Synaptotagmin-1 and -7 are functionally overlappingCa2� sensors for exocytosis in adrenalchromaffin cellsJean-Sebastien Schonn*†, Anton Maximov‡§¶�, Ye Lao‡§¶, Thomas C. Sudhof‡§¶**, and Jakob B. Sørensen*,**

*Max Planck Institute for Biophysical Chemistry, Am Fassberg 11, D-37077 Gottingen, Germany; and Departments of ‡Neuroscience and §Molecular Geneticsand ¶Howard Hughes Medical Institute, University of Texas Southwestern Medical Center, Dallas, TX 75390

Contributed by Thomas C. Sudhof, December 31, 2007 (sent for review November 29, 2007)

Synaptotagmin-1, the canonical isoform of the synaptotagminfamily, is a Ca2� sensor for fast synchronous neurotransmitterrelease in forebrain neurons and chromaffin cells. Even thoughdeletion of synaptotagmin-1 abolishes fast exocytosis in chromaf-fin cells, it reduces overall secretion by only 20% because of thepersistence of slow exocytosis. Therefore, another Ca2� sensordominates release in these cells. Synaptotagmin-7 has a higherCa2� affinity and slower binding kinetics than synaptotagmin-1,matching the proposed properties for the second, slower Ca2�

sensor. Here, we examined Ca2�-triggered exocytosis in chromaf-fin cells from KO mice lacking synaptotagmin-7, and from knockinmice containing normal levels of a mutant synaptotagmin-7 whoseC2B domain does not bind Ca2�. In both types of mutant chromaffincells, Ca2�-triggered exocytosis was decreased dramatically. More-over, in chromaffin cells lacking both synaptotagmin-1 and -7, onlya very slow release component, accounting for �30% of WTexocytosis, persisted. These data establish synaptotagmin-7 as amajor Ca2� sensor for exocytosis in chromaffin cells, which, to-gether with synaptotagmin-1, mediates almost all of the Ca2�

triggering of exocytosis in these cells, a surprising result, consid-ering the lack of a role of synaptotagmin-7 in synaptic vesicleexocytosis.

amperometry � calcium-binding protein � capacitance �dense core vesicle � fusion

Release of neurotransmitters and hormones is triggered byCa2� influx, which leads to exocytosis of synaptic or secre-

tory vesicles. Synaptotagmins constitute a family of proteinscontaining a single transmembrane region and two Ca2�- andphospholipid-binding C2 domains. Synaptotagmin-1, the proto-typical synaptotagmin, functions as a vesicular Ca2� sensor forfast exocytosis in forebrain neurons and chromaffin cells (1–5).In the absence of synaptotagmin-1, the fastest phase of exocy-tosis is abolished; however, a slower form of exocytosis persists(3, 5–7). In chromaffin cells, this slower form seems to representrelease from a distinct ‘‘slowly releasable’’ pool (SRP) (or state)of vesicles that have not yet reached the ‘‘readily releasable’’state (readily releasable pool or RRP) (3, 8). In fact, exocytosisin chromaffin cells occurs predominantly via the slower pathwaywhen triggered by a step elevation of [Ca2�]i. As a result, deletionof synaptotagmin-1 leads to only a �20% reduction in theamount of exocytosis measured after 5 s (3, 8).

The mild effect of the synaptotagmin-1 deletion on overallexocytosis in chromaffin cells raises important questions. Thisobservation supports the notion that synaptotagmins might notbe important for regulating overall release but only for timing therelease relative to the Ca2� trigger (6). An alternative view holdsthat synaptotagmins do actively trigger release but that there isa second synaptotagmin Ca2� sensor in chromaffin cells that isresponsible for slower forms of exocytosis (1). Of the 15 mam-malian synaptotagmins, only 8 display Ca2�-dependent lipidbinding (synaptotagmin-1, -2, -3, -5, -6, -7, -9, and -10). Recentstudies have identified synaptotagmin-1, -2, and -9 as alternative

and parallel Ca2� sensors in synaptic exocytosis (9), and synap-totagmin-1 and -9 have been shown to account for essentially allrelease in PC12 cells, a phaeocytochroma cell line derived fromthe adrenal medulla (10, 11). However, synaptotagmin-2 and -9are not expressed in chromaffin cells (9, 12, 13), and even thoughkinetic differences between synaptotagmin-1, -2 and -9 isoformsare detectable (8, 9), they are all relatively fast acting isoforms(14) and thus cannot explain radically slower forms of exocytosis.Thus, the identity of the slow Ca2� sensor, which is crucial forexocytosis in chromaffin cells, remains obscure.

Synaptotagmin-7 is unusual in that it is ubiquitously expressedearly in development but restricted to secretory cells withCa2�-regulated exocytosis in the first postnatal weeks (15). As apotential Ca2� sensor for slower forms of exocytosis, synapto-tagmin-7 is of special interest because it shows maximal Ca2�-dependent phospholipid binding in the low micromolar Ca2�-range [�3 �M; (16)], which fits well with the higher apparentCa2� affinities of slower exocytosis in chromaffin cells (17). Inaddition, synaptotagmin-7 displays slower lipid-unbinding kinet-ics than synaptotagmin-1 (14) and can confer Ca2� dependenceon SNARE-dependent vesicle fusion (18), consistent with afunction as a slow sensor.

Several studies suggested a role for synaptotagmin-7 in Ca2�-dependent exocytosis of secretory vesicles and lysosomes (19,20), but results are contradictory as to whether synaptotagmin-7is required for lysosome exocytosis per se (21, 22) or forrestricting fusion pore expansion (23). In PC12-cells, overex-pressed exogenous synaptotagmin-7 can stimulate exocytosis(24–26), but the function of endogenously expressed synapto-tagmin-7 in these cells is questioned by the finding that synap-totagmin-1 and synaptotagmin-9 may account for all exocytosis(11). Moreover, the question whether Ca2� binding to endoge-nous synaptotagmin-7 is necessary for its function, as would beexpected for a protein acting as a Ca2� sensor, has not beenaddressed. Furthermore, the finding that synaptotagmin-7 is notrequired for any form of synaptic vesicle exocytosis (27), despiteits synaptic localization (19) and the effect of its overexpressionon synaptic vesicle recycling (28), raises doubts about thepossible Ca2�-sensor function of synaptotagmin-7 in exocytosis.Thus, we have in the present study examined the possible role of

Author contributions: J.-S.S., A.M., T.C.S., and J.B.S. designed research; J.-S.S., A.M., and Y.L.performed research; J.-S.S., A.M., T.C.S., and J.B.S. analyzed data; and T.C.S. and J.B.S. wrotethe paper.

The authors declare no conflict of interest.

†Present address: Institut de Biologie Physico-Chimique, Centre National de la RechercheScientifique UPR 1929, F-75005 Paris, France.

�Present address: Department of Cell Biology, Scripps Research Institute, 10550 NorthTorrey Pines Road, La Jolla, CA 92037.

**To whom correspondence may be addressed. E-mail: thomas.sudhof@utsouthwestern.edu or jsoeren@gwdg.de.

This article contains supporting information online at www.pnas.org/cgi/content/full/0712373105/DC1.

© 2008 by The National Academy of Sciences of the USA

3998–4003 � PNAS � March 11, 2008 � vol. 105 � no. 10 www.pnas.org�cgi�doi�10.1073�pnas.0712373105

Dow

nloa

ded

by g

uest

on

Aug

ust 1

, 202

1

synaptotagmin-7 in chromaffin granule exocytosis. Our resultsunequivocally establish that synaptotagmin-7 is an essentialCa2�-sensor for this type of exocytosis, revealing an unexpectedspecialization of different synaptotagmin isoforms for differenttypes of exocytosis.

ResultsImmunoblotting results indicated that chromaffin cells in miceexpress synaptotagmin-1 and -7, but appear to lack synaptotag-min-2, and -9 (Fig. 1A) and (9, 29). Thus, we analyzed Ca2�-triggered exocytosis in chromaffin cells from mutant mice thatlack synaptotagmin-7 (KO), either alone or in conjunction witha deletion of synaptotagmin-1, or that contain normal levels ofa mutant synaptotagmin-7 containing amino acid substitutions inits C2B-domain that prevent Ca2�-binding to this domain (27).

Depolarization-Induced Release Is Moderately Impaired by Deletion ofSynaptotagmin-7. We first investigated basic properties of exocy-tosis in chromaffin cells obtained from young adult (1–2-monthsof age) synaptotagmin-7 KO (�/�) mice and their WT (WT,�/�) littermates. After establishment of a whole-cell recordingconfiguration, chromaffin cells were stimulated by a train ofdepolarization pulses, and exocytosis was monitored by mem-brane capacitance measurements and amperometry (Fig. 1B). Insynaptotagmin-7-deficient cells, exocytosis was moderately de-pressed when assayed by capacitance measurements (�25%; P �0.055, Mann–Whitney test), and more dramatically depressedwhen measured by amperometry (�54%, P � 0.007). Measure-ments of the Na� and Ca2� currents did not reveal significantdifferences between synaptotagmin-7 KO and WT mice (Fig. 1Cand data not shown). However, Ca2� measurements indicatedthat the [Ca2�]i concentration was elevated to slightly higherlevels in KO cells (P � 0.05), which could have obscured theeffect of the deletion of synaptotagmin-7.

Deletion of Synaptotagmin-7 Affects the Extent and Kinetics ofExocytosis. In a second round of experiments, we elicited exocy-tosis by flash photolysis of caged Ca2�, which causes a step-like,homogeneous increase of [Ca2�]i. In WT cells, this stimulationprotocol resulted in a fast increase in capacitance (the secretoryburst), which corresponds to the fusion of two pools of releasablevesicles, the slowly and the readily releasable pool (Fig. 2A). A

later, much slower part of the capacitance increase—the sus-tained component—is triggered by release of granules that arebeing primed during the stimulation period. The time integral ofthe amperometric current generally mirrored the capacitancesignal, except for a diffusional delay (Fig. 2 A Lower).

We found that in synaptotagmin-7-deficient cells, both the exo-cytotic burst and the sustained component were reduced (Fig. 2 Aand D–E) by �50%. Importantly, postflash [Ca2�]i were closelymatched between KO and WT cells (Fig. 2A Top), which explainsthe clearer phenotype of the KO during these measurements. Therelationship between the cubic root of the amperometric charge andthe square root of the capacitance change was approximately linearin both cases (KO and WT) with tightly overlapping regression lines(Fig. 2B). This finding indicates no difference in the releasedamount of catecholamines per unit area of vesicle membrane afterremoval of synaptotagmin-7. These data agree with findings in thesynaptotagmin-1 KO (8).

To investigate the kinetics of exocytosis, we examined themean capacitance traces from the KO and WT cells within thefirst 500 ms after the flash (Fig. 2C). The capacitance tracesfollowed each other during the first �20 fF increase, after whichtime the KO trace deviated. Consequently, if we normalize thetraces to the same amplitude at 0.5 s after the flash, the KO leadsthe WT trace (dotted line in Fig. 2C), indicating relative fasterrelease kinetics in the KO. We next estimated the release kineticsby fitting individual capacitance traces with a sum of exponen-tials. The amplitudes represent the vesicle pool sizes, whereas thetime constants identify their fusion kinetics. We found that insynaptotagmin-7-deficient cells, the fast burst was significantlysmaller (�58%, P � 0.0001, Fig. 2D), its time constant slightlyfaster (�25%, P � 0.0009, Fig. 2F), and the delay between theflash stimulation and the start of the capacitance increaseshorter (�44%, P � 0.0003, Fig. 2F) than in WT cells. The slowburst was significantly smaller in KO cells (�40%, P � 0.0065,Fig. 2D), but the fusion time course was not affected (Fig. 2F).

The apparently faster fusion kinetics of the RRP in theabsence of synaptotagmin-7 allows at least two alternativeinterpretations: Either the KO of synaptotagmin-7 causes somevesicles to fuse with bona fide faster kinetics, or the decrease inthe size of the RRP has uncovered a subpool with slightly fasterkinetics, which is also present in WT cells. For two reasons, wefavor the second interpretation. First, the capacitance trace of

+/+ +/+-/- -/-

Syt 7

Syt 1

Rab 3

VCP

AdrenalsBrain

L -

S -

A B

-1.0

0.0

)An( I

700600time (ms)

-0.1

0.0

Ia

C)

An(

300

200

100

0∆Cm

)Ff(

86420time (s)

40

20

0

aC[

+2] i )

Mµ(

50

25

0

Qp

mA

)Cp(

-1.5-1.0-0.50.0)

An( I

*

Syt 7 KO

WT

p=0.055

**

C

-0.2

Fig. 1. Depolarization experiments in the synaptotagmin-7 KO. (A) Expression of synaptotagmin-7 (Syt 7; L and S, long and short splice variants),synaptotagmin-1 (Syt 1), Rab 3, and VCP (loading control) in the brains and adrenal medulla of synaptotagmin-7 KO mice and their WT littermates was analyzedby immunoblotting. Note that only the short synaptotagmin-7 splice variants are expressed in adrenal medulla. (B) Chromaffin cells were stimulated bydepolarization trains (18 � 100 ms at 0 mV, the bottom graph shows the current) and exocytosis was monitored by simultaneous capacitance (second graph),and amperometry measurements (third graph, shown is the time integral of the amperometric current). The Ca2� concentration was measured simultaneouslyby dual-dye microfluorimetry (top graph). In the synaptotagmin-7 KO cells, exocytosis (gray traces, mean of n � 36 cells, N � 7 mice) was slightly depressed whencompared with their WT littermates (black traces, n � 41, N � 7). The KO cells displayed a significantly higher Ca2� buildup during the trains (*, P � 0.05), eventhough resting [Ca2�]i before the beginning of stimulation was unchanged (P � 0.92). (C) (Upper) Mean current during the first depolarization for KO (gray) andWT (black). (Lower) The amplitude of the Ca2� currents was similar in WT and KO cells.

Schonn et al. PNAS � March 11, 2008 � vol. 105 � no. 10 � 3999

NEU

ROSC

IEN

CE

Dow

nloa

ded

by g

uest

on

Aug

ust 1

, 202

1

the KO followed the WT trace for the first �20 fF (Fig. 2C),corresponding to the size of the RRP in KO cells (Fig. 2D).Second, we found that larger RRP amplitudes correlated weaklybut significantly with longer time constants in both KO and WTcells (Fig. 2G). Therefore, WT cells with smaller RRP sizes alsodisplay a slightly faster kinetics. We interpret this as evidence forheterogeneity within the RRP, where a small, faster pool appearsto be disguised by a larger and somewhat slower pool in cells witha large RRP size. In synaptotagmin-7 KO cells, only the fastestsubpool of the RRP is present.

Deletion of Synaptotagmin-7 May Impair the Refilling of the Releas-able Pools. Deletion of synaptotagmin-7 reduced the sustainedcomponent of release by 40% (P � 0.0012; Fig. 2 A and E). Thisnear-linear component reports on the vesicle priming process,which is much slower than the triggering step at these Ca2�

concentrations. Examination of the relative amplitudes of twoconsecutive flashes applied at a 90-sec interval also providesuseful information about the refilling of the releasable pools. InWT cells, the second flash elicited smaller responses [77% of thefirst-f lash burst amplitude, n � 65, N � 6 (n � number of cellsrecorded; N � number of mice used for each experimentalseries)]. This secretory depression was, however, more pro-nounced in absence of synaptotagmin-7 (48% of the first burstamplitude, P � 0.002; n � 51, N � 6, Fig. 2H). This observation

suggests that deletion of synaptotagmin-7 impaired releasablepools refilling between stimuli.

Ca2� Sensitivity of Exocytosis in the Low-Micromolar Range Is NotAltered by Deletion of Synaptotagmin-7. Using flash stimulation,we showed that deletion of synaptotagmin-7 modifies exocytosisat intermediate (20–30 �M) Ca2� levels. Because synaptotag-min-7 is proposed to be a high-affinity Ca2� sensor, we nextstudied the Ca2� dependence of exocytosis rates in the low-micromolar range (0.3–10 �M) using Ca2� ramps created bygradual Ca2� release from caged Ca2� (2).

We first examined [Ca2�]i-threshold values, i.e., the [Ca2�]i atwhich exocytosis shows maximal acceleration (Fig. 3A). Thisvalue is increased after mutation of synaptotagmin-1, whichlowers Ca2�-affinity (2). We found no significant change in thisparameter in synaptotagmin-7 KO mice, which is a first indica-tion of no major change in the Ca2� sensitivity of release (Fig.3B). We next calculated the rate of capacitance change as afunction of [Ca2�]i and normalized the rates to the remainingpool size (Fig. 3C). This normalization leads to the identificationof the pool size-independent fusion rate constant. No differencebetween KO and WT cells was found (Fig. 3C). Thus, deletionof synaptotagmin-7 does not profoundly modify the exocytoticrate constant in the low-micromolar Ca2� concentration range.

0

3020100

aC[

+2] i

)M

µ(

300300

200

100

0

∆Cm

)Ff(

200

100

WT

Syt 7 KO

Flash #1 Flash #2

100

50

02dn

1fo%(

hs alfts

)hsalf

**

H

A

543210time (s)

50

25

0Qp

mA

)Cp(

543210time (s)

50

25

0

F15

10

5

0

sm

Fast τ

*** 0.2

0.1

0.0

s

Slow τ

4

3

2

1

sm

***

Delay

E 40

30

20

10

0

s/Ff

Sustained

**

D 80

60

40

20

0Fast burst Slow burst

******

Ff

C60

0

∆)

F f(m

C

0.520.50time (s)

1

150 KOscaled

WT

Syt 7 KO

B6

5

4

3

3√

Qma

p(3

√)

Cp

1.00.80.60.40.22√Cm (

2√pF)

tsaF

τ) s

m(

Fast burst amplitude (fF)

1

10 100

10

G

Fig. 2. Ca2� uncaging reveals changes in several kinetic phases of exocytosis. (A) Flash photorelease of Ca2� (at arrows) resulted in a secretory burst, followedby a sustained component as measured by both capacitance (Middle) and amperometry (Bottom, time-integrated traces). Shown are averaged traces (WT, blacktraces, n � 72, N � 6; Syt 7 KO, gray traces, n � 62, N � 6). The second flash stimulation (Right) was delivered 90 s after the first one (WT, n � 65, N � 6; KO, n �51, N � 6). (B) The cubic root of the amperometric charge and the square root of the capacitance increase are both expected to be proportional with vesicle size.The overlapping regression lines from KO and WT measurements indicate no difference in the released amount of catecholamines per unit area of vesiclemembrane. (C) Expanded view of the capacitance increase during the first 20 ms after the uncaging flash. The dashed line is the KO trace normalized to WTamplitude at 500 ms after the flash, i.e., after exhaustion of the exocytotic burst (note the split axis). (D–F) Kinetic analysis of capacitance traces. Individualcapacitance traces were fitted with a triple-exponential function. Information on pool amplitudes and kinetics was extracted from all cells used to build theaverage traces in A, except for cells presenting a fast or slow burst amplitude �10 fF, where the pool size is too small for a reliable estimation of the kineticsin the presence of noise (rms noise was typically 2–3 fF). For these cells, only the pool amplitudes were included. This leaves for the fast-burst time constant: WT,n � 63; KO, n � 42; and for the slow burst time constant: WT, n � 66; KO, n � 47. In KO animals, the fast and slow bursts (D) and the sustained component (E)were reduced in size, whereas both the fast time constant and the delay between the UV flash and the beginning of capacitance increase were significantlyreduced (F). (G) A weak correlation between fast time constant of release and amplitude was found in both KO and WT cells. (H) Comparing the second flashburst size with the first one reports on the replenishment of the releasable vesicle pools. Replenishment was inhibited in KO cells. *, P � 0.05; **, P � 0.01; ***,P � 0.001.

4000 � www.pnas.org�cgi�doi�10.1073�pnas.0712373105 Schonn et al.

Dow

nloa

ded

by g

uest

on

Aug

ust 1

, 202

1

Ca2� Binding to the C2B Domain of Synaptotagmin-7 Is Essential forTriggering Exocytosis. Because deletion of synaptotagmin-7 altersseveral phases of exocytosis, we next asked whether these effectsdepend on Ca2� binding. Thus, we analyzed exocytosis fromsynaptotagmin-7 knockin (KI) mice, which express normal levels ofsynaptotagmin-7 with a mutated C2B-domain (27). Importantly,flash experiments in homozygous KI (�/�) mice and their WT(�/�) littermates closely reproduced the phenotype identified insynaptotagmin-7 KO mice (Fig. 4). Overall, exocytosis was reducedin the KI (Fig. 4A), because of changes in both the burst phase andthe sustained phase of release. Kinetic analysis showed that the fastburst was significantly reduced in amplitude (Fig. 4B), but displayedfaster kinetics (Fig. 4C), as in synaptotagmin-7 KO mice. Other thanin the KO mice, the slow burst was not significantly changed (Fig.4 B and C); however, a tendency to a reduction was still seen.Finally, the sustained component of release was decreased (Fig.4B). Overall, these findings demonstrate that the main changes inthe synaptotagmin-7 KO are reproduced when synaptotagmin-7is unable to bind Ca2� by its C2B domain, indicating thatsynaptotagmin-7 acts as a Ca2� sensor during exocytosis of dense-core vesicles (see also Table 1).

Synaptotagmins-1 and -7 Functionally Cooperate in Exocytosis. Thesearch for a second Ca2� sensor in chromaffin cell exocytosis wasmotivated by the modest effect of deleting synaptotagmin-1,which only leads to the elimination of the fast burst of release (3,8). To test whether synaptotagmin-7 is responsible for theresidual slower exocytosis in synaptotagmin-1 deficient chromaf-fin cells, we generated a double-KO mouse line. Because syn-aptotagmin-1 null mice are not viable, we used chromaffin cellsisolated from embryonic day (E)18 embryos. Embryonic chro-maffin cells might, in certain cases, deviate from adult cells byexpressing different protein isoforms (30). We therefore re-peated the experiments with synaptotagmin-7 KO using embry-onic mice and found qualitatively the same phenotype as in adultcells [supporting information (SI) Fig. 6]. Next, we preparedchromaffin cells from E18 embryos that resulted from crossingsof mice that were synaptotagmin-1 heterozygotes and synapto-tagmin-7 hetero- or homozygotes. Flash stimulation of controlchromaffin cells (heterozygous for synaptotagmin-7) resulted innormal fast exocytosis (Fig. 5, black trace). Note that exocytosisin these cells (Fig. 5) were similar to synaptotagmin-7 WTembryonic cells (compare SI Fig. 6), indicating that the het-erozygous condition was without effect. In agreement withprevious data, deletion of synaptotagmin-1 (in the synaptotag-min-7 heterozygous background) resulted in the elimination offast-burst exocytosis, whereas the slow exocytotic burst and thesustained component persisted (3, 8) (Fig. 5, red trace). Strik-

ingly, deletion of synaptotagmin-7 in the synaptotagmin-1 KObackground almost completely eliminated slow-burst exocytosis(Fig. 5, gray trace). The remaining exocytosis amounted to only20–30% of WT secretion and was almost linear in shape, withoutany signs of an initial faster (burst-like) phase.

Because of the low levels of exocytosis, kinetic analysis of

WTSyt 7 KO

2

1

0

aC[

+2)

Mµ( dlohserht ]

BA C

0.001

0.01

0.1

1

10

s/1 etar .mr o

N

10-6

10-5

[Ca2+

]i (M)

∆ )

Ff( m

C aC[

+2] i

)M

µ(

300

200

100

0

1086420time (s)

-80-40

04080

∆ s.Ff( "

mC

2-)

10

5

0

Remaining pool

Fig. 3. Deletion of synaptotagmin-7 does not alter the fusion rate constant in the low-micromolar range of Ca2�. (A) (Upper) Measurement from a chromaffin cellstimulated by a Ca2� ramp (blue symbols, right axis). The capacitance signal (red trace, left axis) showed a sigmoid-like increase. The black trace shows the size of theremaining releasable vesicle pool, which was used to calculate normalized release rates in C. (Lower) The second derivative of the capacitance trace, after smoothing,was used to calculate Ca2� threshold values in B. (B) Threshold [Ca2�] values ([Ca2�] at the point of maximum acceleration of the capacitance trace; WT, black, n � 28,N � 5; Syt 7 KO, gray, n � 31, N � 5). No difference was found between the two populations of cells (P � 0.8). (C) Normalized release rates were deduced from absoluterates by normalization to the remaining pool size at each data point. No difference was observed between WT and KO cells.

A

300

200

100

0

∆Cm

)Ff(

543210time (s)

3020100

aC[

+2] i

)M

µ(

50

25

0Qp

mA

)Cp(

WT

Syt 7 KI

B 40

30

20

10

0

s/Ff

Sustained

**Ff

C

120

100

80

60

40

20

0

***

Fast burst Slow burst

0.4

0.3

0.2

0.1

0.0

s

4

2

0

sm

Delay

*

30

20

10

0

sm ***

Fast τ Slow τ

Fig. 4. Ca2� binding to the C2B domain is required for synaptotagmin-7function. Chromaffin cells from KI mice expressing a C2B domain deficient forCa2� binding (synaptotagmin-7 KI) present the same exocytosis impairment assynaptotagmin-7 KO cells. (A) Averaged traces recorded from KI mice and WTlittermates (n � 23, N � 3; n � 23, N � 3, respectively). (B and C) Kinetic analysiswas conducted as for Fig. 2; similar features to that of synaptotagmin-7 KOwere found in KI cells (smaller and faster RRP, reduced sustained component),except that in the KI, the size of the slow burst was not significantly reduced.See also Table 1.

Schonn et al. PNAS � March 11, 2008 � vol. 105 � no. 10 � 4001

NEU

ROSC

IEN

CE

Dow

nloa

ded

by g

uest

on

Aug

ust 1

, 202

1

individual recordings was not possible in the double-KO cells.Instead, we fitted the averaged double-KO capacitance trace.Exocytosis in the double KO could be described by a sustainedcomponent (50% of the control), and a very small slow burstcomponent (13% of control cells; Table 1) but without the useof a fast burst component (as in synaptotagmin-1 KO cells). Thisconfirms that synaptotagmin-7 is responsible for the slowerphase of the exocytotic burst in synaptotagmin-1 KO chromaffincells. Thus, synaptotagmin-1 and synaptotagmin-7 display con-text-dependent redundancy as Ca2� sensors for exocytosis inchromaffin cells.

DiscussionWe have investigated the role of synaptotagmin-7 in Ca2�-triggered exocytosis using mutant mice that either entirely lacksynaptotagmin-7 expression or contain normal levels of a mutantsynaptotagmin-7 that is unable to bind Ca2� via its C2B domain(27). Our results establish that synaptotagmin-7 is important forexocytosis in chromaffin cells and that synaptotagmin-1 and -7together account for most of the Ca2�-triggered exocytosis in

these cells. Our conclusions are based on the following principalobservations:

1. Elimination of synaptotagmin-7 alone impairs Ca2�-triggeredexocytosis, with an �50% decrease observed by capacitancemeasurements or amperometry in good agreement withrecent data from insulin-secreting cells (31, 32).

2. Elimination of synaptotagmin-7 causes an apparent acceler-ation of the fast burst of exocytosis, probably by uncoveringa faster, synaptotagmin-1-responsive subpool of the RRP.

3. The synaptotagmin-7 KO phenotype is essentially phenocop-ied when synaptotagmin-7 is unable to bind Ca2� by its C2Bdomain.

4. The double KO of both synaptotagmin-1 and -7 abolishesboth the fast- and slow-burst components, leaving only a veryslow, linear secretory component

The apparent acceleration of the fast exocytotic burst in thesynaptotagmin-7 KO appears to happen because of heteroge-neity within the RRP (see above), such that the elimination ofsynaptotagmin-7 leaves a faster releasing subpool behind. Thissubpool might agree or overlap with the high Ca2�-sensitivevesicle pool described by the Gillis laboratory (33). Conse-quently, during mild stimulation regimes (Figs. 1 and 3), theeffect of the synaptotagmin-7 deletion is modest, which agreeswell with the viability of the synaptotagmin-7 KO mice.

During stronger stimulus regimes, recruitment processes domi-nate exocytosis. The synaptotagmin-7 KO led to a �2-fold de-crease in the sustained-release component, regardless of whethersynaptotagmin-1 was present or not. The synaptotagmin-1 KO, incontrast, does not change the sustained component or the size ofthe slow exocytotic burst, representing fusion of slowly releasablevesicles. A model that can explain these findings is that synapto-tagmin-7 engages the secretory machinery first, during vesiclepriming, and that it is then partly or completely replaced bysynaptotagmin-1 as the vesicle further matures and moves into thereadily releasable pool characterized by a higher release rate. In thiscompetition model, it would be expected that overexpression ofsynaptotagmin-1 leads to a speed-up in the apparent fusion kineticsof the burst, because of a larger fast component, which is exactlywhat we have found (8). The model is also consistent with theobservations that the synaptotagmin-1 KO slows down the exocy-totic burst and that the synaptotagmin-7 KO in the synaptotagmin-1null background eliminates the residual slow burst.

Synaptotagmin-7 has very different effects on the slow exo-cytotic burst, depending on whether synaptotagmin-1 is presentor not: In the absence of synaptotagmin-1, the burst seems to

Table 1. Comparison of KOt and KI mouse strains

Strain/genotype†

Secretory component Time constants, ms

Delay, ms No. of cells‡

Fast burst,fF

Slowburst,

fF

Sustainedcomponent,

fF/sFastburst

Slowburst

Syt 7 KO CTL 62 � 9 60 � 9 30.5 � 3.3 15.3 � 1.0 157 � 18 3.6 � 0.5 71 (72)KO 26 � 3*** 36 � 3*** 18.1 � 2.1** 11.4 � 1.5*** 195 � 23 2.0 � 0.3*** 57 (62)

Syt 7 KI CTL 91 � 15 69 � 9 25.5 � 3.2 20 � 3 291 � 57 3.3 � 0.6 20KI 32 � 4*** 56 � 8 17.8 � 4.2** 13 � 1** 183 � 18 1.5 � 0.5* 19

Syt 1/Syt 7 CTL 95 � 17 60 � 12 35.2 � 3.9 24 � 3 165 � 17 3.2 � 0.9 16 (33)DKO§ (�) �8 �17 (�) �200 (�) 58

Syt 1 KO �1 51 � 9 32.6 � 4.0 (�) 250 � 38 31.3 � 7.3 19 (26)

*, P � 0.05; **, P � 0.01; ***, P � 0.001.†Strain/genotype: For Syt1/Syt7 strain, CTL: �/� or �/� for synaptotagmin-1 and �/� for synaptotagmin-7, DKO: �/� for synaptotagmin-1 and �/� forsynaptotagmin-7. Syt 1 KO cells were �/� for synaptotagmin-1 and �/� for synaptotagmin-7.

‡The first number (without parentheses) gives the number of cells with capacitance responses that could be fitted with a sum of exponential functions. Inparentheses is given the total number of cells. All cells were used to calculate mean traces for figures and for estimating the sustained component.

§Fit to mean capacitance trace.

300

200

100

0

543210

time (s)

40

20

0

40

20

0

∆Cm

)Ff(

aC[

+2] i

)M

µ( Q

pm

A)

Cp(

Syt1/Syt7 DKO

control

Syt 1 KO

Fig. 5. Synaptotagmin-1 and -7 both determine the secretory-burst compo-nents. Mean responses to flash stimulation in Syt-1/Syt-7 double-KO (dKO)cells (Syt 1 �/�, Syt 7 �/�, gray, n � 58, N � 6), control littermates (Syt 1 �/�or �/�, Syt 7 �/�, black, n � 33, N � 4), and Syt 1 KO (Syt 1 �/�, Syt 7 �/�, redtrace, n � 26, N � 3) are shown; the residual exocytosis in synaptotagmin 1/7dKO cells was almost devoid of a secretory burst and was linear in appearance.

4002 � www.pnas.org�cgi�doi�10.1073�pnas.0712373105 Schonn et al.

Dow

nloa

ded

by g

uest

on

Aug

ust 1

, 202

1

depend almost entirely on synaptotagmin-7, but in the presenceof synaptotagmin-1, part of the burst is clearly independent ofsynaptotagmin-7. This indicates that the slow burst in thepresence and absence of synaptotagmin-1 has different origins,even though the amplitude and apparent fusion kinetics appearsto stay constant. Thus, the slow burst is not a state attributableto a specific Ca2� sensor but is, rather, a consequence of somevesicles being in one of several possible conditions that aresuboptimal for fast release.

The elimination of burst-like exocytosis and reduction ofoverall exocytosis by �70% in the synaptotagmin-1 and -7double-KO mice shows that exocytosis in chromaffin cells issimilar to neuronal exocytosis in that it requires the presence ofCa2�-binding synaptotagmin isoforms. These data demonstratethat synaptotagmins are strictly required for the release processitself and not just for synchronizing release that would happenanyway, as suggested (6). We demonstrate that at least inchromaffin cells, synaptotagmin-7 functions by binding to Ca2�

and, thus in this sense, is a Ca2� sensor. The finding that Ca2�

binding to the C2B domain of synaptotagmin-7 is essential is inagreement with similar findings for synaptotagmin-1 (7, 34–36).This similarity makes it likely that synaptotagmin-7 acts to triggerexocytosis in a manner similar to synaptotagmin-1, by binding tophospholipids and SNARE proteins (37, 38). Our data thusindicate that in chromaffin cells, synaptotagmin-1 and -7 arecompeting dual Ca2�-sensors sharing a single common mecha-nism but with distinct kinetic properties.

Materials and MethodsGeneration of Mouse Lines. The generation and analysis of synaptotagmin-7mutant mice has been described (27).

Preparation of Embryonic and Adult Mouse Chromaffin Cells. Embryonic (E18)adrenal glands were dissected and chromaffin cells prepared by enzymatic(papain) digestion as explained (39). For adult adrenals, the following mod-ifications were made: The glands were gently opened with tweezers beforedigestion, and, after trituration with a 1-ml pipette tip, the cell suspension wasbriefly centrifuged (4 min, 180 � g) and the pellet gently dispersed in culturemedium. After plating, cells were incubated at 8% CO2, 37°C, 95% relativehumidity and used within 1–4 days. Details of the cell isolation technique arefound in SI Materials and Methods.

Capacitance, Amperometry, and Intracellular Ca2� Measurements. Membranecapacitance whole-cell recordings were conducted essentially as described(17) by using an EPC-9 amplifier (HEKA Elektronik). Amperometric recordingswere carried out by using 10-�m carbon fibers (P-100S; Amoco) insulated byelectrodepositing insulating paint. Fibers cut at their tips were held in closecontact with the cell and held at 800 mV versus the reference electrode. Fordepolarization experiments, the cells were filled with a solution containing112 mM glutamic acid, 36 mM Hepes, 17 mM NaCl, 1 mM MgCl2, 2 mM Mg-ATP,0.3 mM GTP, 0.4 mM Fura-4F, 0.4 mM Furaptra 0.4 (pH 7.2) with CsOH;osmolarity �300 mOsm. Ca2�-uncaging experiments, intracellular [Ca2�] mea-surements by dual-dye ratiometric fluorimetry, and data analysis were per-formed as described (17, 39). The basal free-Ca2� concentration was �300–600 nM and adjusted if needed to the desired preflash level by photo-releasingsmall amounts of Ca2� by brief monochromator illumination. All experimentswere conducted at room temperature.

Statistical Analysis. Significance levels of observed differences were assessedby using nonparametric Mann–Whitney test (*, P � 0.05, **, P � 0.01, ***, P �0.001). Values are presented as mean values � SEM.

ACKNOWLEDGMENTS. We thank Ina Herfort and Dirk Reuter for experttechnical assistance and Erwin Neher for ongoing support and comments ona previous version of the manuscript. This work was supported by DeutscheForschungsgemeinschaft Grant SO 708/1-1/2 (to J.B.S.) and by a fellowshipfrom the Federation of European Biochemical Societies (to J.-S.S.).

1. Maximov A, Sudhof TC (2005) Autonomous function of synaptotagmin 1 in triggeringasynchronous release independent of asynchronous release. Neuron 48:547–554.

2. Sorensen JB, Fernandez-Chacon R, Sudhof TC, Neher E (2003) Examining synaptotag-min 1 function in dense core vesicle exocytosis under direct control of Ca2�. J GenPhysiol 122:265–276.

3. Voets T, et al. (2001) Intracellular calcium dependence of large dense-core vesicleexocytosis in the absence of synaptotagmin I. Proc Natl Acad Sci USA 98:11680–11685.

4. Fernandez-Chacon R, et al. (2001) Synaptotagmin I functions as a calcium regulator ofrelease probability. Nature 410:41–49.

5. Geppert M, et al. (1994) Synaptotagmin I: A major Ca2� sensor for transmitter releaseat a central synapse. Cell 79:717–727.

6. Nishiki T, Augustine GJ (2004) Synaptotagmin I synchronizes transmitter release inmouse hippocampal neurons. J Neurosci 24:6127–6132.

7. Yoshihara M, Littleton JT (2002) Synaptotagmin I functions as a calcium sensor tosynchronize neurotransmitter release. Neuron 36:897–908.

8. Nagy G, et al. (2006) Different effects on fast exocytosis induced by synaptotagmin 1and 2 isoforms and abundance but not by phosphorylation. J Neurosci 26:632–643.

9. Xu J, Mashimo T, Sudhof TC (2007) Synaptotagmin-1, -2, and -9: Ca2�-sensors for fast releasethat specify distinct presynaptic properties in subsets of neurons. Neuron 54:567–581.

10. Zhu D, et al. (2007) Synaptotagmin I, IX function redundantly in controlling fusion poreof large dense core vesicles. Biochem Biophys Res Commun 361:922–927.

11. Lynch KL, Martin TF (2007) Synaptotagmins I, IX function redundantly in regulatedexocytosis but not endocytosis in PC12 cells. J Cell Sci 120:617–627.

12. Marqueze B, et al. (1995) Cellular localization of synaptotagmin I, II, and III mRNAs inthe central nervous system and pituitary and adrenal glands of the rat. J Neurosci15:4906–4917.

13. Geppert M, Archer BT, III, Sudhof TC (1991) Synaptotagmin II. A novel differentiallydistributed form of synaptotagmin. J Biol Chem 266:13548–13552.

14. HuiE,etal. (2005)Threedistinctkineticgroupingsof thesynaptotagminfamily:Candidatesensors for rapid and delayed exocytosis. Proc Natl Acad Sci USA 102:5210–5214.

15. Han W, et al. (2004) N-glycosylation is essential for vesicular targeting of synaptotag-min 1. Neuron 41:85–99.

16. Sugita S, Shin O-H, Han W, Lao Y, Sudhof TC (2002) Synaptotagmins form hierarchy ofexocytotic Ca2�-sensors with distinct Ca2�-affinities. EMBO J 21:270–280.

17. Voets T (2000) Dissection of three Ca2�-dependent steps leading to secretion inchromaffin cells from mouse adrenal slices. Neuron 28:537–545.

18. Bhalla A, Tucker WC, Chapman ER (2005) Synaptotagmin isoforms couple distinctranges of Ca2�, Ba2�, and Sr2� concentration to SNARE-mediated membrane fusion.Mol Biol Cell 16:4755–4764.

19. Sugita S, et al. (2001) Synaptotagmin VII as a plasma membrane Ca2�-sensor inexocytosis. Neuron 30:459–473.

20. Martinez I, et al. (2000) Synaptotagmin VII regulates Ca(2�)-dependent exocytosis oflysosomes in fibroblasts. J Cell Biol 148:1141–1149.

21. Roy D, et al. (2004) A process for controlling intracellular bacterial infections inducedby membrane injury. Science 304:1515–1518.

22. Chakrabarti S, et al. (2003) Impaired membrane resealing and autoimmune myositis insynaptotagmin VII-deficient mice. J Cell Biol 162:543–549.

23. Jaiswal JK, Chakrabarti S, Andrews NW, Simon SM (2004) Synaptotagmin VII restrictsfusion pore expansion during lysosomal exocytosis. PLoS Biol 2:E233.

24. Wang P, Chicka MC, Bhalla A, Richards DA, Chapman ER (2005) Synaptotagmin VII istargeted to secretory organelles in PC12 cells, where it functions as a high-affinitycalcium sensor. Mol Cell Biol 25:8693–8702.

25. Fukuda M, Kanno E, Satoh M, Saegusa C, Yamamoto A (2004) Synaptotagmin VII istargeted to dense-core vesicles and regulates their Ca2�-dependent exocytosis in PC12cells. J Biol Chem 279:52677–52684.

26. Tsuboi T, Fukuda M (2007) Synaptotagmin VII modulates the kinetics of dense-corevesicle exocytosis in PC12 cells. Genes Cells 12:511–519.

27. Maximov, et al. (2008) Genetic analysis of synaptotagmin-7 function in synaptic vesicleexocytosis. Proc Natl Acad Sci USA 105:3986–3991.

28. Virmani T, Han W, Liu X, Sudhof TC, Kavalali ET (2003) Synaptotagmin 7 splice variantsdifferentially regulate synaptic vesicle recycling. EMBO J 22:5347–5357.

29. Pang ZP, Sun J, Rizo J, Maximov A, Sudhof TC (2006) Genetic analysis of synapto-tagmin 2 in spontaneous and Ca2�-triggered neurotransmitter release. EMBO J25:2039 –2050.

30. Speidel D, et al. (2005) CAPS1 regulates catecholamine loading of large dense-corevesicles. Neuron 46:75–88.

31. Li Y, et al. (2007) Regulation of insulin secretion and GLUT4 trafficking by the calciumsensor synaptotagmin VII. Biochem Biophys Res Commun 362:658–664.

32. Gauthier BR, et al. (2008) Synaptotagmin VII splice variants �, �, and � are expressed inpancreatic �-cells and regulate insulin exocytosis. FASEB J 22:194–206.

33. Yang Y, Udayasankar S, Dunning J, Chen P, Gillis KD (2002) A highly Ca2�-sensitive poolof vesicles is regulated by protein kinase C in adrenal chromaffin cells. Proc Natl AcadSci USA 99:17060–17065.

34. ShinO-H,etal.(2003)Sr2�bindingtotheCa2�bindingsiteofthesynaptotagmin1C2Bdomaintriggers fast exocytosis without stimulating SNARE interactions. Neuron 37:99–108.

35. Nishiki T, Augustine GJ (2004) Synaptotagmin I synchronizes transmitter release inmouse hippocampal neurons. J Neurosci 24:8542–8550.

36. Mackler JM, Drummond JA, Loewen CA, Robinson IM, Reist NE (2002) The C2B Ca2�-binding motif of synaptotagmin is required for synaptic transmission in vivo. Nature418:340–344.

37. Tang J, et al. (2006) A complexin/synaptotagmin 1 switch controls fast synaptic vesicleexocytosis. Cell 126:1175–1187.

38. Li C, et al. (1995) Ca2�-dependent and Ca2�-independent activities of neural andnonneural synaptotagmins. Nature 375:594–599.

39. Sorensen JB, et al. (2003) Differential control of the releasable vesicle pools by SNAP-25splice variants and SNAP-23. Cell 114:75–86.

Schonn et al. PNAS � March 11, 2008 � vol. 105 � no. 10 � 4003

NEU

ROSC

IEN

CE

Dow

nloa

ded

by g

uest

on

Aug

ust 1

, 202

1