A icle
The Rocke elle Uni e si y P ess $30.00
J. Gen. Physiol. Vol. 135 No. 4 379–392
www.jgp.o g/cgi/doi/10.1085/jgp.200910302 379
INTRODUCTION
Oxygen- egula ed K+ channels, ini ially desc ibed in
he abbi ca o id body (CB) glomus cell (López-Ba neo
e al., 1988; Gan o nina and López-Ba neo, 1991), a e
belie ed o play a undamen al ole in chemosenso y
ansduc ion. I is b oadly accep ed ha educ ion o
glomus cell K+ conduc ance in hypoxemia is he majo
e en leading o depola iza ion and Ca2+ channel open-
ing, ise o cy osolic [Ca2+], and ansmi e elease. These
ansmi e s s imula e a e en ne e ibe s ac ing on
b ains em espi a o y neu ons o e oke hype en ila ion
(López-Ba neo e al., 1993; Buckle and Vaughan-Jones,
1994; U eña e al., 1994; Mon o o e al., 1996; o ecen
e iews see P abhaka , 2006; López-Ba neo e al., 2008).
Di e en unc ional sub ypes o O2- egula ed K+ chan-
nels ha e been epo ed in glomus cells om se e al
mammalian species (Pee s, 1990; S ea and Nu se, 1991;
Gan o nina and López-Ba neo, 1992; Wya and Pee s,
1995; Buckle , 1997; Pé ez-Ga cía e al., 2004) as well as
P. O ega-Sáenz and K.L. Le i sky con ibu ed equally o his pape .
Co espondence o José López-Ba neo: [email p o ec ed]
Abb e ia ions used in his pape : 4-AP, 4-aminopy idyne; CB, ca o id
body; RT, e e se ansc ip ion; TH, y osine hyd oxylase.
in o he neu osec e o y cell classes acu ely esponding
o hypoxia ( o e iew see López-Ba neo e al., 2001;
Nu se e al., 2006).
Al hough he unde s anding o he cellula bases o
CB chemo ansduc ion has ad anced conside ably, he
p ecise molecula na u e o he O2 senso (s) and he
e ec o K+ channel(s) is unknown (see Kemp, 2006).
P og ess in his ield is hampe ed by me hodological
limi a ions de i ed om he gaseous na u e o he
s imulus and he delicacy o he O2-sensing appa a us,
which can be al e ed du ing cell dissocia ion (O ega-
Sáenz e al., 2007). Addi ionally, he small size o he CB
has p ecluded la ge-scale biochemical analyses. These
limi a ions can be pa ially o e come by he use o
gene ically modi ied mice, in which he unc ional con-
sequences o a ge ed molecula abla ion can be un-
ambiguously demons a ed (e.g., O ega-Sáenz e al.,
2006; Mulkey e al., 2007). To his end, we de eloped
Ca o id body chemosenso y esponses in mice de icien
o TASK channels
Pa icia O ega-Sáenz, Kons an in L. Le i sky, Ma ía T. Ma cos-Alma az, Vic o ia Bonilla-Henao,
Albe o Pascual, and José López-Ba neo
Ins i u o de Biomedicina de Se illa (IBIS) and Cen o de In es igación Biomédica en Red sob e En e medades Neu odegene a i as
(CIBERNED), Hospi al Uni e si a io Vi gen del Rocío, Consejo Supe io de In es igaciones Cien i icas, Uni e sidad de Se illa,
41013 Se illa, Spain
Backg ound K+ channels o he TASK amily a e belie ed o pa icipa e in senso y ansduc ion by chemo ecep o
(glomus) cells o he ca o id body (CB). Howe e , s udies on he sys emic CB-media ed en ila o y esponse o hy-
poxia and hype capnia in TASK1- and/o TASK3-de icien mice ha e yielded con lic ing esul s. We ha e cha ac-
e ized he glomus cell pheno ype o TASK-null mice and s udied he esponses o indi idual cells o hypoxia and
o he chemical s imuli. CB mo phology and glomus cell size we e no mal in wild- ype as well as in TASK1/ o
double TASK1/3/ mice. Pa ch-clamped TASK1/3-null glomus cells had signi ican ly highe memb ane esis-
ance and less hype pola ized es ing po en ial han hei wild- ype coun e pa . These elec ical pa ame e s we e
p ac ically no mal in TASK1/ cells. Sensi i i y o backg ound cu en s o changes o ex acellula pH was d as i-
cally diminished in TASK1/3-null cells. In con as wi h hese obse a ions, esponsi eness o hypoxia o hype -
capnia o ei he TASK1/ o double TASK1/3/ cells, as es ima ed by he ampe ome ic measu emen o
ca echolamine elease, was appa en ly no mal. TASK1/3 knockou cells showed an enhanced sec e o y a e in ba-
sal (no moxic) condi ions compa ible wi h hei inc eased exci abili y. Responsi eness o hypoxia o TASK1/3-null
cells was main ained a e pha macological blockade o maxi-K+ channels. These da a in he TASK-null mouse
model indica e ha TASK3 channels con ibu e o he backg ound K+ cu en in glomus cells and o hei sensi i -
i y o ex e nal pH. They also sugges ha , al hough TASK1 channels migh be dispensable o O2/CO2 sensing in
mouse CB cells, TASK3 channels (o TASK1/3 he e ome s) could media e hypoxic depola iza ion o no mal glo-
mus cells. The abili y o TASK1/3/ glomus cells o main ain a powe ul esponse o hypoxia e en a e blockade
o maxi-K+ channels, sugges s he exis ence o mul iple senso and/o e ec o mechanisms, which could con e
upon he cells a high adap abili y o main ain hei chemosenso y unc ion.
© 2010 O ega-Sáenz e al. This a icle is dis ibu ed unde he e ms o an A ibu ion–
Noncomme cial–Sha e Alike–No Mi o Si es license o he i s six mon hs a e he publi-
ca ion da e (see h p://www. up ess.o g/ e ms). A e six mon hs i is a ailable unde a
C ea i e Commons License (A ibu ion–Noncomme cial–Sha e Alike 3.0 Unpo ed license,
as desc ibed a h p://c ea i ecommons.o g/licenses/by-nc-sa/3.0/).
The Jou nal o Gene al Physiology
on Sep embe 22, 2014jgp. up ess.o gDownloaded om
Published Ma ch 29, 2010
h p://jgp. up ess.o g/con en /suppl/2010/03/22/jgp.200910302.DC1.h ml
Supplemen al Ma e ial can be ound a :
380 Glomus cell O2/CO2 sensing in TASK1/3-null mice
ano he labo a o y has obse ed in TASK1, bu no
in TASK3, knockou mice dec eased en ila ion and
a e en sinus ne e discha ges in esponse o hypoxia
and hype capnia (T app e al., 2008). He ein, we
epo he basic elec ophysiological p ope ies and
in insic chemosenso y ac i i y o indi idual glomus
cells om TASK-de icien animals. We show ha glo-
mus cells om TASK1-null animals appea o be no -
mal, bu TASK1/3 knockou cells exhibi cha ac e is ic
elec ophysiological al e a ions and dec eased sensi-
i i y o ex e nal pH. None heless, he sec e o y e-
sponses o TASK1/3/ cells o hypoxia as well as o he
chemosenso y s imuli emain essen ially unal e ed.
The implica ions o hese indings o CB O2 sensing
a e discussed.
MATERIALS AND METHODS
Animals
Fo he expe imen s, we used young adul (2–6-mo-old) TASK-
null mice (ei he TASK1/, TASK3/, o double TASK1/3/)
and he co esponding wild- ype li e ma es p o ided by M. Alle
(Ins i u o de Neu ociencias, Alican e, Spain). The non unc ional
alleles we e gene a ed as desc ibed in de ail p e iously (Alle e al.,
2005; B ickley e al., 2007). Mice we e geno yped as desc ibed
p e iously (Alle e al., 2005; B ickley e al., 2007). Animal ca e
and expe imen a ion we e pe o med acco ding o he ins i u-
ional animal ca e commi ee guidelines.
RNA analysis
Fou animals o e e y geno ype we e killed by sodium pen oba bi al
o e dose (in ape i oneally [i.p.]), and he CBs we e dissec ed,
pooled, and s o ed in liquid ni ogen. mRNA was ex ac ed using
Dynabeads mRNA DIRECT mic o ki (In i ogen). Fi s -s and cDNA
was syn hesized om o al mRNA ex ac ion using he Supe sc ip
i s -s and syn hesis sys em o e e se ansc ip ion (RT)-PCR (In i -
ogen). PCR ampli ica ions o TASK1, TASK3, and GAPDH mRNAs
we e pe o med using he ollowing p ime s: TASK1 (Kcnk3; 515 bp):
5-CACCGTCATCACCACAATCG-3 and 5-TGCTCTGCATCAC-
GCTTCTC-3; TASK3 (Kcnk9; 413 bp), 5-ATGAGATGCGCGAG-
GAGGAGAAAC-3 and 5-ACGAGGCCCATGCAAGAAAAGAAG-3;
and GAPDH (255 bp): 5-CAAAATGGTGAAGGTCGGTGTG-3 and
5-TTTGATGTTAGTGGGGTCTCGC-3. Fo quan i a i e RT-PCR
analysis, ou g oups o h ee young adul double TASK1/3/ o
con ol mice we e killed by pen oba bi al o e dose (i.p.), and he CBs
we e p ocessed as desc ibed abo e. Real- ime PCR was pe o med in
an ABI P ism 7500 Sequence De ec ion Sys em (Applied Biosys ems)
using SYBR G een PCR Mas e mix (Applied Biosys ems) and he
he mocycle condi ions ecommended by he manu ac u e . Each
sample was analyzed o cyclophilin o no malize o RNA inpu
amoun s and o pe o m ela i e quan i ica ions. To no malize mRNA
le els in knockou mice o hose in con ol samples, we calcula ed an
a e age cycle h eshold o he con ol samples and p ocessed all he
samples in he expe imen ela i e o his a e age cycle h eshold.
P ime s we e designed using he compu e p og am P ime Exp ess
(Applied Biosys ems). The ollowing p ime s we e used: maxi-K+
channel  subuni (Kcnma; 76 bp): 5-CATGGCTTTCAACGT-
GTTCTTC-3 and 5-GCCAGAACCACAGCTTATCATTG-3; TASK5
(Kcnk15; 53 bp): 5-GCCTACTACTACTGCTTCATCACTCTCA-3
and 5-ACGAAGTCGCCGAAGCCT-3; and cyclophilin A (Ppia;
75 bp): 5-GCACTGGTGGCAAGTCCAT-3 and 5-GCCAGGACCT-
GTATGCTTCAG-3. Mel ing cu e analysis showed a single sha p
peak wi h he expec ed Tm o all samples.
he mouse CB hin slice p epa a ion, whe e ep oducible
esponses o glomus cells o chemosenso y s imuli can
be ou inely ob ained (Pi ua e al., 2004; O ega-Sáenz
e al., 2007).
He e, we ha e e alua ed he chemosensi i i y o CB
glomus cells om mice de icien o TASK channels. These
belong o he andem po e domain (K2P) amily o
channels and con ibu e o he leak o backg ound K+
conduc ance in a b oad a ie y o cells. TASK1 (Kcnk3
o K2P3.1) and TASK3 (Kcnk9 o K2P9.1), he ele an
membe s o he TASK channel class (Dup a e al., 1997;
Kim e al., 2000; Rajan e al., 2000), can o m he e ome s
(Czi ják and Enyedi, 2002) and ha e been p oposed o
be in ol ed in pe iphe al and cen al chemo ecep ion
(Bayliss e al., 2001; Feldman e al., 2003; Mulkey e al.,
2004). Recombinan TASK1 channel ac i i y is educed
upon exposu e o low O2 ension (Kemp e al., 2004; Lee
e al., 2006; howe e , o con as ing esul s see Johnson
e al., 2004), and hese channels appea o media e he
hypoxic depola iza ion o ce ebella g anule cells (Plan
e al., 2002). In a CB glomus cells, an O2-sensi i e
TASK-like s anding K+ cu en wi h weak ou wa d ec i-
ica ion (in physiological asymme ical K+) and blocked by
ex acellula Ba2+ bu esis an o he classical K+ channel
blocke s TEA and 4-aminopy idyne (4-AP), has been e-
po ed (Buckle , 1997). The CB s anding K+ cu en sha es
o he pha macological p ope ies, such as ac i a ion by
he ola ile anes he ic halo hane and inhibi ion by anan-
damide, wi h cu en s media ed by TASK1 channels
(Buckle e al., 2000). Mo eo e , backg ound single K+
channel cu en ac i i y in glomus cells shows licke ing
kine ics and slope conduc ance compa ible wi h hose
epo ed o ecombinan , he e ologously exp essed
TASK1 channels (Williams and Buckle , 2004). The e o e,
a popula iew is ha TASK1 channels may be unda-
men al o CB O2 sensing (Dup a e al., 2007). Howe e ,
a de ailed ecen s udy by Kim e al. (2009) has shown
ha he e ome ic TASK1/TASK3 a e he majo O2-sensi-
i e backg ound K+ channels in a CB glomus cells.
In ecen yea s, TASK1- and/o TASK3-de icien
mice ha e been independen ly gene a ed in wo labo-
a o ies ha epo ed he animals o be heal hy and
wi h no mal li espan (Alle e al., 2005; B ickley e al.,
2007; Mulkey e al., 2007). The absence o TASK1 and/o
TASK3 esul s only in a mino pheno ype in cen al
neu ons, despi e he loss o acid sensi i i y in some
neu onal g oups and he comple e disappea ance o
halo hane e ec on memb ane cu en s o conduc ance.
In some mice s ains, in alida ion o TASK1 channels
dis up s ad enal gland zona ion and p oduces hype -
aldos e onism (Hei zmann e al., 2008). Howe e , he
impac o TASK de iciency on pe iphe al chemo ecep-
ion is a subjec o con o e sy. Al hough one g oup
has epo ed no mal en ila o y esponses o hypoxia
and hype capnia in he double TASK1 and TASK3
(TASK1/3/) knockou animals (Mulkey e al., 2007),
on Sep embe 22, 2014jgp. up ess.o gDownloaded om
Published Ma ch 29, 2010
O ega-Sáenz e al. 381
70 K2SO4, 30 KCl, 2 MgCl2, 1 EGTA, and 10 HEPES, pH 7.2.
The s anda d ba h solu ion con ained (in mM): 140 NaCl, 2.5 KCl,
10 HEPES, 10 glucose, 2.5 CaCl2, and 4 MgCl2, pH 7.4. Fo he pH
expe imen s, he ex e nal solu ions con ained (in mM): 3 KCl,
118 NaCl, 1 MgCl2, 1.5 CaCl2, 25 HEPES, and 10 glucose, wi h he
pH adjus ed o he desi ed le el using NaOH o HCl. We also
added 10 mM TEA and 5 mM 4-AP o he solu ion o block ol age-
dependen K+ channels. Es ima ed alues o es ing po en ial a e
gi en a e co ec ion o junc ion po en ials. Mac oscopic Ca2+,
Na+, and K+ cu en s we e eco ded in dialyzed glomus cells. The
solu ions used o he eco ding o whole cell Na+ and Ca2+ cu -
en s con ained (in mM): ex e nal: 140 NaCl, 9 BaCl2, 1 CaCl2, 10
HEPES, and 10 glucose; pH 7.4 and osmolali y 300 mOsm/kg;
and in e nal: 110 CsCl, 30 CsF, 10 EGTA, 10 HEPES, and 4 ATP-Mg;
pH 7.2 and osmolali y 285 mOsm/kg. The solu ions used o he
eco ding o whole cell K+ cu en s con ained (in mM): ex e nal:
140 NaCl, 2.5 KCl, 10 HEPES, 10 glucose, 2.5 CaCl2, and 4 MgCl2,
pH 7.4); and in e nal: 80 po assium glu ama e, 50 KCl, 1 MgCl2,
10 HEPES, 4 MgATP, and 5 EGTA, pH 7.2.
Ampe ome ic eco ding o single-cell ca echolamine
sec e ion in slices
CB slices we e used because he mos ep oducible single–glomus
cell esponses o hypoxia a e ob ained in his p epa a ion (Pa dal
e al., 2000). Mice CB dissec ion, slicing, and cul u e, as well as
he measu emen o ca echolamine sec e ion, we e pe o med ol-
lowing he same p ocedu es desc ibed p e iously (O ega-Sáenz
e al., 2003, 2006). CBs we e esec ed, cleaned o connec i e issue,
and included in aga ose. A e moun ing he piece on he s age o a
ib a ome, 150-µm hick slices we e cu , placed in a Pe i dish wi h
cul u e medium (Dulbecco’s modi ied Eagle’s medium wi h 10%
e al bo ine se um, 1% penicillin/s ep omycin, 1% l-glu amine,
and 84 U o insulin pe ml), and main ained a 37°C in a 5% CO2 in-
cuba o o 24–48 h. Slices we e ans e ed o a eco ding chambe
and con inuously pe used wi h a solu ion con aining (in mM):
117 NaCl, 4.5 KCl, 23 NaHCO3, 1 MgCl2, 2.5 CaCl2, 5 glucose, and
5 suc ose. The osmolali y o he solu ion was 280 mOsm/kg.
The “no moxic” solu ion was bubbled wi h a gas mix u e o
5% CO2, 20% O2, and 75% N2 (O2 ension, 145 mm Hg).
Immunocy ochemis y and mo phological s udies
Fo e e y geno ype, h ee animals we e killed by sodium pen o-
ba bi al o e dose (i.p.), and he ca o id bi u ca ions we e dissec ed,
washed wi h PBS, ixed 2 h a 4°C in 4% pa a o maldehyde, and
equilib a ed o 12 h in a 30% suc ose solu ion. Bi u ca ions we e
included in OCT (op imal cu ing empe a u e; Tissue Tek; Saku a)
and snap- ozen by quenching in d y ice. 10-µm hick slices we e
cu wi h a c yos a . Sec ions we e s ained wi h he an i– y osine hy-
d oxylase (TH) polyclonal an ibody (1:1,000; Pel-F eez Biologi-
cals). The En ision+ ki (Dako) was used o immunohis ochemis y
acco ding o he manu ac u e ’s ecommended p o ocol. The sig-
nal was de eloped wi h DAB (Dako). Images we e acqui ed unde
a mic oscope (BX-61; Olympus). Es ima ion o CB olume was
pe o med ac oss he en i e CB pa enchyma using he CAST G id
Sys em. A Ca alie i size o 912.7 µm2 was used.
Pa ch clamp eco dings
Mac oscopic cu en s we e eco ded om dispe sed mouse glomus
cells using ei he he pe o a ed pa ch o he whole cell con igu-
a ions o he pa ch clamp echnique as adap ed o ou labo a o y
(Muñoz-Cabello e al., 2005; Ga cía-Fe nández e al., 2007). P ep-
a a ion o dispe sed mouse CB cells was pe o med as desc ibed
p e iously (Pi ua e al., 2004; O ega-Sáenz e al., 2006). Pa ch
elec odes (1.5–2.5 MΩ) we e pulled om capilla y glass ubes
(1.5–1.6 mm OD; Kimax; Kimble P oduc s), i e polished on a
mic o o ge MF-830 (Na ishige), and coa ed wi h silicone elas ome
(Sylga d 184; Co ning) o dec ease capaci ance. Vol age clamp
eco dings we e ob ained wi h an EPC-8 pa ch clamp ampli ie
(HEKA) using s anda d ol age clamp p o ocols designed wi h
Pulse so wa e (HEKA). Unless o he wise no ed, holding po en-
ial was 80 mV. Da a we e il e ed a 10 kHz, digi ized a a sam-
pling in e al o 20 µs wi h an ITC-16 A/D con e e (HEKA),
and s o ed on a Macin osh compu e . O line analysis o da a was
pe o med using cus om so wa e and Pulse Fi (HEKA). All ex-
pe imen s we e conduc ed a oom empe a u e, 23–26°C. Expe i-
men s designed o es ima e he cell’s es ing po en ial and inpu
esis ance, as well as he pH dependence o he backg ound K+
cu en s, we e made using pe o a ed pa ches wi h ampho e icin B
in he pipe e solu ion. This solu ion also con ained (in mM):
Figu e 1. Molecula and his ological cha ac e iza ion o TASK1- and double TASK1/3–de icien mice. (A) RT-PCR analysis showing he
absence o Task1 o Task3 mRNA exp ession in he CB o TASK1- o TASK1/3-null mice. (B) CB ana omy in young TASK1- o TASK1/3-
de icien animals. Rep esen a i e sec ions o con ol (wild ype; le ), TASK1/ (middle), and TASK1/3/ ( igh ) ca o id bodies. CB
glomus cells a e s ained wi h an an ibody agains TH. IC, in e nal ca o id a e y. (C) G aph ep esen ing he CB olume occupied by
TH-posi i e cells. Da a a e n = 3 o each expe imen al condi ion.
on Sep embe 22, 2014jgp. up ess.o gDownloaded om
Published Ma ch 29, 2010
382 Glomus cell O2/CO2 sensing in TASK1/3-null mice
sum o he ime in eg al o successi e ampe ome ic e en s (Pa dal
and López-Ba neo, 2002). Sec e ion a e (in ei he em o- o pico-
coulombs/min) was calcula ed as he amoun o cha ge ans-
e ed o he eco ding elec ode du ing a gi en ime pe iod.
S a is ical analysis
Unless o he wise speci ied, da a a e exp essed as mean ± SE, wi h
he numbe (n) o expe imen s indica ed. S a is ical analysis was
pe o med by unpai ed S uden ’s es . A alue o P < 0.05 was
conside ed as s a is ically signi ican .
Online supplemen al ma e ial
Con ocal luo ime ic eco dings o in acellula Ca2+ concen a-
ion in dispe sed mouse glomus cells a e shown in Fig. S1. The
da a indica e ha in bo h wild- ype and TASK1/3/ glomus cells,
hypoxia induces an inc ease o in acellula Ca2+ concen a ion.
In Fig. S2, we show in dispe sed ch oma in cells he inhibi ion o
The “hypoxic” solu ion was bubbled wi h 5% CO2 and 95% N2 o
each an O2 ension in he chambe o 15 mm Hg. To pe o m
dose– esponse cu es, he solu ions we e also bubbled wi h 12
and 6% O2, keeping CO2 a 5%. When hese solu ions we e used,
he app oxima e alues o O2 ension in he chambe we e, e-
spec i ely, 90 and 50 mm Hg. Expe imen s o sec e ion induced by
hype capnia we e done wi h he same con ol solu ion bubbled
wi h 10% CO2, 20% O2, and 70% N2 o 20% CO2, 20% O2, and
60% N2 (10 o 20% hype capnia, espec i ely). In hese condi-
ions, ex acellula pH dec eased om 7.3 in 5% CO2 o 7.1 in
10% CO2 o 6.8 in 20% CO2. All he expe imen s we e made a a
empe a u e in he chambe o 36°C. Sec e o y e en s we e e-
co ded wi h a pola ized (+750 mV) 10-µm ca bon ibe elec ode
posi ioned nea a cell unde isual con ol and connec ed o he
cu en - o- ol age con e e o an EPC-8 pa ch clamp ampli ie .
Ampe ome ic cu en s we e il e ed a 100 Hz and s o ed on a
compu e . The cumula i e sec e ion signal was ob ained by he
TABLE I
Elec ophysiological pa ame e s o wild- ype and TASK-null mouse glomus cells
Pa ame e Wild ype Task1/Task1/3/
Cell capaci ance (pF) 2.9 ± 0.1 (25) 2.7 ± 0.1 (29) 2.9 ± 0.1 (28)
Inpu esis ance (GOhm) 7.8 ± 0.9 (25) 8.4 ± 0.8 (29) 10.8 ± 0.9* (31)
Memb ane po en ial (mV) 57.0 ± 1.2 (24) 54.5 ± 1.8 (29) 51.5 ± 0.8* (33)
Cu en densi y (20 mV, pH 8.2, pA/pF) 7.9 ± 2.7 (13) — 1.6 ± 0.5* (14)
Peak K+ cu en (+30 mV, pA/pF) 362 ± 17 (13) 390 ± 21 (11) 262 ± 17** (21)
Peak Na+ cu en (+10 mV, pA/pF) 26.2 ± 5.0 (11) 24.5 ± 5.2 (12) 27.5 ± 3.4 (22)
Peak Ca2+ cu en (+20 mV, pA/pF) 7.5 ± 0.9 (21) 9.1 ± 1.9 (12) 3.3 ± 0.4** (33)
Reco dings we e pe o med using pe o a ed pa ches wi h he excep ion o hose done o measu e mac oscopic ol age-dependen K+, Na+, and Ca2+
cu en densi y, in which we used whole cell (dialyzed) pa ch-clamped cells. Values a e gi en as mean ± SE, wi h he numbe o expe imen s in pa en heses.
As e isks indica e s a is ical signi icance (*, P < 0.05; **, P < 0.01) wi h espec o co esponding alues in wild- ype cells.
Figu e 2. Sensi i i y o ex e nal
pH o backg ound po assium cu -
en s om wild- ype and TASK-
null mouse glomus cells. (A; op)
Vol age amp p o ocol applied o
dispe sed glomus cells s udied wi h
he pe o a ed pa ch echnique.
10 mM TEA and 5 mM 4-AP we e
added o he ex e nal solu ion o
block ol age-dependen K+ chan-
nels. (Middle) Cu en s eco ded
in wild- ype glomus cells exposed o
ex e nal solu ions wi h pH 6.3 and
8.2. (Bo om) Cu en s eco ded in
double TASK1/3–null glomus cells
exposed o he same ex e nal pH
shi . (B) Quan i a i e analysis o he
po assium cu en densi y (20 mV)
a di e en pH o wild- ype (n = 13)
and TASK1/3/-de icien (n = 14)
cells. *, P < 0.05. (C) Quan i a i e
analysis o he e ec o pH on inpu
cell esis ance o wild- ype (n = 13)
and TASK 1/3/ (n = 14) cells.
**, P < 0.01.
on Sep embe 22, 2014jgp. up ess.o gDownloaded om
Published Ma ch 29, 2010
O ega-Sáenz e al. 383
expe imen s we e pe o med on glomus cells om
he TASK1/3 double knockou , al hough, when neces-
sa y, expe imen s we e also done on single TASK1-null
cells because hese las channels a e hose p oposed o
pa icipa e mo e speci ically in CB O2 sensing (see
Dup a e al., 2007; T app e al., 2008). In ou expe i-
men al condi ions, he es ima ed es ing memb ane po-
en ial o dispe sed wild- ype cells eco ded wi h he
pe o a ed pa ch echnique was 57 ± 1.2 mV (n = 24);
his alue was only sligh ly changed in TASK1-de icien
cells (54.5 ± 1.8 mV; n = 29) bu dec eased signi ican ly
( o 51.5 ± 0.8 mV; n = 33; P < 0.05) in TASK1/3-null
cells. Glomus cells om TASK1/3 knockou mice also
had a s a is ically signi ican inc ease o inpu esis ance
compa ed wi h con ols (Table I).
Recombinan TASK1 and TASK3 channels a e highly
sensi i e o ex acellula pH due o a p o on-sensing
his idine esidue loca ed a he ex e nal mou h o he
channels (Rajan e al., 2000). Based on pha macological
expe imen s, i has been p oposed ha TASK1 channels
con ibu e o pH sensing in glomus cells (Buckle e al.,
2000). In cells ba hed in ex e nal solu ion con aining
TEA and 4-AP o minimize he ion luxes h ough ol -
age-dependen K+ channels, cu en s e oked by depo-
la izing amps we e inc eased by pH alcaliniza ion
(swi ching om pH 6.3 o 8.2) and, as expec ed, his e -
ec was small o e en negligible in TASK1/3-null cells
(Fig. 2, A and B). Upon exposu e o pH 8.2, ou wa d K+
cu en densi y a 20 mV in wild- ype cells was on a e -
age i e old la ge han in cells om TASK1/3-de icien
animals (Fig. 2 B and Table I). In pa allel wi h he e ec
on K+ cu en densi y, abla ion o he TASK1/3 channel
Ca2+-dependen mac oscopic K+ cu en s by paxilline, a selec i e
blocke o maxi-K+ channels. Figs. S1 and S2 a e a ailable a
h p://www.jgp.o g/cgi/con en / ull/jgp.200910302/DC1.
RESULTS
Mo phology o TASK-null ca o id bodies
In wild- ype animals, we con i med by PCR he exp es-
sion o TASK1 and TASK3 mRNAs in he CB and hei
comple e disappea ance in TASK-de icien animals
(Fig. 1 A). Immunocy ochemical analyses wi h an i-TH
an ibodies demons a ed in all he animal ypes ha he
CBs we e no mal and appea ed o ganized in he cha -
ac e is ic clus e s (glome uli) o TH-posi i e glomus cells
(Pa dal e al., 2007) (Fig. 1 B). The olume occupied by
he CB pa enchyma was simila in TASK1 o TASK1 and
TASK3 double (TASK1/3) knockou mice compa ed
wi h wild- ype li e ma es (Fig. 1 C). The size o indi id-
ual CB glomus cells, as indica ed by he alue o o al
capaci ance measu ed in pe o a ed pa ch-clamped cells,
was also simila in he h ee animal s ains (see Table I).
Elec ophysiological pa ame e s and pH sensi i i y
o TASK-de icien glomus cells
TASK1 and TASK3 channels a e exp essed in nume ous
neu al and non-neu al issues, as well as in he CB. These
channels a e open o e a b oad ange o memb ane
ol ages and con ibu e o se he cell’s es ing po en ial
and memb ane esis ance. Hence, we sough o see
whe he he absence o TASK channels esul ed in mod-
i ica ions in he elec ophysiological pa ame e s o glo-
mus cells. Fo he sake o simplici y, some o hese
Figu e 3. Vol age-dependen K+
cu en s in mouse glomus cells e-
co ded wi h he whole cell con igu-
a ion o he pa ch clamp echnique.
(A) Rep esen a i e amily o K+ cu -
en s eco ded a a ious memb ane
po en ials. The holding po en ial
was 80 mV, and he depola iza ion
ol ages (mV) a e indica ed nea
each ace. (B) Po assium cu en –
ol age ela ionship ob ained om
wild- ype (n = 13), TASK1/ (n = 11),
and TASK1/3/ (n = 21) glomus
cells. (C) Le el o exp ession o he
mRNA o maxi-K+ channel  sub-
uni es ima ed by quan i a i e PCR
om CB issue o wild- ype and
TASK1/3-null animals. Da a a e
scaled wi h espec o alues in wild-
ype animals (n = 4 expe imen s).
on Sep embe 22, 2014jgp. up ess.o gDownloaded om
Published Ma ch 29, 2010

384 Glomus cell O2/CO2 sensing in TASK1/3-null mice
channels, he closes ela i e o TASK1 and TASK3 wi hin
he TASK amily (Dup a e al., 2007), we e no signi i-
can ly exp essed in he CB issue. The maxi-K+ channel 
subuni , unc ionally exp essed in mouse CB (Yamaguchi
e al., 2004; O ega-Sáenz e al., 2006), appea ed o be
down- egula ed in TASK1/3/ animals (Fig. 3 C).
Dispe sed mouse glomus cells exhibi ed ela i ely la ge
Na+ and/o Ca2+ inwa d cu en s (Fig. 4 A). The cu en –
ol age ela ion o he Na+ cu en was unal e ed in
TASK-de icien cells, and TASK1/ cells also had
Ca2+ cu en s o no mal ampli ude (Table I and Fig. 4, B
and C). Howe e , in TASK1/3/ cells, he peak Ca2+ cu -
en densi y dec eased o 50% o con ol alues, al-
hough he ol age dependence o he cu en emained
unchanged (Table I and Fig. 4 C). As hei abbi coun-
e pa s (U eña e al., 1989), mouse glomus cells ha e
wo well- ep esen ed dis inc popula ions o Ca2+ chan-
nels ha we e easily sepa a ed by hei deac i a ions
genes also abolished he dec ease o memb ane esis-
ance induced by alcaliniza ion (Fig. 2 C). Collec i ely,
hese da a sugges ha he lack o TASK channels in CB
glomus cells (pa icula ly TASK3) does indeed esul in
consis en al e a ions o hei elec ophysiological pa-
ame e s, as well as he esponsi eness o changes in
ex acellula pH.
To u he in es iga e he TASK-null CB pheno ype, we
measu ed he densi y o ol age-dependen K+, Na+, and
Ca2+ cu en s in wild- ype and TASK1 o TASK1/3 knock-
ou glomus cells. La ge ol age-dependen ou wa d K+
cu en s we e eco ded in he h ee glomus cell ypes
(Fig. 3 A). The ampli ude o he cu en was unchanged
in TASK1/ cells, bu a clea educ ion o K+ cu en
densi y (30%) was obse ed in TASK1/3/ p epa a-
ions (Fig. 3 B and Table I). We ha e es ed by quan i a-
i e RT-PCR whe he he mRNA exp ession o o he K+
channel genes is al e ed in TASK1/3/ CB cells. TASK5
Figu e 4. Vol age-dependen Na+ and
Ca2+ cu en s in mouse glomus cells
eco ded wi h he whole cell con igu-
a ion o he pa ch clamp echnique.
(A) Rep esen a i e mac oscopic so-
dium (INa) and calcium (ICa) cu en s
eco ded in a cell du ing a depola iza-
ion o +20 mV om a holding po en-
ial o 80 mV. The decay o he ail
cu en gene a ed on epola iza ion
o 70 mV e lec s he closing ime
cou se o he channels open du ing he
pulse. (B) Peak sodium cu en – ol -
age ela ionship o wild- ype (n = 11),
TASK1/ (n = 12), and TASK1/3/
(n = 22) glomus cells. (C) Peak cal-
cium cu en – ol age ela ionship o
wild- ype (n = 18), TASK1/ (n = 12),
and TASK1/3/ (n = 24) glomus cells.
(D) Single-exponen ial unc ions i ed
o he Ca2+ ail cu en (da k g ay, as
deac i a ing componen ; ligh g ay,
slowly deac i a ing componen ). The as
and slow ime cons an alues in his ex-
ample a e, espec i ely, 0.11 and 1.21 ms.
(E and F) Quan i a i e analysis o he
cu en densi y o as and slowly deac i-
a ing componen s o Ca2+ ail cu en s
om wild- ype (n = 25), TASK1/
(n = 12), and TASK1/3/ (n = 34) glo-
mus cells. *, P < 0.05; **, P < 0.01.
on Sep embe 22, 2014jgp. up ess.o gDownloaded om
Published Ma ch 29, 2010
O ega-Sáenz e al. 385
he cellula le el a e almos simila o he co ela ion
be ween a e ial PO2 and he a e en discha ges o he
CB sinus ne e eco ded in i o o in he explan ed CB
in i o (Fig. 6, A and B).
kine ics (a 70 mV:  as = 0.123 ± 0.009 ms and  slow =
1.60 ± 0.12 ms; n = 25; Fig. 4 D). These co espond o
high h eshold, o as deac i a ing, and low h eshold,
o slowly deac i a ing, Ca2+ channels, which in neu o-
sec e o y cells can con ibu e o ansmi e elease (see
Ca abelli e al., 2007; Le i sky and López-Ba neo, 2009).
The deac i a ion ime cons an s o hese channels ypes
we e also indis inguishable be ween wild- ype and TASK-
null cells (TASK1/:  as = 0.093 ± 0.006 ms and  slow =
1.47 ± 0.09 ms; n = 12; TASK1/3/:  as = 0.106 ± 0.005 ms
and  slow = 1.66 ± 0.09 ms; n = 34). The dec ease o
Ca2+ cu en densi y obse ed in TASK1/3/ glomus
cells was mainly due o dec ease o he high ol age–
ac i a ed cu en , he mos p edominan componen
in glomus cells (Fig. 4 E). Ne e heless, he densi y o he
low ol age–ac i a ed cu en was also signi ican ly e-
duced in he TASK1/3 knockou cells (Fig. 4 F).
Chemosenso y esponses o hypoxia o glomus cells
om TASK-null mice
Responsi eness o CB glomus cells o hypoxia in TASK-
de icien mice was s udied using hin CB slices, whe e
he in insic O2 sensi i i y o in ac glomus cells can be
e alua ed sepa a ely om he o he s eps along he chemo-
senso y pa hway in ol ed in he hypoxic en ila o y
esponses (Pi ua e al., 2004; O ega-Sáenz e al., 2006).
Rep esen a i e eco dings o ca echolamine elease
om indi idual glomus cells subjec ed o low O2 en-
sion (PO2, 15 mm Hg) a e illus a ed in Fig. 5 A, and a
quan i a i e summa y o he sec e ion a e in no moxic
and hypoxic condi ions is shown in Fig. 5 (B and C).
Sec e ion a e induced by hypoxia in wild- ype glomus
cells (5,133 ± 1,010 C/min; n = 7) was simila o he al-
ues obse ed in cells de icien o ei he TASK1 (5,175 ±
719 C/min; n = 12) o TASK1/3 (5,939 ± 1,595 C/min;
n = 5) channels (Fig. 5 B). Basal sec e ion in no moxic
condi ions was, howe e , o e wo old highe in TASK1/3-
null cells han in he wo o he glomus cell ypes (Fig. 5 C).
As he Ca2+ channel densi y was educed in TASK1/3/
cells, we also es ed ha , simila o o he oden glomus
cells (Pa dal e al., 2000; Pi ua e al., 2004), he sec e-
o y esponse o hypoxia was abolished by blockade o
Ca2+ channels wi h Cd2+ (Fig. 5 A, inse , bo om), hus
sugges ing ha i was igge ed by ansmemb ane Ca2+
in lux. P elimina y expe imen s pe o med on dispe sed
glomus cells also indica e ha hypoxia induces simila
ises o cy osolic [Ca2+] in wild- ype and TASK1/3-null
cells (Fig. S1).
Sensi i i y o hypoxia o TASK-null glomus cells was
u he s udied by dose– esponse expe imen s in which
mouse CB slices we e exposed o a ious PO2 le els.
Glomus cells showed a g aded esponse upon exposu e
o p og essi ely highe hypoxia wi h a cha ac e is ic
hype bolic co ela ion ha was p ac ically simila in all
he animal s ains s udied (wild ype, TASK1 null, and
TASK1/3 null). These dose– esponse ela ionships a
Figu e 5. Responsi eness o acu e hypoxia o CB glomus cells
om wild- ype, TASK1-, and double TASK1/3–null mice. (A) Am-
pe ome ic eco dings and co esponding cumula i e sec e ion
signals (in pC) o ca echolamine elease induced by hypoxia (O2
ension, 15 mm Hg) in CB glomus cells om he di e en mice
s ains s udied. The inse a he bo om o he panel shows ha
he sec e o y esponse induced by hypoxia in TASK1/3-null glo-
mus cells was inhibi ed by blockade o Ca2+ channels wi h 0.2 mM
Cd2+. (B) Quan i ica ion o he inc ease in sec e ion a e (pC/min)
induced by hypoxia. S a is ical signi icance (P < 0.05) wi h espec
o basal alues in no moxia (O2 ension, 145 mm Hg; n = 7,
12, and 5 o wild- ype, TASK1/, and TASK1/3/ cells, espec-
i ely). (C) Basal sec e ion a e o glomus cells om he a ious mice
s udied in no moxic condi ions. Basal sec e ion a e (in C/min)
in TASK1/3-null cells was signi ican ly di e en (P < 0.05) wi h
espec o alues in wild- ype cells.
on Sep embe 22, 2014jgp. up ess.o gDownloaded om
Published Ma ch 29, 2010
386 Glomus cell O2/CO2 sensing in TASK1/3-null mice
s udied whe he in he mu an animals he O2-sensi i e
maxi-K+ cu en was necessa y o compensa e o he lack
o TASK channels. In hese expe imen s, maxi-K+ chan-
nels we e blocked wi h paxilline, a b oadly used selec i e
an agonis (G ibko e al., 1996; Sheehan e al., 2009)
ha in ou expe imen al condi ions ully inhibi ed
maxi-K+ channels (Fig. S2). In he p esence o high con-
cen a ions o paxilline, he sec e o y esponse o hypoxia
o TASK1/3-de icien cells emained unal e ed (Fig. 7 B).
Chemosenso y esponses o hype capnia and
hypoglycemia o glomus cells om TASK-null mice
The CB is a polimodal chemosenso y ecep o ha be-
sides hypoxia o pH is also ac i a ed by hype capnia and
hypoglycemia (Pa dal and López-Ba neo, 2002; Ga cía-
Fe nández e al., 2007; Zhang e al., 2007; Fi zge ald e al.,
2009). The esponsi eness o wild- ype and TASK1/3
double knockou glomus cells o hype capnia was ana-
lyzed keeping cons an O2 ension and changing CO2
concen a ion om 5 o 10 and 20%. Exposu e o high
CO2 ension e oked a su ge o ansmi e elease om
glomus cells ha , al hough weake han he esponse o
hypoxia, was also concen a ion dependen (Fig. 8 A).
As no iced in a p e ious se o expe imen s, basal sec e-
ion a e was highe in TASK1/3-de icien han in con ol
glomus cells; howe e , sensi i i y o hype capnia (ei he
10 o 20%) was unal e ed by abla ion o he TASK
genes (Fig. 8 B). TASK1/3-de icien glomus cells exhib-
i ed a sec e o y esponse o emo al o glucose quali a-
i ely simila o ha desc ibed p e iously in he a CB
(Pa dal and López-Ba neo, 2002), which was also po en-
ia ed upon concomi an exposu e o hypoxia (Fig. 9).
DISCUSSION
Mo phological and elec ophysiological p ope ies
o TASK1- and TASK1/3-null glomus cells
In his pape , we show ha TASK1- o TASK1/3-null
mice ha e mo phologically no mal CB, wi h TH-posi i e
glomus cells a anged in clus e s (glome uli) ypical o he
CB pa enchyma (Pa dal e al., 2007). TASK1-de icien
glomus cells had no mal passi e elec ophysiological pa-
ame e s. In con as , glomus cells lacking bo h TASK1
and TASK3 channels showed a clea elec ophysiological
pheno ype cha ac e ized by an inc ease o memb ane
esis ance and cell depola iza ion. These obse a ions
sugges ha TASK3 channels (o he e ome s o TASK1
and TASK3 channels) con ibu e o se he es ing po en-
ial o no mal mouse glomus cells (see Kim e al., 2009).
Backg ound K+ cu en s eco ded om TASK1/3-null
glomus cells we e less sensi i e o changes in ex acellu-
la pH han hose in con ol cells. This inding i s well
wi h he p oposal ha TASK-like channels pa icipa e in
ex acellula acid sensing in he CB (Buckle e al.,
2000; o discussion on CB acid sensing see López-López
To gain u he insigh in o he ole o TASK channels
in CB O2 sensing, we pe o med ampe ome ic expe i-
men s in no mal (wild- ype) mouse glomus cells exposed
o acidic pH o minimize he ac i i y o pH-sensi i e
backg ound channels. Low pH, which p esumably inhibi s
TASK channels (see Fig. 2), induced sec e o y ac i i y
in glomus cells bu did no p e en a ull esponse
upon subsequen exposu e o hypoxia. Indeed, low pH
appea ed o po en ia e he e ec o hypoxia (Fig. 7 A).
Al hough TASK1/3/ glomus cells showed down- egu-
la ion o bo h mac oscopic K+ cu en ampli ude (Fig. 3 B)
and maxi-K+ channel mRNA exp ession (Fig. 3 C), we
Figu e 6. Responsi eness o CB glomus cells o di e en le -
els o hypoxia in wild- ype, TASK1-, and double TASK1/3–null
mice. (A) Ampe ome ic eco dings o ca echolamine sec e ion
induced by di e en le els o O2 ension in CB glomus cells om
he indica ed mouse s ains (con ol, 21% O2, 145 mm Hg; 12%
O2, 90 mm Hg; 6% O2, 50 mm Hg; hypoxia, 15 mm Hg). (B) Dose–
esponse cu es es ima ed om glomus cells o he a ious mouse
s ains exposed o solu ions equilib a ed wi h a ious le els o O2
(150, 90, 50, and 15 mm Hg). Da a a e om h ee o se en expe i-
men s. The a e age sec e ion a e alues in wild- ype, TASK1/,
and TASK1/3/ glomus cells a each O2 ension we e no signi i-
can ly di e en (P > 0.05).
on Sep embe 22, 2014jgp. up ess.o gDownloaded om
Published Ma ch 29, 2010
O ega-Sáenz e al. 387
ni ican popula ion o low h eshold T- ype Ca2+ channels
ha , as desc ibed in neona al o hypoxic adul ch oma in
cells (Ca abelli e al., 2007; Le i sky and López-Ba neo,
2009), could be ac i a ed by small memb ane depola -
iza ions o induce exocy osis. Indeed, we ha e obse ed
an inc eased basal sec e ion a e in TASK1/3-null cells
(see below), which is compa ible wi h he sus ained de-
pola iza ion and inc eased exci abili y o his cell ype.
Vol age-dependen (Na+, K+, and Ca2+) cu en densi-
ies we e no mal in TASK1 knockou glomus cells; how-
e e , K+ and Ca2+ cu en s we e ma kedly educed in
TASK1/3-de icien cells. These obse a ions u he sup-
po he iew ha al hough TASK1 channels migh be
dispensable, TASK3 channels (ei he as homome s o
TASK1/TASK3 he e ome s) a e absolu ely equi ed o
he main enance o he physiological pheno ype o glo-
mus cells. The dec ease o Ca2+ and K+ cu en densi y in
TASK1/3 knockou cells could be he esul o an elec o-
physiological emodeling induced by he pe sis en
depola iza ion o hese cells. In his ega d, i is known
ha L- ype Ca2+ channel  subuni is down- egula ed by
ch onic depola iza ion in PC12 and smoo h muscle cells
and Pé ez-Ga cía, 2007). Ou da a also ag ee wi h p e i-
ous s udies epo ing ha he absence o TASK1 in ce -
ebella g anule neu ons has no e ec on any o he
pa ame e s (memb ane po en ial and esis ance, as well
as densi y o ol age-dependen channels) analyzed (Alle
e al., 2005; Mulkey e al., 2007). Abla ion o he TASK1
gene does no seem o induce up- egula ion o o he
TASK channels; howe e , some cen al neu ons lose
hei sensi i i y o pH o halo hane, hus sugges ing ha
o he K+ channels (o eplacemen o TASK1/TASK3
he e ome s by TASK3 homome s) compensa e o he
lack o TASK1 (Alle e al., 2005; Mulkey e al., 2007).
In con as , neu ons wi hou TASK3 (o TASK1/3) chan-
nels exhibi inc eased exci abili y and al e a ion in hei
i ing equency due o Na+ channel inac i a ion (B ickley
e al., 2007). These esul s, in acco d wi h he elec o-
physiological pheno ype obse ed in TASK1/3-null
glomus cells, sugges ha in some neu ons, TASK3 chan-
nels ( ha ha e la ge conduc ance han TASK1) a e ab-
solu ely equi ed o he main enance o hei no mal
es ing po en ial. In he case o mouse glomus cells, his
could be o c i ical impo ance because hey ha e a sig-
Figu e 7. Main enance o he sec e-
o y esponse o hypoxia in wild- ype
and TASK1/3/ glomus cells a e
blockade o di e en po assium chan-
nels. (A) Ampe ome ic eco ding and
co esponding cumula i e sec e ion
signal (in pC) o ca echolamine e-
lease induced by hypoxia (15 mm Hg)
in wild- ype CB glomus cells be o e and
a e blockade o TASK channels wi h
acidic pH. The composi ion o he con-
ol solu ion (pH 7.4) is indica ed in
Ma e ials and me hods and con ained
23 mM HCO3
. In he acidic (pH 6.5)
solu ion, his was eplaced wi h 10 mM
HCO3
 plus 12 mM NaH2PO4 and
1 mM Na2HPO4. Sec e ion a e du ing
hypoxia and acidic hypoxia we e 3,694
and 5,888 C/min, espec i ely. (B) Se-
c e o y esponse o hypoxia and co e-
sponding cumula i e sec e ion signal
(in pC) in TASK1/3-null mice glomus
cells in he p esence o he maxi-K+ chan-
nel blocke paxilline (500 nM). Sec e ion
a e du ing hypoxia was 3,193 C/min
and changed o 4,095 C/min in he
p esence o paxilline.
on Sep embe 22, 2014jgp. up ess.o gDownloaded om
Published Ma ch 29, 2010