NP_001085547, XP_419462, XP_971600, NP_680769and Q84BQ9, respectively) were aligned and the conservation of residues highlighted by shades of blue; the darker the colour, the more conserved the residue 25. A green line over residues in the human sequence signies the peptides that were identied by MS / MS of the partially puried lamb CaM KMT. Red letters above the human sequence are the putative residues involved in AdoMet binding based on structural comparisons with other class I MTs; the rst of these represents the conserved Gly-rich loop found in all class I MTs. The yellow highlighted Phe (F) residue represents the SNP-altered residue. Amino-acid residues that correspond to Motif I, Post-Motif I, Motif II and Motif III in Class I MTs are delineated by lines coloured the same as in the model.

Figure 4 | Representation of human genes and deletions in locus 2p21 associated with different diseases. Using a screenshot of Homo sapiens chromosome 2 (GRCh37 primary reference assembly) between ~44,400 and 45,000 K nucleotides in the NCBI sequence viewer, a representation of the genes that are affected in cystinuria type I, hypotonia-cystinuria syndrome (HCS), atypical HCS and the 2p21 deletion syndrome was constructed. The boxes (gene symbol within) and the spacing between the boxes (intergenic region) are scaled appropriately. The arrow within the box represents the direction of transcription for mRNA. Coloured lines indicate the genes that are affected (by total or partial deletion) in each disease; cystinuria type I,HCS, atypical HCS and the 2p21 deletion syndrome are represented by green, red, yellow and light blue lines, respectively. These lines are notto scale regarding the size of deletions that have been characterized by two different research groups, Parvari and Creemers 15,16,2628. Full names of

the genes according to the Human Genome Organisation are as follows: PPM1B, protein phosphatase, Mg 2+ /Mn2+ -dependent, 1B; SLC3A1, solute carrier family 3 (cystine, dibasic and neutral amino-acid transporters, activator of cystine, dibasic and neutral amino-acid transport), member 1; PREPL, prolyl endopeptidase-like; and c2orf34, identied herein as CaM KMT, calmodulin lysine methyltransferase.

substrate-binding mode of PrmA 19. These regions ( Hs N-terminus, Ala-38 to Asp-55, Leu-57 to Asn-69; C-terminus Leu-299, Tyr-307 to Lys-321) are also conserved across all CaM KMT homologues

( Fig. 3 ) and may show the highly conserved nature of regions involved in CaM binding.

Th ere are some annotated discrepancies with HsCaM KMT (Isoform 1); the cDNA (accession BC 053733) has a single predicted amino-acid alteration (Val-209 to Phe). However, according to amino-acid sequence alignment, the valine residue is conserved among other mammalian and amphibian CaM KMTs, and residues with similar properties are present in both insect and avian enzymes ( Fig. 3 ). Only At CaM KMT has a glutamate residue at position 209 and it is the form with the highest kcat (Table 1).

Although the initially expressed form of HsCaM KMT with Phe-209 was catalytically inactive, alteration of the sequence to correspond to the gene-predicted valine restored catalytic activity to amounts similar to those observed for other CaM KMTs ( Table 1 ). To date, of th 2,000 reported single-nucleotide polymorphisms (SNPs) for this gene, this SNP (rs17855699) has no frequency data established; thus, the signicance of this discrepancy remains undetermined, although several other SNPs are apparently associated with specic human genetic backgrounds ( http://www.ncbi.nlm. nih.gov/sites/entrez?db=snp ). Several gene expression analyses suggest that there could be signicant tissue-specic dierences in c2orf34 expression, as well as dierences between normal and cancer cell lines ( http://www.genecards.org/cgi-bin/carddisp. pl?gene=C2ORF34).

Although

HsCaM KMT is part of two human deletion syndromes, other genes are involved ( Fig. 4 ) and a denitive assignment of a role for HsCaM KMT in the clinical observations associated with these disorders requires further genetic analysis. However, our identication of c2orf34 as Hs CaM KMT may facilitate new genetic and biochemical approaches for further characterization of the role of HsCaM KMT.

NATURE COMMUNICATIONS | 1:43 | DOI: 10.1038/ncomms1044 | www.nature.com/naturecommunications

2010 Macmillan Publishers Limited. All rights reserved.

4

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms1044

ARTICLE

Th e discovery of CaM KMT will enable a well-dened molecular and genetic approach to address whether post-translational methylation of CaM functions as a dynamic cytosolic control mechanism regulating its interaction with CaM-binding proteins, similar to the well-described eects of post-translational methylation of histones.

Methods

CaM afnity column preparation. A mass of 3.2mg of unmethylated CaM( BtCaM) puried from bacteria (described below) was coupled to 7 ml cyanogen bromide-activated Sepharose 4 Fast Flow ( Sigma Aldrich ) as described by Klee and Krinks 20. Alternatively, a pET28a BtCaM clone, engineered to have an N-terminal His-tag, was expressed and the protein puried using HisTrap HP columns ( GE Healthcare ). Subsequently, recombinant BtCaM was biotinylated using EZ-Link Iodoacetyl-LC-Biotin (Thermo Scientic ) on the only cysteine within the engineered His-tag tail of this CaM and then bound to strepavidin agarose ( Millipore ).

CaM cloning, expression and purication. Bt and

TcCaMs were cloned from a cDNA ( Open Biosystems through Th ermo Fisher Scientic ) and from Tc cDNAs (kindly provided by Dr. S.R. Palli), respectively. The amplied sequences, aer TA cloning, were subcloned to pET-28a ( Novagen ) and transformed into appropriate bacterial protein expression lines. Bacteria were grown at 37 C until OD 600 reached0.6, and were then induced with 0.5 mM isopropyl -d-1-thiogalactopyranoside (IPTG) and collected a er 4 h of additional growth at 37 C. CaMs were puried using HisTrap HP columns ( GE Healthcare ), desalted by dialysis overnight, followed by treatment with a Phenyl Sepharose 6 Fast Flow column (high capacity;

GE Healthcare ), similar to the procedure described in Roberts et al18. An

AtCaM2

clone in a modied pET5a vector (kindly provided by Dr. R.E. Zielinski) was used for expression and the recombinant protein was puried as the other CaMs.

CaM KMT partial purication. Following a previously published protocol 12, CaM KMT was partially puried from lamb testicles obtained from a local butcher shop; only the nal gel ltration on Superose 12HR was omitted.

[ 3H-methyl] AdoMet photolabelling. S-adenosylmethionine (AdoMet; Sigma) was puried before use 21 and diluted with [ 3H-methyl] AdoMet (GE Healthcare) to the concentrations required in each experiment. A CaM KMT-enriched fraction was incubated with 7 M [3H-methyl] AdoMet (120 10 6d.p.m.nmol 1) in 25mM HEPES (4-(2-hydro xyethyl)-1-piperazineethanesulfonic acid; pH 7.5), 200 mM NaCl, 0.05 % (w / v) CHAPS, 2 mM dithiothreitol (DTT) and 2 mM MgCl 2 on ice for 10 min. It was then transferred to a ceramic plate and crosslinked for 1 h using a 4 W UV lamp (254 nm; UVG-11 by UVP ) positioned over the ceramic plate ( ~ 7 mm between the sample and the lamp). Th e sample was loaded on a 12.5 % SDS PAGE gel and, a er electrophoresis, transferred to a PVDF membrane ( Millipore ) that was subse quently exposed to a storage phosphor screen ( GE Healthcare ) for 60 h.

In-gel digestion for liquid chromatography / MSMS. SDSPAGE gel slices containing stained polypeptides were cut into 1 mm cubes, destained by two 30 min washes with 50 mM NH 4HCO 3/50% CH3CN, followed by 10 min of vortexing, then dried in a vacuum centrifuge. Proteins were reduced by addition of 50 mM NH 4HCO 3 containing 10 mM DTT and incubation at 57 C for 30 min. A er discarding the reducing liquid, proteins were alkylated by addition of 50 mM NH 4HCO 3 containing 50 mM iodoacetamide and incubated for 30 min in the dark at room temperature. Th e gel was washed twice with 50 mM NH 4HCO 3, once with CH 3CN and partially dried in a vacuum centrifuge. Th e dried gel was rehydrated on ice for 1h in 50mM NH4HCO 3 containing 10ngl1 of modied trypsin ( Promega ;

sequencing grade). An additional 50 mM of NH 4HCO 3 was added to cover the sample, and the gel was incubated for 18 h at 37 C. When digested with Asp-N, the gel was rehydrated in 50 mM NH 4HCO 3 containing 40ngl1 of Asp-N and incubated immediately at 37 C for 3 5 h. Peptides were extracted from the gel in 0.1 % formic acid by sonicating for 10 min, followed by vortexing for 10 min. The extraction was repeated with 50 % CH 3CN/0.1% HCOOH, the extracts combined and the volume reduced to eliminate most of the acetonitrile. Th e volume of the peptide solution was adjusted to 12 l with 0.1 % HCOOH and the solution was ltered though a 0.45 m Millex lter (Millipore) before liquid chromatography (LC)/MSMS analysis.

LC/MSMS analysis. Nano-ow reverse-phase LC-MS/MS was performed usinga nanoLC system ( Eksigent ) coupled with a QSTAR XL quadrupole time-of-ight mass spectrometer ( AB Sciex ) through a nanoelectrospray ionization source (Protana). Analyst QS so ware (AB Sciex) was used for system control and data collection. Th e desired volume of protein solution was injected by the autosampler and desalted on a C18 trap column (300 m1mm, Vydac) for 6min at a ow rate of 10lmin1. Th e sample was subsequently separated by a C18 reverse-phase column (75 m15cm, Vydac) at a ow rate of 220 nl min 1. Th e mobile phases consisted of water with 0.1 % formic acid (A) and 90 % acetonitrile with 0.1 % formic acid (B), respectively. A 90-min linear gradient from 5 to 50 % B was typically used. A er LC separation, the sample was introduced into the mass spectrometer

through a 10 m silica tip ( New Objective ) adapted with a nanoelectrospray source ( Protana ). Data were acquired in information-dependent acquisition mode. Each cycle typically consisted of a 1 s time-of-ight mass spectrometry survey from 4001,600 (m/z) and two 2 s MS / MS scans with a mass range of 65 1,600 ( m/z).

Th e LC-MS / MS data were submitted to a local MASCOT server (Matrix Science Inc.) for MS/MS ions search.

CaM KMTs cloning, expression and purication. Th e cDNA clones for Hs,Rn and XlCaM KMT ( Open Biosystems ) were used as templates for amplifying the coding regions. A er subcloning to pET-23b (Novagen) from pGEM-T Easy (Promega), these constructs were transformed into BL21(DE3)pLysS (Promega). Bacterial cell cultures were grown at 37 C until OD 600 reached 0.4, and then adjusted to 0.25 mM IPTG and the culture le to grow at 20 C for 3 h. Bacteria were lysed in 20 mM Tris (pH 8), 150 mM NaCl, 2 mM CaCl 2, 5mM DTT and loaded onto a BtCaM-coupled column (prepared as described above). The column was washed extensively with 20 mM Tris (pH 8), 0.6 M NaCl, 5 mM DTT and 0.5mM CaCl2, and the CaM KMTs were eluted using 20 mM Tris (pH 8), 1 M NaCl, 5 mM DTT and 1 mM EGTA.

Tc and AtCaM KMTs were cloned from Tc and

At cDNAs that were kindgi s from Dr S.R. Palli and Dr A.B. Downie, respectively. Clones were generated in a N-terminal hexahistidine (His 6)-Smt3 fusion vector and transformed into

Rosetta(DE3)pLysS ( Novagen ). Bacteria were grown and induced with 0.5 mM IPTG once OD 600 reached 0.6 and the culture was le to grow overnight at 17 C. Tc and

AtCaM KMT were puried using a HisTrap HP column ( GE Healthcare ), followed by an overnight digestion with Ulp1 protease to remove His-tag and Smt3. Finally, aer affi nity chromatography on a BtCaM-coupled column as described above, the CaM KMTs were desalted by dialysis overnight and then used for kinetic assays.

CaM KMT assays. Using a modied protocol for Rubisco large subunit methyltransferase 22, CaM KMT enzyme assays, in a nal volume of 20 l, were carried out using 100 mM bicine (pH 8), 150 mM KCl, 2 mM MgCl 2, 2.5mM MnCl2,0.01% Triton X-100, 0.1mM CaCl2, 2mM DTT, [3H-methyl] AdoMet (3943M, 59106d.p.m.nmol 1), 0.2g CaM KMT and various concentrations of bacterially expressed CaM. For Hs, Rn and XlCaM KMT assays, recombinant BtCaM was used because CaM is identical among mammals and amphibians, whereas recombinant AtCaM2 and TcCaM1 were used for At and TcCaM KMT assays, respectively. All reactions were performed at 37 C for 2 min and terminated by protein precipitation with 500 l of 10% (v/v) trichloroacetic acid (TCA). Th e precipitated protein pellet was dissolved in 150 l of 0.1 N NaOH and precipitated again with the same volume of TCA before dissolution in 50 l of formic acid, followed by the addition of Bio-Safe II ( Research Products International ) and liquid scintillation analysis( Packard Tri-Carb 2200CA). Preparation of samples for phosphorimage analyses followed a similar protocol, except that, a er the second TCA precipitation, samples were dissolved in SDS PAGE loading buer, electrophoresed on 12.5 % SDS PAGE gels and transferred to a PVDF membrane before exposure.

Thin-layer chromatographic product analysis. BtCaM (21.6g) was incubated with 1 g of RnCaM KMT and [ 3H-methyl] AdoMet (200 M, 1.6106 d.p.m. nmol 1) in a nal volume of 20 l containing 100 mM bicine (pH 8) and 20 mM MgCl 2 at 30 C for 3 h. Because of its instability, more [ 3H-methyl] AdoMet (same concentration as before) was added a er 3 h. Th e volume was adjusted to 40 l using concentrated reaction buer to maintain the buer strength and incubated for an additional 3 h. TCA (10 % v / v) was added and the sample processed as described above for CaM KMT assays, except that the nal protein pellet from TCA precipitation was subjected to complete hydrolysis with 6 N HCl, followed by analyses for methylated lysyl derivatives by thin-layer chromatography 23.

Modelling CaM KMT . A molecular model of the catalytic core of HsCaM KMT (Isoform 1) was generated on the basis of its homology with ribosomal L11 PKMT (PrmA; 2ZBQ.pdb chain A). The ESyPred3D 14 server used residues 69292 for modelling, and the image was rendered with PyMOL (http://www.pymol.org/).

References

1. Taverna, S. D., Li, H., Ruthenburg, A. J., Allis, C. D. & Patel, D. J.

How chromatin-binding modules interpret histone modications: lessons from professional pocket pickers . Nat. Struct. Mol. Biol. 14, 10251040 (2007).

2. Klose, R. J., Kallin, E. M. & Zhang, Y. JmjC-domain-containing proteins and histone demethylation.Nat. Rev. Genet. 7, 715727 (2006).

3. Klose, R. J. & Zhang, Y. Regulation of histone methylation by demethylimination and demethylation . Nat. Rev. Mol. Cell. Biol. 8, 307318 (2007).

4. Shi, Y. & Whetstine, J. R. Dynamic regulation of histone lysine methylation by demethylases.Mol. Cell. 25, 114 (2007).

5. Rupp, B., Marshak, D. R. & Parkin, S. Crystallization and preliminary X-ray analysis of two new crystal forms of calmodulin . Acta Crystallogr. D Biol. Crystallogr. 52, 411413 (1996).

6. Oh, S. H. & Roberts, D. M. Analysis of the state of posttranslational calmodulin methylation in developing pea plants . Plant Physiol. 93, 880887 (1990).

5

NATURE COMMUNICATIONS | 1:43 | DOI: 10.1038/ncomms1044 | www.nature.com/naturecommunications

2010 Macmillan Publishers Limited. All rights reserved.

ARTICLE

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms1044

7. Rowe, P. M., Wright, L. S. & Siegel, F. L. Calmodulin N-methyltransferase. Partial purication and characterization . J. Biol. Chem. 261, 70607069 (1986).

8. Takemori, N., Komori, N., Th ompson, J. N. Jr., Yamamoto, M. T. & Matsumoto,H. Novel eye-specic calmodulin methylation characterized by protein mapping in Drosophila melanogaster. Proteomics 7, 26512658 (2007).9. Roberts, D. M., Burgess, W. H. & Watterson, D. M. Comparison of the NAD kinase and myosin light chain kinase activator properties of vertebrate, higher plant, and algal calmodulins . Plant Physiol. 75, 796798 (1984).

10. Roberts, D. M. et al. Expression of a calmodulin methylation mutant aects the growth and development of transgenic tobacco plants . Proc. Natl Acad. Sci. USA 89, 83948398 (1992).

11. Rea, S. et al. Regulation of chromatin structure by site-specic histone H3 methyltransferases.Nature 406, 593599 (2000).

12. Han, C. H., Richardson, J., Oh, S. H. & Roberts, D. M. Isolation and kinetic characterization of the calmodulin methyltransferase from sheep brain . Biochemistry 32, 1397413980 (1993).

13.Petrossian, T. C. & Clarke, S. G. Multiple motif scanning to identify methyltransferases from the yeast proteome . Mol. Cell Proteomics 8, 15161526 (2009).

14.Lambert, C., Leonard, N., De Bolle, X. & Depiereux, E. ESyPred3D: prediction of proteins 3D structures . Bioinformatics 18, 12501256 (2002).

15. Martens, K., Jaeken, J., Matthijs, G. & Creemers, J. W. Multi-system disorder syndromes associated with cystinuria type I . Curr. Mol. Med. 8, 544550 (2008).

16.Parvari, R. & Hershkovitz, E. Chromosomal microdeletions and genes functions: a cluster of chromosomal microdeletions and the deleted genes functions.Eur. J. Hum. Genet. 15, 997998 (2007).

17.Wright, L. S., Bertics, P. J. & Siegel, F. L. Calmodulin N-methyltransferase. Kinetics, mechanism, and inhibitors . J. Biol. Chem. 271, 1273712743 (1996).

18. Roberts, D. M. et al. Chemical synthesis and expression of a calmodulin gene designed for site-specic mutagenesis.Biochemistry 24, 50905098 (1985).

19. Demirci, H., Gregory, S. T., Dahlberg, A. E. & Jogl, G. Multiple-site trimethylation of ribosomal protein L11 by the PrmA methyltransferase . Structure 16, 10591066 (2008).

20.Klee, C. B. & Krinks, M. H. Purication of cyclic 3 ,5-nucleotide phosphodiesterase inhibitory protein by affi nity chromatography on activator protein coupled to Sepharose . Biochemistry 17, 120126 (1978).

21. Chirpich, T. P., Zappia, V., Costilow, R. N. & Barker, H. A. Lysine 2,3-aminomutase. Purication and properties of a pyridoxal phosphate and S-adenosylmethionine-activated enzyme.J. Biol. Chem. 245, 17781789 (1970).

22. Magnani, R., Nayak, N. R., Mazarei, M., Dirk, L. M. & Houtz, R. L. Polypeptide substrate specicity of PsLSMT. A set domain protein methyltransferase . J. Biol. Chem. 282, 2785727864 (2007).

23.Dirk, L. M.et al. Kinetic manifestation of processivity during multiple methylations catalyzed by SET domain protein methyltransferases . Biochemistry 46, 39053915 (2007).

24. Waterhouse, A. M., Procter, J. B., Martin, D. M., Clamp, M. & Barton, G.J. Jalview version 2a multiple sequence alignment editor and analysis workbench.Bioinformatics 25, 11891191 (2009).25.Livingstone, C. D. & Barton, G. J. Protein sequence alignments: a strategy for the hierarchical analysis of residue conservation . Comput. Appl. Biosci. 9, 745756 (1993).

26. Chabrol, B. et al. Deletion of C2orf34, PREPL and SLC3A1 causes atypical hypotonia-cystinuria syndrome.J. Med. Genet. 45, 314318 (2008).

27. Martens, K. et al. Global distribution of the most prevalent deletions causing hypotonia-cystinuria syndrome.Eur. J. Hum. Genet. 15, 10291033 (2007).

28. Parvari, R. et al. Th e 2p21 deletion syndrome: characterization of the transcription content.Genomics 86, 195211 (2005).

Acknowledgments

Mass spectrometric analyses were performed by Dr C.M. Beach at the Universityof Kentucky, Center for Structural Biology Protein Core Facility. Th is core facility is supported in part by funds from the NIH National Center for Research Resources (NCRR) Grant P20 RR020171. Th e research reported here was supported by Department of Energy Grant DE-FG02-92ER20075 and Kentucky Science & Engineering Foundation Grant KSEF-1526-RDE-010 to R.L.H., and by NIH Grant R01 GM073839 to R.C.T.

We are grateful to Dr R.E. Zielinski for providing the AtCaM2 clone and to Drs S.R. Palli and A.B. Downie for the cDNAs from Tribolium castaneum and Arabidopsis thaliana, respectively.

Author contributions

R.M. and L.M.A.D. performed the experiments. R.C.T. generated the model and analysed the class I methyltransferase motifs. R.L.H. designed the study. R.M., L.M.A.D., R.C.T. and R.L.H. wrote the paper.

Additional information

Competing nancial interests: R.L.H., R.M., and L.M.A.D. are inventors on a pending patent application on nucleotide and amino-acid sequences coding for CaM KMTs.

Reprints and permission information is available online at http://npg.nature.com/ reprintsandpermissions/

How to cite this article: Magnani, R. et al. Calmodulin methyltransferase is an evolutionarily conserved enzyme that trimethylates Lys-115 in calmodulin. Nat. Commun. 1:43 doi: 10.1038/ncomms1044 (2010).

6

NATURE COMMUNICATIONS | 1:43 | DOI: 10.1038/ncomms1044 | www.nature.com/naturecommunications

2010 Macmillan Publishers Limited. All rights reserved.