Limited trypsinolysis at K60 there are 10 other K/R residues in CR I – II confirms that the linker of CR I – II is solvent-exposed and that other potential sites are protected by regular
Trang 1Structural and biochemical characterization of neuronal calretinin domain I – II (residues 1 – 100)
Małgorzata Palczewska1, Patrick Groves1, Attila Ambrus2,*, Agata Kaleta1, Katalin E Ko¨ve´r3, Gyula Batta4 and Jacek Kuz´nicki1,5
1 Department of Molecular and Cellular Neurobiology, Nencki Institute of Experimental Biology, Warsaw, Poland;
2 Department of Biochemistry and Molecular Biology, University of Debrecen, Hungary;3Department of Inorganic Chemistry,
and4Research Group for Antibiotics, Department of Chemistry, University of Debrecen, Hungary;
5 International Institute of Molecular and Cell Biology, Warsaw, Poland
This study characterizes the calcium-bound CR I – II
domain (residues 1 – 100) of rat calretinin (CR) CR, with
six EF-hand motifs, is believed to function as a neuronal
intracellular calcium-buffer and/or calcium-sensor The
secondary structure of CR I – II, defined by standard NMR
methods on 13C,15N-labeled protein, contains four helices
and two short interacting segments of extended structure
between the calcium-binding loops The linker between the
two helix – loop – helix, EF-hand motifs is 12 residues long
Limited trypsinolysis at K60 (there are 10 other K/R
residues in CR I – II) confirms that the linker of CR I – II is
solvent-exposed and that other potential sites are protected
by regular secondary structure.45Ca-overlay of glutathione
S-transferase (GST) – CR(1 – 60) and GST – CR(61 – 100)
fusion proteins confirm that both EF-hands of CR I–II have
intrinsic calcium-binding properties The primary sequence
and NMR chemical shifts, including calcium-sensitive glycine
residues, also suggest that both EF-hand loops of CR I–II bind
calcium NMR relaxation, analytical ultracentrifugation, chemical cross-linking and NMR translation diffusion measurements indicate that CR I – II exists as a monomer Calb I– II (the homologous domain of calbindin D28k) has the same EF-hand secondary structures as CR I– II, except that helix B is three residues longer and the linker has only four residues [Klaus, W., Grzesiek, S., Labhardt, A M., Buckwald, P., Hunziker, W., Gross, M D & Kallick, D A (1999) Eur
J Biochem 262, 933 – 938] In contrast, Calb I – II binds one calcium cation per mono-meric unit and exists as a dimer Despite close homology and similar secondary structures,
CR I – II and Calb I – II probably have distinct tertiary structure features that suggest different cellular functions for the full-length proteins
Keywords: calretinin; calcium; calbindin D28k; EF-hand; NMR secondary structure
Calretinin (CR) and calbindin D28k(Calb) are homologous
calcium-binding EF-hand proteins with 59% sequence
identity (rat forms) [1 – 3] The proteins contain six
helix – loop – helix motifs (EF-hands) in which the loops carry the calcium-binding ligands The rat sequences are
271 (CR) and 261 (Calb) amino acids long [4,5] Both proteins have distinct, predominantly neuronal cellular distributions [3,6 – 9] and Calb is additionally found in the digestive system [10] Both proteins are used as markers for
a subset of neurons and several neurodegenerative diseases [11 – 13] Immunohistochemistry of CR is used to distin-guish adenocarcinomas from mesothelioma [14,15], with
CR playing a role in the early stages of mesothelioma [16] In contrast, Calb is found in subpopulations of neuro-endocrine phenotypes of some carcinoids and small-cell carcinomas [17]
EF-hand proteins act as intracellular calcium-sensors, linked to protein signaling cascades, and/or calcium-buffers [3] For example, calmodulin binds to more than 100 different proteins in a calcium-specific manner, while parvalbumin is considered important in maintaining nontoxic levels of free intracellular calcium through its buffering ability [3] The function of CR is unclear 2 there is evidence for both buffer and sensor roles (reviewed in [18]) but no definite target protein has yet been identified for CR to support a sensor role Calb appears to play a buffer role in neurons [19] and possibly facilitates calcium uptake through the digestive system [10] However, there is evidence that intestinal brush border membrane alkaline phosphatase
Correspondence to J Kuznicki, Department of Molecular and Cellular
Neurobiology, Nencki Institute of Experimental Biology, 3 Pasteur
Street, 02-093 Warsaw, Poland Fax: 1 48 22 822 53 42,
Tel.: 1 48 22 659 31 43, E-mail: jacek@nencki.gov.pl or to G Batta,
Research Group for Antibiotics, Department of Chemistry, Egyetem te´r.
1, University of Debrecen, PO Box 70, H-4010 Debrecen, Hungary.
Fax: 1 36 52 512 914, Tel.: 1 36 52 512 900 extn 2370,
E-mail: batta@tigris.klte.hu
*Present address: The University of Arizona, Department of Chemistry,
1306 E University, Tucson, AZ 85721 – 0041, USA.
Note: a web page is available at
http://www.nencki.gov.pl/labs/cbplab/kuzhome.htm
(Received 28 June 2001, revised 24 September 2001, accepted
4 October 2001)
Abbreviations: CR, calretinin (rat isoform, unless otherwise stated); CR
I – II, calretinin residues 1 – 100 consisting of the first pair of EF-hand
motifs; Calb, calbindin D 28k (rat isoform, unless otherwise stated); Calb
I – II, calbindin D 28k 1 – 93 consisting of the first pair of EF-hand motifs;
Calb III – IV, calbindin D 28k 79 – 193, consisting of the second pair of
EF-hand motifs; CSI, chemical shift index; GST, glutathione
S-transferase; DOSY, diffusion ordered spectroscopy.
Trang 2and caspase-3 might serve as functional targets for Calb
[20,21]
The biochemical characterizations of CR [22 – 25] and
Calb [26 – 29] suggest both proteins undergo significant
calcium-dependent structural changes In particular,
calcium-modulated exposure of hydrophobic surfaces,
perhaps the most important sensor characteristic, has been
reported for both CR and Calb [22,24,30] The biochemical
properties of CR and Calb are similar to those of
calmodulin, the best-known EF-hand calcium sensor It is
difficult to localize the biochemical properties of CR and
Calb to particular domains but this information can be
obtained by studying protein fragments For example, the
structure of the calcium-bound first domain of Calb,
comprising the first two EF-hands (Calb I – II, residues
1 – 93), was recently reported [31] Calb I – II binds a single
calcium and forms a homodimer [31] We have expressed
and purified the homologous domain of CR (CR I – II,
residues 1 – 100) [23,32] This has allowed us to analyze the
biochemical and structural properties of CR I – II and to
compare them with Calb I – II Our data show that CR I – II
and Calb I – II share virtually the same EF-hand structures
However, a shorter helix B in CR I – II accentuates a longer
linker loop between the EF-hands of CR I – II compared to
Calb I – II In contrast to Calb I – II, CR I – II shows no
tendency to dimerize and both EF-hands of CR I – II bind
calcium We conclude that the significant structural and
biochemical differences between CR I – II and Calb I – II
are related to a region of poor sequence identity (residues
54 – 79 of CR) We expect the differences between CR I – II
and Calb I – II will be present in the full-length proteins and
that these differences may play a role in the distinct cellular
functions of CR and Calb
E X P E R I M E N T A L P R O C E D U R E S
Expression and purification of CR and CR fragments
Recombinant CR and CR fragments were expressed from
Escerichia coli as glutathione S-transferase (GST) fusion
proteins and were purified as described previously [23,33]
GST – CR(1 – 60) and GST – CR(61 – 100) were prepared
using the same techniques as described for the preparation
of other GST – CR fragments [23,33] Extracellularly
expressed CR I – II using Pichia pastoris was purified by
DEAE chromatography [32] and the proteins eluted were
desalted by dialysis The resulting protein solutions were
lyophilized to dryness Reconstituted protein concentrations
were established using the Bradford method, with BSA as
standard [34] CR I – II from P pastoris was used in all
experiments unless stated otherwise
15N-labeled CR I – II was produced from P pastoris
using 98% 15N-ammonium sulfate (Martek Biosciences)
and the same protocols as for unlabeled material [32]
13C,15N-labeled CR I – II was prepared as for 15N-labeled
material but with 5 g:L21glucose (day 1) and 2 g:L2198%
13C-glucose (day 2) (Martek Biosciences) as the sole carbon
source during the growth phase and the addition of 0.5%
13C-methanol each day (Euris-top, France) during the
expression phase (additional 5 days) Protein purification
followed the same steps as for unlabeled protein
Sequence/structure analysis Protein sequences for rat CR and rat Calb were retrieved from GenBank (accession numbers X66974 and M31178) The relevant parts of the N-terminal modified CR I – II and Calb I – II sequences were aligned with SeqWeb (Oxford Molecular Group) using theBESTFITalgorithm
NMR determination of secondary structure
A standard method of assigning protein backbone HN, NH,
CA, HA, CB, HB and CO chemical shifts was used, based
on protocols described in Cavanagh et al [35] and Sattler [36] The sample contained 1 mM13C,15N-labeled CR I – II,
50 mM Tris pH 7.7, 25 mM NaCl, 10 mM CaCl2 The experiments included HNCO, HN(CA)CO, HNCA,
HNHA, and HSQC The experiments were processed usingFELIX (MSI, San Diego) TheAUTOASSIGNprogram [37] was used to obtain automatic assignments All assignments were manually checked for residue type and compared to Calb I – II assignments [31] in homologous regions using FELIX (MSI software) or SPARKY [38] Full experimental details are included in BMRB entry 5156 Spectra at pH 5.6 were more complete but of poorer quality (more overlapping peaks in the area of the HSQC corresponding to unstructured peptide) The assignment at
pH 7.7 was extended by a series of experiments at pH 6.7 as few of the signals undergo large pH dependent changes in chemical shift A restricted set of triple resonance data was collected on a sample in which the pH was lowered to 6.7, leading to a more complete assignment
Limited trypsinolysis of CR I – II Tryptic digests of CR I – II were investigated using conditions and protocols previously described for CR [39] Briefly, 40 mg protein in 80 mL 150 mM KCl, 50 mM
Tris pH 7.5 was subjected to 1 : 100 trypsin (w/w) for
0 – 50 min Uncleaved (as control) and cleaved proteins were then separated by Tris/tricine/PAGE on 15% acrylamide gels [40] The identity of the CR polypeptides were confirmed by MALDI-TOF MS (CR (1 – 60), 7382 Da with 7377 Da expected, and CR(61 – 99) 4570 Da with
4572 Da, unknown loss of last residue) and N-terminal amino acid sequencing (GSMAG for CR(1 – 60), residues
22 to 1 3, and SDNFG for CR(61 – 100), residues 61 – 65)
Calcium-binding to the CR II EF-hand motif Samples of GST, CR, GST – CR I, GST – CR II and standards were separated on 10% acrylamide Tris/tricine/gels in the presence of SDS and then electroblotted to nitrocellulose (0.45 mm, Bio-Rad) in 20% methanol, 48 mM Tris, 39 mM
glycine, 0.37% SDS buffer (pH 8.3) for 60 min at 0.1 A The method of Maruyama et al [41] was followed with the blot incubated in 2 mM45Ca (10.5 Ci:g21Ca; ICN) in 5 mM
MgCl2, 60 mM KCl, 10 mM imidazole pH 6.8 for 10 min followed by washing in 50% ethanol (1 min) Subseqeuntly, blots were subjected to autoradiography on Hyperfilm (Amersham Life Science) and developed after 24 h
Trang 3Measurement of correlation time and translation
diffusion of CR I – II
The 15N T1, T2 and NOE relaxation data in a 1.5-mM
15N-labeled CR I – II sample with 10% D2O (for field lock),
33 mMNaCl, 10 mMCaCl2, 50 mMsodium acetate pH 5.4
at 310 K were measured in a series of two-dimentional
heteronuclear correlated spectra using the
sensitivity-enhanced gradient pulse schemes of Farrow et al [42]
Spectral widths were 13 and 28 p.p.m in F2 and F1,
respectively, on a Bruker DRX 500 spectrometer The1H
carrier frequency was placed on the water resonance and
at 118 p.p.m for nitrogen An in-house written MATLAB
routine was applied to fit NMR relaxation data to
experimental parameters
A sample of unlabeled 0.25 mM CR I – II was prepared
in 50 mM deuterated Tris, 25 mM NaCl pH 8.1 The
stimulated-echo with longitudinal eddy current delay
(STE-LED) method was applied to obtain diffusion ordered
spectroscopy (DOSY) spectra at 278 K [43,44] We
imple-mented Bruker’s LEDGS2S pulse sequence [45] with a
combined watergate and presaturation sequence which
allowed measurements in H2O/D2O solution The mean of
the diffusion profile arising from CR I – II protons between 0
and 4 p.p.m was determined The relative diffusion rate
measurements were carried out using known proteins as
external references or internal TSP
[3-(trimethylsilyl)pro-pionic acid, sodium salt] Calibration proteins included
ubiquitin (8.7 kDa), Spo0F (14.4 kDa) and HP-RNAse
dimer (24.7 kDa)
Analytical ultracentrifugation of CR I – II
Experiments were carried out on a Beckman XL-A
instrument in a sedimentation equilibrium experiment at
39 000 g as described for Calb I – II [31] Analytical
ultracentrifugation was performed at 293 K with a sample
concentration of 0.233 mg:mL21, < 20 mM as calculated
by a molar extinction coefficient of 1 ¼ 11 460 cm21:M21
The buffer used was 1 mM CaCl2, 40 mM potassium
phosphate pH 6.4
Chemical cross-linking of CR I – II
Reactions were carried out in 20 mL volumes modified from
the method described by Staros et al [46] six micrograms
of protein and 1 mMCaCl2were initially prepared in 10-mL
A stock solution of 10 mL 100 mM NaCl, 100 mM Tris
pH 8.0, with 20 mM N-hydroxysuccinimide and 8 mM
1-ethyl-3-(3-dimethylaminopropyl)-1-napthalene-sulfonic
acid (EDC), was then added After 60 min in the dark at
room temperature, the reactions were terminated by the
addition of 4 SDS sample buffer and heating at 95 8C for
5 min The reaction mixtures were separated by Tris/tricine/
SDS/PAGE [40], together with a low-range molecular
weight marker (Promega), on 10% acrylamide gels and
stained with Coomassie blue Bovine S100B, bovine
a-lactalbumin and horse skeletal myoglobin (all from
Sigma) were used as control protein samples Myoglobin
and a-lactalbumin produce single bands of monomeric
weight and S100B provides a positive control with a yield of
cross-linked dimer of < 30%
R E S U L T S
Definition of CR I – II structure The CR I – II and Calb I – II sequences have high sequence identity, Fig 1, and many other residues display conservative changes The region between residues 54 and
79 of CR I – II, encompassing the linker, helix C and start of the second calcium-binding loop, has least identity to Calb
I – II In this region, identity persists only for hydrophobic residues of helix C that are expected to form part of the hydrophobic core
Standard protocols were followed to obtain spectra suitable for backbone assignment of calcium-bound CR
I – II (full sample details, acquired NMR experiments and assignments at pH 6.7 and 7.7 are available from BMRB entry 4749) TheAUTOASSIGN program [37] was used to obtain automatic assignments By running the program with several sets of parameters, we were able to obtain a consistent auto-assignment for residues T17-I36, E70-I83 and T94-N97 that were subsequently confirmed manually Other assignments, most notably for helix B, were obtained
by manual assignment The assignment was extended at
pH 6.7 for residues between L13 and F98, except for K39, E40, K53, M58 and S61 Fig 2 shows an assigned HSQC spectrum of CR I – II at pH 7.7
A summary of the secondary structure elements defined
by pH 7.7 NMR data is given in Fig 3 Four helices are defined by chemical shift index (CSI), medium-range NOEs and measured3JHNHavalues smaller than 5 Hz (A18 – D31, G38 – K53, E66 – K77 and A86 – L92) Three short stretches
of extended structure (E15 – T17, Y35 – E37 and K82 – E84) are suggested by CSI However, the extended structure is supported only by consistently very strong HNi,Ha(i-1)
NOEs and 3JHNHa values 7 Hz for the latter two segments The short segments of extended structure found
Fig 1 Pairwise alignment of CR I – II (1 – 100) and Calb I – II (1 – 93) sequences (both of rat origin) Additional residues at the
N terminus, that are the result of molecular biology manipulations, are underlined An EF-hand consensus sequence is placed between the aligned EF-hand I and EF-hand II sequences [3] The helices are shown
as boxes with conserved hydrophobic (h) and other amino acids (.), the calcium-binding ligands are designated by their coordinates: x, y, z, -y, -x and -z [60] Residues discussed in the text are given in bold (G34, K60 and G81 of CR I – II and G29, G66, R68, G71 and G74 of Calb
I – II) The SeqWeb alignment is shown between the aligned EF-hand sequences: identical residues (|), homologous residues (:) and similar residues (.) are marked, as scored by the program The helices defined
by Klaus et al for Calb I – II are boxed [31] and the secondary structure elements are labeled (helices A – D; helices pre-A and D0found in Calb
I – II; the linker between EF-hand motifs; Cal I and Cal II denote the two calcium-binding loops) For comparison, the position of helices A – D is also boxed for the CR I – II sequences, as determined by the NMR data described in this paper (see Fig 3) Five asterisks in the linker at the start of EF-hand II of the Calb I – II sequence denote a five-residue long gap compared to the CR I – II sequence.
Trang 4in the calcium-binding loops interact in a b-sheet-like
manner This interaction between residues Y35 – E37 and
K82 – E84 is supported by interstrand NOEs (I36HN –
E84HA and I36HN – I83HN) The linker could not be
assigned at pH 6.7 The CSI scores at pH 6.7 are added to
Fig 3 as this data could not be confirmed unambiguously at
pH 7.7
Our secondary structure assignment is consistent with
known capping motifs T13 and Q16 provides an N-terminal
helix capping motif in ShK toxin, which plays a role in helix
stabilization [47] CR I – II (T17 and Q20) and Calb I – II (T12 and Q16) contain analogous residues at the N-terminus
of helix A The residues around G54 suggest that it forms the C-capping motif of helix B [48], reinforcing the NMR data that helix B of CR I – II is three residues shorter than that of Calb I – II The residues around G65 are also con-sistent with an N-capping motif and agree with the Calb I – II assignment, although N63 could provide an alternative capping site [48] A C-capping motif for the L92P93 pair of residues, as suggested by Prieto and Serrano [49], agrees with our secondary structure determination of CR I – II and
is also found in Calb I – II
Fig 4 shows the result of limited trypsinolysis of CR
I – II Using the conditions used previously to cleave CR, we found that CR I – II was fully digested in < 5 min, similar to that reported previously for CR [39] The resulting tryptic fragments remained relatively resistant to further cleavage ( 30 min) in the presence of calcium MALDI-TOF and N-terminal amino acid sequencing confirmed that the upper band is CR(1 – 60) and the lower band is CR(61 – 100) Preferential cleavage of CR I – II at K60 is consistent with the secondary structure presented in Fig 3 and the previous results for CR [39] All other potential trypsin cleavage sites are found either in helical regions (K26, K50, R52, K53, K67, K69, K74) or in the calcium-binding loops (K39, K77, K82)
Both EF-hands of CR I – II bind calcium The 12-residue long calcium-binding loops of EF-hands are readily identified from their primary sequence Significant deviations from the consensus sequence usually lead to a low-affinity calcium-binding site Here we provide experi-mental evidence that both EF-hands of CR I – II bind calcium CR fragments corresponding to the first (residues
1 – 60) and second (residues 61 – 100) EF-hands were expressed as GST fusion proteins We found thrombin cleaved CR fragments difficult to purify and blot Therefore,
45
Ca-overlay experiments were performed on the fusion products together with CR and GST as controls, Fig 5 Ponceau red staining of the electroblot indicated that some cleavage of the fusion proteins occurred and a lower band, most probably GST, was observed We observed that CR, but not GST (also the lower bands of lanes 2 and 3), binds
45Ca under the conditions used GST – CR(1 – 60) and GST – CR(61 – 100) provide 45Ca-overlay bands of similar inten-sity when blotted in similar quantities This is consistent with both EF-hand sequences containing a single calcium-binding site
NMR data also provides evidence suggesting that both EF-hands of CR I – II bind calcium The calcium-bound
Fig 2. 1H,15N HSQC of15N-labeled CR I – II with assignments Of
note are residues G34 and G81, the ‘position 6’ glycine residues.
Residues I36, E37, I83, E84 and M85 have1H and15N chemical shifts
characteristic of extended conformation (confirmed by other
assign-ments) Other resonances were assigned at pH 6.7 These assignments
are missing from this figure as they fall in the crowded part of the
spectrum and their assignments could not be transferred
unambigu-ously, even though some of these peaks are still observed in this HSQC.
A pH titration indicated that the majority of resonances did not undergo
pH-induced changes in chemical shift 0.1 p.p.m (1H) or
0.5 p.p.m ( 15
N) Lines connect the unassigned glutamine and
asparagine side-chain resonances.
Fig 3 Summary of the secondary structure assignments extracted
from NMR data at pH 7.7 Medium-range NOEs are given as lines
connecting assignments between HA and i 1 3 or i 1 4 HN resonances
(marked as NOE i 1 3 and NOE i 1 4, derived from a NOESY-HSQC
experiment) CSI parameters, scored as 1 1, 0 or 2 1, are calculated
from the backbone assignment according to Wishart et al [61] Note:
the CSI data for the linker region was determined from NMR data at
pH 6.7 3 J HNHa values (marked as 3J) are calculated from a HNHA
spectrum using a routine.
Fig 4 Limited trypsinolysis of CR I – II Incubation times are given above the lanes The two proteolytic fragments correspond to CR(1 – 60) and CR(61 – 100) for the upper and lower bands, respectively.
Trang 5NMR assignments of conserved glycine amide protons, at
position six of the 12-residue long calcium-binding loops of
EF-hand proteins, have chemical shifts 10 p.p.m for
filled loops and chemical shifts of < 8 p.p.m for empty
loops [50] The chemical shifts of CR I – II, Calb I – II and
several reference EF-hand proteins are given in Fig 6 The chemical shifts agree with biochemical data that only certain calcium-binding sites of the reference proteins are filled [31,50 – 56]
CR I– II is monomeric
15N NMR relaxation experiments can be used to determine global correlation times of proteins that are related to the size of the studied molecule Using the isotropic form of the Lipari – Szabo analysis [57], a global correlation time
of 6.7 ^ 0.5 ns was determined for calcium-bound
15N-labeled CR I – II at pH 5.4 and 310 K (Fig 7A) Also, the diffusion rate of an unlabeled CR I – II sample was evaluated under different conditions on an arbitrary log10 scale at pH 8.1 and 278 K (Fig 7B) Both sets of NMR data for CR I – II clearly fit with a monomeric mass (12 kDa) rather than a dimeric mass (24 kDa) when compared to reference proteins (Fig 7A,B) The NMR data indicate that
CR I – II exists as a monomer even at protein concentrations
as high as 1.5 mM
Fig 5 CR(1 – 60) and CR(61 – 100) bind calcium Proteins were
separated by 10% Tris/tricine/SDS/PAGE and electroblotted to
nitrocellulose as described in Experimental procedures (A) Ponceau
red stain of blotted proteins Lane 1, GST; lane 2, GST – CR(1 – 60); lane
3, GST – CR(61 – 100); lane 4, CR The fusion proteins in lanes 2 and 3
appear to be partly digested; the lower band corresponds to GST (B)
45
Ca-overlay of the blot shown in part A indicating that CR and GST
fusion products of CR fragments bind 45 Ca.
Fig 6 ‘Glycine 6’ of the calcium-binding loop
is a sensitive indicator of calcium loading Filled
bars denote EF-hand loops that bind calcium and
empty bars those that do not, as determined by
biochemical or crystal structure characterization
[31,50 – 56] (A) Frequenin (G33, G78, G114,
G162) [51,52] (B) GCAP-2 (G38, G72, G110,
G163) [53,54] (C) Calb I – II (G29, G71) [31] (D)
Nereis diversicolor sarcoplasmic Ca21-binding
protein (G21, G55, N109, G143) [55,56] (E)
Calerythrin (G22, G73, G117, G151) [50] (F) CR
I – II (G34, G81).
Fig 7 CR I – II is monomeric (A) Plot of global
correlation time determined from NMR relaxation
data for three reference proteins vs their known
molecular weights (X) The reference proteins are
ubiquitin (Martek Biosciences Corporation,
Columbia), Spo0F [62] and the dimeric HP-RNase
[63] The experimentally determined correlation
time of CR I – II is plotted against its predicted
monomeric mass (B) and the predicted dimeric
mass (A) (B) Plot of relative diffusion coefficients
for several reference proteins, as in (A) vs known
molecular weight (X) CR I–II experimental
DOSY data is plotted against its predicted
monomeric mass (B) and predicted dimeric mass
(A).
Trang 6Analytical ultracentrifugation showed earlier that
calcium-bound Calb I – II is dimeric [31] and this method was used
under identical conditions to establish the extent of
dimerization of CR I – II Analytical ultracentrifugation
revealed that $ 90% of 20 mMCR I – II exists as a monomer
in the presence of 1 mM CaCl2 Possible dimer, trimer and
tetramer species accounted for , 4% each sample
In cross-linking experiments, a single band is observed
for CR I – II (lane 4, Fig 8) at the molecular mass expected
for monomeric protein A small amount of cross-linked CR
I – II, < 3% according to densitomeric analysis, is probably
the result of nonspecific cross-linking as the same amount of
cross-linked dimer species was detected for the monomeric
proteins a-lactalbumin and myoglobin S100B, which forms
both covalent and noncovalent dimers, provides a yield of
30 – 40% This is an acceptable yield for this method [46]
D I S C U S S I O N
CR I– II EF-hand structure: comparison to Calb I – II
Only a partial backbone assignment of CR I – II was possible
at pH 7.7 because the intensity of some resonances are
diminished or unobserved in HSQC-based spectra (Fig 2)
Moreover, several weak resonances do not provide sufficient
intensity in three-dimensional NMR experiments to enable
positive assignment It is difficult to assign spectra of CR
I – II at lower pH due to the overlap of broad resonances
When the pH was lowered to pH 6.7, it was possible to
assign most of the resonances between L13 and F98 Many
of the new pH 6.7 assignments resonated in the region
where residues in unstructured segments of protein are
usually found and their CSI scores confirmed this The
N terminus of CR I – II could not be assigned even at pH 6.7
However, NMR data for Ala2 and Gly3 at pH 6.7 suggests
that these residues are in an unordered state The NMR data
of CR I – II is consistent with an unordered N terminus and
linker
Klaus et al described the N-terminal region as containing
a pre-A helix and noted that it is poorly formed [31] This
region was not detected in the NMR spectra of CR I – II
(Fig 3) There is also no sequence homology between CR
I – II and Calb I – II for this helix (Fig 1) The C termini of
the CR I – II and Calb I – II domains contain a consensus
sequence (ENFLLXF, where X is a variable residue), which
is only partly observed in Fig 1 This segment is present in
Calb and CR sequences of all species as well as the recently
described secretagogin sequence [58] We do not observe helix D0 in CR I – II (Fig 1) possibly because CR I – II is three residues shorter than Calb I – II
The secondary structures of CR I – II and Calb I – II are almost identical within the EF-hand regions (see Fig 1) Helices A, C and D, together with the three-residue b-type cross-strand interaction between the two calcium-binding loops, are in identical positions However, helix B of CR
I – II is three residues shorter than helix B in Calb I – II A number of identified helix capping motifs are consistent with the determined secondary structure of CR I – II, includ-ing the shorter helix B of CR I – II Limited trypsinolysis of
CR I – II leads to fragmentation that is consistent with the defined secondary structure: K60 is in the linker whereas the other 10 potential cleavage sites are found in helical regions
or in the binding loops that scaffold the calcium-binding ligands and are consequently protected Helix B of
CR I – II is still longer than predicted by the EF-hand consensus sequence, an unusual feature first described for Calb I – II The CR I – II sequence also has a five-residue insertion in the linker compared to Calb I – II (see Fig 1) Therefore, the linker of CR I – II is 12 residues long compared to only four residues in Calb I – II
CR I– II binds two calcium cations Based on a comparison with the consensus sequence, both EF-hands of CR I – II are predicted to bind calcium On the other hand, the residues at positions 66, 68 and 74 (G66, R68 and G74) of Calb I – II lack the side-chain oxygens expected
to provide calcium-binding ligands The absence of calcium-binding to the EF-hand II motif of Calb has been shown in the individual polypeptide, Calb I – II, Calb mutants and full-length Calb using a number of experi-mental techniques in several laboratories [26 – 29] The analogous residues of CR I – II (D76, N78 and E84), as well
as the whole second loop sequence, have suitable ligands with side-chain oxygens where the consensus sequence requires them for calcium binding The second EF-hand of
CR I – II has a potential high calcium affinity
The principal limited proteolysis products of CR I – II conveniently correspond to individual EF-hand motifs (residues 1 – 60 and 61 – 100) We found that the PAGE separated limited trypsinolysis products of CR I – II bound
45Ca However, GST fusion products of the same fragments were produced in order to clarify our data due to the poor blotting efficiency of the small CR fragments Both CR(1 – 60) and CR(61 – 100) bind 45Ca as GST fusion products Individual synthetic Calb EF-hands I, III, IV and V bind calcium [29] This data of A˚ kerfeldt et al [29] agrees with the large body of independent data collected on Calb and truncated Calbs by Kumar’s group This latter work revealed that EF-hands II and VI of Calb do not bind calcium and that Calb binds a total number of four calcium cations [26 – 28] Calcium-dependent NMR data for
‘position 6’ glycine residues also support calcium-binding
to both EF-hands of CR I – II (Fig 6) but only to EF-hand I
of Calb I – II These glycine chemical shifts report the formation of a strong interaction between the glycine amide and a carboxylate side-chain This has been shown empirically to occur when calcium ions are bound to the site
In summary, the45Ca-binding data for CR fragments and
Fig 8 10% Acrylamide gel run under Tris/tricine/SDS/PAGE
conditions and stained with Coomassie blue showing the products of
chemical cross-linking of 0.3 mg:mL21protein in the presence of
1 m M CaCl 2 Lane 1, S100B; lane 2, a-lactalbumin; lane 3, myoglobin;
lane 4, CR I – II; molecular weight standards are indicated to the left.
Trang 7NMR data for CR I – II indicate that the stoichiometry of
calcium binding in CR I – II is 2 : 1
Calcium-bound CR I – II is monomeric
We speculate that the Calb I – II homodimer structure
mimics the interaction between Calb I – II and other parts of
the Calb sequence [31], in accordance with the Linse model
where all six EF-hand motifs of Calb interact within a single
globular domain structure [59] In contrast to Calb I – II, CR
I – II is monomeric under various conditions We have
not observed an interaction between CR I – II and various
complement CR fragments using fluorescence, chemical
cross-linking and size exclusion chromatography
(M Palczewska, P Groves & J Kuz´nicki, unpublished
data) Therefore, we believe that CR does not form a single
globular domain Perhaps the different oligomeric states of
CR I – II and Calb I – II are an indication of different
organizations of EF-hands within full-length CR and Calb
CR I– II and Calb I– II may have different structures
CR I – II and Calb I – II have different biochemical
properties, despite their similar secondary structures The
four-residue linker of Calb I – II could restrict the relative
positions of helices B and C to a greater degree than the
12-residue linker of CR I – II The low-resolution NMR
structure of calcium-bound Calb I – II suggests that the
second calcium-binding loop may contain a distorted
con-formation [31] We show that the second calcium-binding
loop of CR I – II binds calcium and is therefore expected to
adopt a characteristic conformation found in other EF-hand
protein structures We predict that an overlay of future
high-resolution CR I – II and Calb I – II structures will reveal
that the largest structural differences lie in the orientation
of helix C, together with the adjacent linker and second
calcium-binding loop structures There is a region of poor
sequence identity localized to residues 54 – 79 of CR I – II
that supports this prediction (Fig 1) High resolution
structures of both apo- and calcium-bound domains are
essential to understand fully the differences in CR I – II and
Calb I – II structure and properties
Are the functions of CR and Calb distinct?
A large body of physiological data supports a calcium-buffer
role for both Calb and CR However, biochemical (and some
physiological) data suggest sensor roles, as do the
inter-action of Calb with at least two other proteins [20,21] The
ongoing high-resolution structure projects on both CR I – II
and Calb I – II will provide a clearer answer as to the nature
of the calcium-induced structural changes, differences in
structure and clues to the function of the full-length proteins
The present work establishes that the first domains within
CR and Calb have distinct structural features and
bio-chemical properties, and potentially perform different
functions within the full-length proteins
A C K N O W L E D G E M E N T S
We are greatly indebted to Werner Klaus (F Hoffmann-LaRoche AG,
Basel) for discussions about the different oligomeric states of Calb I – II,
the subject of his earlier studies, and CR I – II Francis Mueller and
Eric Kusznir (F Hoffmann-LaRoche AG, Basel) determined the analytical ultracentrifugation data on CR I – II under the same conditions as used earlier for Calb I – II Sa´ndor Ke´ki (Chemistry Department, Debrecen) determined MALDI-TOF data for the limited proteolysis products of CR I – II We thank Walter Chazin (Vanderbilt University, Nashville) for critical evaluation of the manuscript and Barbara Zarzycka (Warsaw) for her technical assistance This work was supported by an International Center for Genetic Engineering and Biotechnology (ICGEB) grant to J K [CRP/Pol97-01(t1)] and G B [CRP/Hun97 – 01(t1)] P G is supported by a grant from the State Committee for Scientific Research, KBN no 6 P04B 01015 G B thanks the support of the Hungarian National Fund, OTKA T-029089 The Bruker DRX 500 spectrometer used in these studies was purchased from the following grants: PHARE-ACCORD H-9112-0198, OMFB MEC 93-0098 and OTKA A084.
R E F E R E N C E S
1 Wasserman, R.H & Taylor, A.N (1966) Vitamin D 3 induced calcium-binding protein in chick intestinal mucosa Science 152,
791 – 793.
2 Rogers, J.H (1987) Calretinin: a gene for a novel calcium-binding protein expressed principally in neurons J Cell Biol 105,
1343 – 1353 (erratum appears in J Cell Biol 110, 1845)
3 Celio, M.R (1996) Guidebook to the Calcium-Binding Proteins, 1st edn Sambrook & Tooze with OUP, Oxford, UK.
4 Strauss, K.I & Jacobowitz, D.M (1993) Nucleotide sequence of rat calretinin cDNA Neurochem Int 22, 541 – 546.
5 Hunziker, W & Schrickel, S (1988) Rat brain calbindin D 28K : six domain structure and extensive amino acid homology with chicken calbindin D 28K Mol Endocrinol 2, 465 – 473.
6 Christakos, S., Rhoten, W.B & Feldman, S.C (1987) Rat calbindin
D 28K : purification, quantitation, immunocytochemical localization, and comparative aspects Methods Enzymol 139, 534 – 551.
7 Resibois, A., Blachier, F., Rogers, J.H., Lawson, D.E & Pochet, R (1990) Comparison between rat brain calbindin- and calretinin-immuno- reactivities Adv Exp Med Biol 269, 211 – 214.
8 Rogers, J.H & Resibois, A (1992) Calretinin and calbindin-D 28k in rat brain: patterns of partial co-localization Neuroscience 51,
843 – 865.
9 Winsky, L & Kuz´nicki, J (1995) Distribution of calretinin, calbindin D 28k , and parvalbumin in subcellular fractions of rat cerebellum: effects of calcium J Neurochem 65, 381 – 388.
10 Nemere, I., Leathers, V & Norman, A.W (1986) 1,25-Dihydroxy-vitamin D 3 -mediated intestinal calcium transport Biochemical identification of lysosomes containing calcium and calcium-binding protein (calbindin-D 28K ) J Biol Chem 261, 16106– 16114.
11 Fonseca, M & Soriano, E (1995) Calretinin-immunoreactive neurons in the normal human temporal cortex and in Alzheimer’s disease Brain Res 691, 83 – 91.
12 Jellinger, K.A (1999) Post mortem studies in Parkinson’s disease –
is it possible to detect brain areas for specific symptoms? J Neural Transm Suppl 56, 1 – 29.
13 Cicchetti, F., Prensa, L., Wu, Y & Parent, A (2000) Chemical anatomy of striatal interneurons in normal individuals and in patients with Huntington’s disease Brain Res Brain Res Rev 34,
80 – 101.
14 Schwaller, B., Meyer-Monard, S., Gander, J.C., Pugin, P., Celio, M.R & Ludwig, C (1998) The calcium-binding protein calretinin-22k is detectable in the serum and specific cells of cancer patients Anticancer Res 18, 3661 – 3667.
15 Ordonez, N.G (1999) The immunohistochemical diagnosis of epithelial mesothelioma Hum Pathol 30, 313 – 323.
16 Saydan, N., Salicio, V., Cappelli-Gotzos, B & Gotzos, V (2001) Expression of calretinin in human mesothelioma cell lines and cell cycle analysis by flow cytometry Anticancer Res 21, 181 – 188.
Trang 817 Katsetos, C.D., Jami, M.M., Krishna, L., Jackson, R., Patchefsky,
A.S & Cooper, H.S (1994) Novel immunohistochemical
localization of 28,000 molecular-weight (M r ) calcium binding
protein (calbindin-D 28k ) in enterochromaffin cells of the human
appendix and neuroendocrine tumors (carcinoids and small-cell
carcinomas) of the midgut and foregut Arch Pathol Lab Med.
118, 633 – 639.
18 Billing-Marczak, K & Kuz´nicki, J (1999) Calretinin – sensor or
buffer – function still unclear Pol J Pharmacol 51, 173 – 178.
19 McMahon, A., Wong, B.S., Iacopino, A.M., Ng, M.C., Chi, S &
German, D.C (1998) Calbindin-D 28k buffers intracellular calcium
and promotes resistance to degeneration in PC12 cells Brain Res.
Mol Brain Res 54, 56 – 63.
20 Leathers, V.L & Norman, A.W (1993) Evidence for calcium
mediated conformational changes in calbindin-D 28K (the vitamin
D -induced calcium binding protein) interactions with chick
intestinal brush border membrane alkaline phosphatase as studied
via photoaffinity labeling techniques J Cell Biochem 52,
243 – 252.
21 Bellido, T., Huening, M., Raval-Pandya, M., Manolagas, S.C &
Christakos, S (2000) Calbindin-D 28k Is expressed in osteoblastic
cells and suppresses their apoptosis by inhibiting caspase-3 activity.
J Biol Chem 275, 26328 – 26332.
22 Kuz´nicki, J., Winsky, L & Jacobowitz, D.M (1994) Ca(21
)-dependent and in)-dependent interactions of calretinin with
hydrophobic resins Biochem Mol Biol Int 33, 713 – 721.
23 Kuz´nicki, J., Strauss, K.I & Jacobowitz, D.M (1995)
Confor-mational changes and calcium binding by calretinin and its
recombinant fragments containing different sets of EF hand motifs.
Biochemistry 34, 15389– 15394.
24 Schwaller, B., Durussel, I., Jermann, D., Herrmann, B & Cox, J.
(1997) Comparison of the Ca21-binding properties of human
recombinant calretinin-22k and calretinin J Biol Chem 272,
29663– 29671.
25 Stevens, J & Rogers, J.H (1997) Chick calretinin: purification,
composition, and metal binding activity of native and recombinant
forms Prot Exp Purif 9, 171 – 181.
26 Gross, M.D., Gosnell, M., Tsarbopoulos, A & Hunziker, W (1993)
A functional and degenerate pair of EF hands contains the very high
affinity calcium-binding site of calbindin-D 28K J Biol Chem 268,
20917– 20922.
27 Kumar, R., Hunziker, W., Gross, M., Naylor, S., Londowski, J.M &
Schaefer, J (1994) The highly efficient production of full-length
and mutant rat brain calcium-binding proteins (calbindins-D 28K ) in
a bacterial expression system Arch Biochem Biophys 308,
311 – 317.
28 Veenstra, T.D., Gross, M.D., Hunziker, W & Kumar, R (1995)
Identification of metal-binding sites in rat brain calcium-binding
protein J Biol Chem 270, 30353– 30358.
29 A ˚ kerfeldt, K.S., Coyne, A.N., Wilk, R.R., Thulin, E & Linse, S.
(1996) Ca21-binding stoichiometry of calbindin D 28k as assessed
by spectroscopic analyses of synthetic peptide fragments.
Biochemistry 35, 3662 – 3669.
30 Berggard, T., Silow, M., Thulin, E & Linse, S (2000) Ca( 21 )- and
H( 1 )-dependent conformational changes of calbindin D(28k).
Biochemistry 39, 6864 – 6873.
31 Klaus, W., Grzesiek, S., Labhardt, A.M., Buchwald, P., Hunziker,
W., Gross, M.D & Kallick, D.A (1999) NMR investigation and
secondary structure of domains I and II of rat brain calbindin D 28k
(1 – 93) Eur J Biochem 262, 933 – 938.
32 Palczewska, M., Groves, P & Kuz´nicki, J (1999) Use of Pichia
pastoris for the expression, purification and characterization of rat
calretinin ‘EF-hand’ domains Prot Exp Purif 17, 465 – 476.
33 Strauss, K.I., Kuz´nicki, J., Winsky, L & Jacobowitz, D.M.
(1994) Expression and rapid purification of recombinant rat
calretinin: similarity to native rat calretinin Prot Exp Purif 5,
187 – 191.
34 Bradford, M (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding Anal Biochem 72, 248 – 254.
35 Cavanagh, J., Fairbrother, W.J., Palmer, A.G & Skelton, N.J (1996) Protein NMR Spectroscopy: Principles and Practice, 1st edn Academic Press, San Diego, CA.
36 Sattler, M., Schleucher, J & Griesinger, C (1999) Heteronuclear multidimensional NMR experiments for the structure determi-nation of proteins in solution employing pulsed field gradients Prog Nucl Magn Reson Spectrosc 34, 93 – 158.
37 Zimmerman, D.E., Kulikowski, C.A., Huang, Y.P., Feng, W.Q., Tashiro, M., Shimotakahara, S., Chien, C.Y., Powers, R & Montelione, G.T (1997) Automated analysis of protein NMR assignments using methods from artificial intelligence J Mol Biol.
269, 592 – 610.
38 Goddard, T.D & Kneller, D.G (1989 – 2000) SPARKY 3 University of California Regents, University of California, San Francisco, CA.
39 Kuz´nicki, J., Wang, T.L., Martin, B.M., Winsky, L & Jacobowitz, D.M (1995) Localization of Ca( 21 )-dependent conformational changes of calretinin by limited tryptic proteolysis Biochem J.
308, 607 – 612.
40 Schagger, H & von Jagow, G (1987) Tricine-sodium dodecyl sulfate-polyacrylamide gel electrophoresis for the separation of proteins in the range from 1 to 100 kDa Anal Biochem 166,
368 – 379.
41 Maruyama, K., Mikawa, T & Ebashi, S (1984) Detection of calcium binding proteins by 45 Ca autoradiography on nitrocellulose membrane after sodium dodecyl sulfate gel electrophoresis.
J Biochem (Tokyo) 95, 511 – 519.
42 Farrow, N.A., Muhandiram, R., Singer, A.U., Pascal, S.M., Kay, C.M., Gish, G., Shoelson, S.E., Pawson, T., Forman-Kay, J.D & Kay, L.E (1994) Backbone dynamics of a free and phosphopep-tide-complexed Src homology 2 domain studied by 15N NMR relaxation Biochemistry 33, 5984 – 6003.
43 Johnson,W.H Jr, Czerwinski, R.M., Stamps, S.L & Whitman, C.P (1999) A kinetic and stereochemical investigation of the role of lysine-32 in the phenylpyruvate tautomerase activity catalyzed by macrophage migration inhibitory factor Biochemistry 38, 16024– 16033.
44 Morris, G & Barjat, H (1997) Methods for Structure Elucidation
by High Resolution NMR (Batta, G., Kover, K.E & Szantay, C Jr, eds) Elsevier, Amsterdam.
45 Altieri, A.S., Hinton, D.P & Byrd, R.A (1995) Association of biomolecular systems via pulsed field gradient NMR self-diffusion measurements J Am Chem Soc 117, 7566 – 7567.
46 Staros, J.V., Wright, R.W & Swingle, D.M (1986) Enhancement
by N-hydroxysulfosuccinimide of water-soluble carbodiimide-mediated coupling reactions Anal Biochem 156, 220 – 222.
47 Lanigan, M.D., Tudor, J.E., Pennington, M.W & Norton, R.S (2001) A helical capping motif in ShK toxin and its role in helix stabilization Biopolymers 58, 422 – 436.
48 Aurora, R., Creamer, T.P., Srinivasan, R & Rose, G.D (1997) Local interactions in protein folding: lessons from the alpha-helix.
J Biol Chem 272, 1413 – 1416.
49 Prieto, J & Serrano, L (1997) C-capping and helix stability: the Pro C-capping motif J Mol Biol 274, 276 – 288.
50 Aitio, H., Annila, A., Heikkinen, S., Thulin, E., Drakenberg, T & Kilpelainen, I (1999) NMR assignments, secondary structure, and global fold of calerythrin, an EF-hand calcium-binding protein from Saccharopolyspora erythraea Protein Sci 8,
2580 – 2588.
51 Kragelund, B.B., Hauenschild, A., Carlstrom, G., Pongs, O & Finn, B.E (2000) 1 H, 13 C, and 15 N assignments of un-myristoy-lated Ca21-frequenin, a synaptic efficacy modulator J Biomol NMR 16, 85 – 86.
52 Ames, J.B., Hendricks, K.B., Strahl, T., Huttner, I.G., Hamasaki, N.
Trang 9& Thorner, J (2000) Structure and calcium-binding properties of
Frq1, a novel calcium sensor in the yeast Saccharomyces
cerevisiae Biochemistry 39, 12149 – 12161.
53 Ames, J.B., Dizhoor, A.M., Ikura, M., Palczewski, K & Stryer, L.
(1999) Three-dimensional structure of guanylyl cyclase activating
protein-2, a calcium-sensitive modulator of photoreceptor guanylyl
cyclases J Biol Chem 274, 19329– 19337.
54 Dizhoor, A.M & Hurley, J.B (1996) Inactivation of EF-hands
makes GCAP-2 (p24) a constitutive activator of photoreceptor
guanylyl cyclase by preventing a Ca21-induced
‘activator-to-inhibitor’ transition J Biol Chem 271, 19346– 19350.
55 Craescu, C.T., Precheur, B., van Riel, A., Sakamoto, H., Cox, J.A.
& Engelborghs, Y (1998) 1 H and 15 N resonance assignment of the
calcium-bound form of the Nereis diversicolor sarcoplasmic
Ca( 21 )-binding protein J Biomol NMR 12, 565 – 566.
56 Vijay-Kumar, S & Cook, W.J (1992) Structure of a sarcoplasmic
calcium-binding protein from Nereis diversicolor refined at 2.0 A˚
resolution J Mol Biol 224, 413 – 426.
57 Lipari, G & Szabo, A (1982) Model-free approach to the
interpretation of nuclear magnetic resonance relaxation in
macromolecules 1 Theory and range of validity J Am Chem.
Soc 104, 4546 – 4559.
58 Wagner, L., Oliyarnyk, O., Gartner, W., Nowotny, P., Groeger, M.,
Kaserer, K., Waldhausl, W & Pasternack, M.S (2000) Cloning and
expression of secretagogin, a novel neuroendocrine- and pancreatic islet of Langerhans-specific Ca21-binding protein J Biol Chem.
275, 24740 – 24751.
59 Linse, S., Thulin, E., Gifford, L.K., Radzewsky, D., Hagan, J., Wilk, R.R & A ˚ kerfeldt, K.S (1997) Domain organization of calbindin D 28k as determined from the association of six synthetic EF-hand fragments Protein Sci 6, 2385 – 2396.
60 Kretsinger, R.H & Nockolds, C.E (1973) Carp muscle calcium-binding protein II Structure determination and general descrip-tion J Biol Chem 248, 3313 – 3326.
61 Wishart, D.S & Sykes, B.D (1994) The13C chemical-shift index: a simple method for the identification of protein secondary structure using 13 C chemical-shift data J Biomol NMR 4, 171 – 180.
62 Feher, V.A., Zapf, J.W., Hoch, J.A., Whiteley, J.M., McIntosh, L.P., Rance, M., Skelton, N.J., Dahlquist, F.W & Cavanagh, J (1997) High-resolution NMR structure and backbone dynamics of the Bacillus subtilis response regulator, Spo0F: implications for phosphorylation and molecular recognition Biochemistry 36, 10015– 10025.
63 el-Joubary, A., Bruix, M., Santoro, J., Cafaro, V., Scognamiglio, R.,
Di Donato, A., D’Alessio, G., Ko¨ve´r, K.E., Batta, G., Szilagyi, L & Rico, M (1999)1H and15N sequential assignment and solution secondary structure of15N labelled human pancreatic ribonuclease.
J Biomol NMR 15, 265 – 266.