Co-operation of domain-binding and calcium-binding sites
in the activation of gelsolin
Emeline Lagarrigue
1
, Sutherland K. Maciver
2
, Abdellatif Fattoum
3
, Yves Benyamin
1
and Claude Roustan
1
1
UMR 5539 (CNRS) Laboratoire de motilite
´
cellulaire (Ecole Pratique des Hautes Etudes), Universite
´
de Montpellier 2, France;
2
Genes and Development Group, Department of Biomedical Sciences, University of Edinburgh, Scotland, UK;
3
Centre de Recherches de Biochimie Macromole
´
culaire, UPR 1086 (CNRS), Montpellier, France
Gelsolin is an abundant calcium dependent actin filament
severing and capping protein. In the absence of calcium the
molecule is compact but in the presence of calcium, as its six
similar domains alter their relative position, a generally more
open configuration is adopted to reveal the three actin
binding sites. It is generally held that a Ôhelical-latchÕ at the
C-terminus of gelsolin’s domain 6 (G6), binds domain 2 (G2)

domains; six in the case of gelsolin itself, adseverin and
villin, and three in capG, fragmin and severin (for a review,
see [2]). The binding of calcium to gelsolin and to actin
bound gelsolin is complex. Free gelsolin binds at least six
calcium ions. These sites, coordinated solely by gelsolin have
been termed type II [3]. The affinity of type II sites varies
greatly. High affinity calcium sites (K
d
 1 l
M
) have been
identified [4–6] and two of these have been localized within
G4–6 [7]. A body of evidence suggests that calcium binding
by G4–6 affords calcium-sensitivity to the whole gelsolin
molecule [8,9]. Sites have been identified by biochemical
means within G4-5 (K
d
 2 l
M
) and G5–6 (K
d
 0.2 l
M
)
[9], and crystallographic studies (S. Kolappan, J. Gooch,
A. Weeds & P. McLaughlin, Wellcome Centre for Cell
Biology, University of Edinburgh, UK, personal commu-
nication, [3]) have shown that calcium ions are bound by
both G5 and G6 (sites IIG5 and IIG6 [3]. Low affinity
calcium-binding sites (K

fluorescein 5-isothiocyanate; 1,5-I-AEDANS, N-iodoacetyl-N¢-(sulfo-
1-naphthyl)-ethylenediamine; BACNHS, biotinamidocaproate
N-hydroxyl-succinimide ester; G-actin, monomeric actin;
F-actin, filamentous actin.
Note: we have adopted the labeling system introduced by Choe et al.
[Choe, H., Burtnick, L.D., Mejillano, M., Yin, H.L., Robinson, R.C.
& Choe, S. (2002) J. Mol. Evol. 324, 691–702.] for the various Calcium-
binding sites so that IG1 is the type I binding site within G2 and IIG6 is
the type II binding site within G6. Type I binding sites are coordinated
by gelsolin and actin whereas type II sites are coordinated solely by
gelsolin residues.
Note: webpages are available at />umr5539/, and
/>(Received 20 December 2002, revised 10 March 2003,
accepted 26 March 2003)
Eur. J. Biochem. 270, 2236–2243 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03591.x
30 l
M
[19], and the most rapid severing rate of actin
filaments occurs at 300 l
M
[20]. Accumulating evidence
[3,12,18,21] suggests the following mechanism for the
activation of gelsolin by calcium. Low calcium concentra-
tions are proposed [18,21,22] to ÔunlatchÕ the connection
between G2 and G6, but higher concentrations are required
to break salt bridges between other domains until the
gelsolin is fully ÔopenÕ, then additional calcium ions are
requiredtobindactintoG1andtoG4forseveringand
capping. The details of the latch helix structure in the
presence of calcium are not clear as there is no density for

pH 7.5 supplemented with either 1 m
M
CaCl
2
or 1 m
M
EGTA in the presence of trypsin (9 lg/mL). The proteolysis
was stopped by addition of an antiprotease mixture
(Complete) (1 : 25 mass/vol.) purchased from Roche SA
(Mannheim, Germany). To identify the cleavage products,
the digest was labeled at cysteine residues by treatment with
1,5-I-AEDANS (10 molar excess) [27] for 4 h at room
temperature, in the presence of 1% SDS. The labeling
reaction was stopped by addition of 0.1
M
b-mercapto-
ethanol. The digest was then analyzed by SDS/PAGE.
Antibodies directed towards gelsolin G4–6 domain were
elicited in rabbits according to [28]. Anti-IgG antibodies
labeled with alkaline phosphatase were purchased from
Sigma (Dorset, UK).
Synthetic peptides derived from gelsolin sequences 159–
193 and 203–225 [17] were prepared on solid phase support
using a 9050 Milligen PepSynthesizer (Millipore) according
to the Fmoc/tBu system. The crude peptides were
de-protected and thoroughly purified by preparative
reverse-phase HPLC. The purified peptides were shown to
be homogenous by analytical HPLC. Electrospray mass
spectra, carried out in the positive ion mode using a Trio
2000 VG Biotech Mass spectrometer (Altrincham), were in

NaCl, 0.05%
tween 20, 10 m
M
Phosphate buffer, pH 7.4. Experiments
with coated fragments were performed in 0.1
M
KCl, 20 m
M
Tris pH 7.2. Binding was monitored at 405 nm using
alkaline phosphatase-labeled anti-IgG antibodies (dilution
1/1000) or alkaline phosphatase labeled streptavidin (dilu-
tion 1/1000). Control assays were carried out in wells
saturated with gelatin and gelatin hydrolysate used alone.
Each assay was conducted in triplicate and the mean value
plotted after subtraction of nonspecific absorption. The
binding parameters (apparent dissociation constant K
d
and
the maximal binding A
max
) were determined by nonlinear
fitting A ¼ A
max
· [L]/(K
d
+[L])(relation1)whereA is
the absorbance at 405 nm and [L] the ligand concentration,
by using the
CURVE FIT
software developed by Kevin Raner

d
)
2
) 4[E] · [L])
0.5
}where
[E] is the concentration of the fluorescent protein. The
maximum fluorescence change (A
max
) at infinite substrate
concentration expressed as percentage variation from initial
fluorescence: F1 – F°/F°
*
100 was calculated by the
relation F 1 – F°/F° ¼ A
max
/F° where F° and F1 are
fluorescence intensities for zero and infinite ligand concen-
trations, respectively.
Analytical methods
Protein concentrations were determined by UV absorbency
using a Varian MS 100 spectrophotometer. Gelsolin
domain concentrations were determined spectrophotomet-
rically using values of A
280
(1 cm
)1
) ¼ 15.5 l
M
for G4–6,

16 kDa band is fluorescent. This result demonstrates that
the cleavage occurs in the loop between G5 and G6 domain
and is in accord with the unpublished results reported
previously [19].
As depicted in Fig. 1B, the tryptic cleavage was faster in
EGTA than in calcium, suggesting that the orientation
between the two domains is different and the junction more
accessible in EGTA.
Calcium induced change in the G4–6 and G2 domains
Conformation changes induced by calcium binding were
monitored by two approaches. First, intrinsic tryptophan
fluorescence of G4–6 domain was measured in the presence
of increasing calcium concentrations (between 1 n
M
and
1m
M
). An increase in fluorescence intensity was observed at
submicromolar calcium concentrations (Fig. 2). In a second
experiment, conformational changes were detected from the
extrinsic fluorescence measurements of FITC-labeled G4–6
domain. A biphasic relationship was observed (Fig. 2)
producing two transitions in fluorescence intensity, one at
 1.5 l
M
, and another around 0.1 l
M
. These transitions
reflecting conformational changes correlate well with the two
binding sites (IIG5 and IIG6, K

]). Calcium concentrations
were determined experimentally (see Material and methods). Inset:
effect of calcium on the G2 tryptophan fluorescence emission. Fluo-
rescence changes were plotted vs. free calcium concentrations.
Fig. 1. Effect of calcium on susceptibility of gelsolin G4–6 domain to
tryptic digestion. (A) Identification of the two fragments (30 kDa and
16 kDa) produced by proteolysis in the presence of 1 m
M
EGTA
followed by 1,5-I-AEDANS labeling as described in Material and
methods. Molecular weight markers (lane T). G4–6 digest chemically
modified by IEADANS and revealed with Coomassie blue (lane 1) or
upon UV lamp (lane 2). (B) Digestion of G4–6 domain by trypsin
(trypsin/G4–6 ratio: 1/25 w/w) for 10 min molecular weight markers
(lane T). G4–6 domain (46 kDa) before proteolysis (lane 1). G4–6
digest in the presence of 0.1 m
M
calcium (lane 2) and in the presence of
0.1 m
M
EGTA (lane 3). Molecular weight marker are phosphory-
lase B (97.4 kDa), bovine serum albumin (66.2 kDa), ovalbumin
(45 kDa), carbonic anhydrase (31 kDa), soybean trypsin inhibitor
(21.5 kDa) and lysozyme (14.4 kDa).
2238 E. Lagarrigue et al. (Eur. J. Biochem. 270) Ó FEBS 2003
sequence 203–225, including one of the actin-binding sites, is
in interaction with the C-terminal a-helix of G6 domain.
The sequence 159–193, including the second actin interface,
appears also in interaction with G6 domain. Therefore, we
tested the interaction of G4–6 domain with G2 domain by

changes in fluorescence were monitored. Saturation curves
were observed in the presence of EGTA or calcium and
apparent K
d
s of 0.8 and 3.5 l
M
, respectively, were deter-
mined (Fig. 5B and Table 1). These values and those
obtained above from ELISA show a better interaction of
G2 in the presence of EGTA than calcium. The differences
in the absolute values observed between the two methods
are likely to be due to the heterogeneous phases used in
ELISA. More interestingly, a significant difference in the
maximum fluorescence enhancement determined in EGTA
and calcium (about 8% and 30%, respectively) possibly
reflects changes in domain conformation and in interfaces
produced under these two conditions. In order to deduce a
more precise relationship between G2 interaction and G4–6
calcium-induced conformation, the maximum fluorescence
enhancement extrapolated for infinite G2 concentrations
was determined at various calcium concentrations. As
shown in Fig. 6, a change in the fluorescence from EGTA to
calcium states was observed in the 0.1 l
M
range. This
important result suggests that the transition would be linked
to the high affinity calcium site in G4–6 [7], that we now
know to be IIG6 [3] (see Discussion). In addition, apparent
K
d

the concentrations indicated. Binding was monitored at 405 nm using
alkaline phosphatase labeled streptavidin. Percentage binding was
plotted vs. G2 concentrations.
Ó FEBS 2003 Calcium activation of gelsolin (Eur. J. Biochem. 270) 2239
show that the two peptides interacted with G4–6 in the
presence of EGTA. In contrast, only the 153–193 peptide
interacted in the presence of calcium (Table 1). These results
were confirmed in solution. For this aim, FITC labeled
G4–6 was mixed with each peptide supplemented with
EGTA or calcium. In the presence of EGTA, changes in the
fluorescence intensity (Fig. 7A,B) were obtained with the
two fragments. A maximum fluorescence quenching of 4%
and fluorescence enhancement of 4% were calculated for
159–193 and 203–225 peptides, respectively. When calcium
is present in the medium, binding of peptide 153–193 to
labeled G4–6 induces an important decrease in fluorescence
intensity (15%) (Fig. 7). The last result demonstrates that
the binding of the latter peptide causes a somewhat different
conformational change in G4–6 domain. In contrast no
effect is observed for the 203–225 fragment, accordingly to
ELISA experiments.
Discussion
Since its discovery as an actin-depolymerizing factor, gelsolin
has now been implicated in a number of important pathways
such as apoptosis, oncogenic transformation, signal trans-
duction and amyloidosis (reviewed [2]). All of these
pathways are likely to involve calcium activation, the process
by which various binding sites become (especially actin)
available for interaction. In this paper we confirm that
calcium causes a large conformational change in the

labeled G4–6 (0.3 l
M
) determined by fluorescence. The experiment
was carried out in 0.1
M
KCl, 20 m
M
Tris buffer pH 7.4 supplemented
with 1 m
M
EGTA (d)or1m
M
CaCl
2
(s).
Fig. 6. Effect of calcium on the fluorescence of the FITC labeled G4–6/
G2 complex. Maximum fluorescence enhancement (% initial fluores-
cence) extrapolated to infinite G2 concentration is plotted vs. free
calcium concentration (pCa). Inset, apparent K
d
s for G2. Interactions
with FITC labeled G4–6 are plotted vs. pCa.
Table 1. Binding of G2 and derived peptide to G4–6.
EGTA K
d
(l
M
) Calcium K
d
(l

moderate affinity (K
d
¼ 2 l
M
) in agreement of this study
and previous work [7], the other IIG6 has a higher affinity.
The slight difference observed between the present study
(K
d
¼ 0.1 l
M
) and the previous study [7] who measured a
K
d
of 0.2 l
M
, may be attributable to conformational
differences in the site in the context of G4–6 compared to
G5–6 or a degree of cocooperativity between calcium site
(see below).
A calcium-dependent conformational change in G2
A number of studies have concluded that calcium induces
conformational changes within G2 [11,36]. A calcium
binding site has been detected in G1–3 [12], and others
[3,11] have located this site within G2 itself. However, this
previous [11] study estimated a low affinity (K
d
¼32 l
M
)

is only marginally calcium sensitive and of higher affinity
(Table 1). The EGTA structure [22] reveals that this peptide
makes salt bridge contacts (R168–D669 and R169–D670)
and that hydrophobic interactions also occur such as those
between V170-V657, with G6. In summary, G2 binds to
G4–6 through two distinct interfaces. Binding site 1 involves
G2 203–225 and G6 744–755 and binding site 2 involves
G2 region around R168-R69 and G6 D669-D670.
‘Unlatching’ and dissociation of the G2 and G6
connection
If the last 23 amino-acids are removed from gelsolin
mutants, the requirement for calcium for actin-binding
is lessened but not abolished [39], similarly it has been
determined that adseverin which naturally lacks the
C-terminal helix has a similar calcium requirement than
the helix minus gelsolin mutant [21]. Together with recent
observations on the structure of gelsolin [3,36], our new
data on binding site 2 is compatible with the following
Fig. 7. Effect of calcium on the interaction of two sequences involved in the G2–G4–6 interface monitored by fluorescence. Binding of FITC labeled
G4–6 (0.3 l
M
) with two synthetic peptides derived from G2 domain: (A) sequence 203–225 and (B), sequence 159–193. The experiments were
performed in 0.1
M
KCl, 20 m
M
Tris buffer pH 7.4 supplemented with 1 m
M
EGTA (d)or1m
M

IIG2 so that it too becomes a high affinity site. We
measure the IIG6 K
d
to be 100 n
M
(Fig. 2) in agreement
with this model.
Acknowledgements
This research was supported by grants from AFM.
References
1. Sun, H.Q., Yamamoto, M., Mejillano, M. & Yin, H.L. (1999)
Gelsolin, a multifunctional actin regulatory protein. J. Biol. Chem.
274, 33179–33182.
2. Kwiatkowski, D.J. (1999) Functions of geloslin: motility, signa-
ling, apoptosis, cancer. Curr. Op. Cell Biol. 11, 103–108.
3. Choe, H., Burtnick, L.D., Mejillano, M., Yin, H.L., Robinson,
R.C. & Choe, S. (2002) The calcium activation of gelsolin: Insights
from the 3A structure of the G4–G6/actin complex. J. Mol. Evol.
324, 691–702.
4. Yin, H.L. & Stossel, T.P. (1980) Purification and structural
properties of gelsolin, a calcium-activated regulator protein of
macrophages. J. Biol. Chem. 255, 9490–9493.
5. Bryan, J. (1988) Gelsolin has three actin-binding sites. J. Cell Biol.
106, 1553–1562.
6. Weeds,A.G.,Gooch,J.,Pope,B.P.&Harris,H.E.(1986)Pre-
paration and characterization of pig plasma and platelet gelsolins.
Eur. J. Biochem. 161, 69–76.
7. Pope, B., Maciver, S. & Weeds, A.G. (1995) Localization of the
calcium-sensitive actin monomer binding site in gelsolin to seg-
ment 4 and identification of calcium-binding sites. Biochemistry

serum gelsolin actin complex. Biochem. J. 293, 813–817.
17. Renoult, C., Blondin, L., Fattoum, A., Ternent, D., Maciver,
S.K., Raynaud, F., Benyamin, Y. & Roustan, C. (2001) Binding of
gelsolin domain 2 to actin. An actin interface distinct from that of
gelsolin domain 1 and from ADF/cofilin. Eur. J. Biochem. 268,
6165–6175.
18. Lin, K M., Mejillano, M. & Yin, H.L. (2000) Ca
2+
regulation of
gelsolin by its C-terminal tail. J. Biol. Chem. 275, 27746–27752.
19. Khaitlina, S. & Hinssen, H. (2002) Ca-dependent binding of actin
to gelsolin. FEBS Lett. 521, 14–28.
20. Kinosian, H.J., Newman, J., Lincoln, B., Selden, L.A., Gershman,
L.C. & Estes, J.E. (1998) Ca
2+
regulation of gelsolin activity:
binding and severing of F-actin. Biophys. J. 75, 3101–3109.
21. Lueck, A., Brown, D. & Kwiatkowski, D.J. (1998) The actin-
binding proteins adseverin and gelsolin are both hightly expressed
but differentially localized in kidney and intestine. J. Cell Sci. 111,
3633–3643.
22. Burtnick, L.D., Koepf, E.K., Grimes, J., Jones, E.Y., Stuart, D.I.,
McLaughlin, P.J. & Robinson, R.C. (1997) The crystal structure
of plasma gelsolin: implications for actin severing, capping and
nucleation. Cell 90, 661–670.
23. Way, M., Pope, B., Gooch, J., Hawkins, M. & Weeds, A.G. (1990)
Identification of a region of segment 1 of gelsolin critical for actin
binding. EMBO J. 9, 4103–4109.
24. Way, M., Pope, B. & Weeds, A.G. (1992) Evidence for functional
homology in the F-actin binding domains of gelsolin and alpha-

Methods Enzymol. 70, 419–439.
2242 E. Lagarrigue et al. (Eur. J. Biochem. 270) Ó FEBS 2003
32. Me
´
jean,C.,Lebart,M.C.,Poyer,M.,Roustan,C.&Benyamin,Y.
(1992) Localization and identification of actin structures involved
in the filamin-actin interactio. Eur. J. Biochem. 209, 555–562.
33. Gill, S.C. & von Hippel, P.H. (1989) Calculation of protein
extinction coefficients from amino acid sequence data. Anal. Bio-
chem. 182, 319–326.
34. Laemmli, U.K. (1970) Cleavage of structural proteins during the
assembly of the head of bacteriophage T4. Nature 227, 680–685.
35. Grynkeiwicz, G., Poenie, M. & Tsien, R.Y. (1985) A new gen-
eration of calcium indicators with greatly improved fluorescence
properties. J. Biol. Chem. 260, 3440–3450.
36. Kazmirski, S.L., Isaacson, R.L., An, C., Buckle, A., Johnson,
C.M., Daggett, V. & Fersht, A.R. (2002) Loss of a metal-binding
site in gelsolin leads to familial amyloidosis-Finnish type. Nature
Struct. Biol. 9, 112–116.
37. Kilhoffer, M C. & Ge
´
rard, D. (1985) Fluorescence study of bre-
vin, the Mr 92,000 actin-capping and -fragmenting protein isolated
form serum. Effect of Ca
2+
on protein conformation. Biochem-
istry 24, 5653–5660.
38. Rouayrenc, J P., Fattoum, A., Me
´
gean, C. & Kassab, R. (1986)