basolateral glycylsarcosine (gly-sar) transport in caco-2 cell monolayers is ph dependent

Basolateral glycylsarcosine (Gly-Sar) transport in Caco-2 cellmonolayers is pH dependentRagna Berthelsen, Carsten Uhd Nielsen and Birger Brodin

Department of Pharmacy, The Faculty of Health and Medical Sciences, University of Copenhagen, Copenhagen, Denmark

Keywordsbasolateral peptide transporter; drug delivery;intestinal epithelium; proton-coupled peptidetransporter

CorrespondenceBirger Brodin, University of Copenhagen,Department of Pharmacy, Copenhagen,Denmark.E-mail: [email protected]

Received September 21, 2012Accepted February 11, 2013

doi: 10.1111/jphp.12061

Abstract

Objectives Transepithelial di/tripeptide transport in enterocytes occurs via theapical proton-coupled peptide transporter, hPEPT1 (SLC15A1) and a basolateralpeptide transporter, which has only been characterized functionally. In this studywe examined the pH dependency, substrate uptake kinetics and substrate specifi-city of the transporter.Methods We studied the uptake of [14C]Gly-Sar from basolateral solution intoCaco-2 cell monolayers grown for 17–22 days on permeable supports, at a rangeof basolateral pH values.Key findings Basolateral Gly-Sar uptake was pH dependent, with a maximaluptake rate at a basolateral pH of 5.5. Uptake of Gly-Sar decreased in the presenceof the protonophore nigericin, indicating that the uptake was proton-coupled.The uptake was saturable, with a maximal flux (Vmax) of 408 � 71, 307 � 25 and188 � 19 pmol/cm2/min (mean � S.E., n = 3) at basolateral pH 5.0, 6.0 and 7.4,respectively. The compounds Gly-Asp, Glu-Phe-Tyr, Gly-Glu-Gly, Gly-Phe-Gly,lidocaine and, to a smaller degree, para-aminohippuric acid were all shown toinhibit the basolateral uptake of Gly-Sar.Conclusions The study showed that basolateral Gly-Sar transport in the intesti-nal cell line Caco-2 is proton-coupled. The inhibitor profile indicated that thetransporter has broad substrate specificity.

Introduction

Transport of nutrients from the small-intestinal lumeninto the bloodstream is mediated by transport proteins inthe apical and basolateral membranes of the enterocytes.Transcellular transport of di- and tripeptides across intes-tinal epithelial cells is mediated by two distinctive systemspresent in the apical and basolateral membranes, respec-tively (i.e. the apical di-/tripeptide transporter hPEPT1and a basolateral peptide transporter). In recent years,PEPT1 has been well described as an apical proton-coupled peptide transporter capable of transporting notonly di- and tripeptides but also a wide range of peptido-mimetic drugs and bioactive peptide derivatives.[1–3] Thebasolateral peptide transporter has received less attentionand has only been characterized at the functional level atpresent, although one study has suggested a candidate pro-tein.[4] The basolateral exit has been described as a possiblerate-limiting step in the transcellular transport of di- andtripeptides.[5] Basolateral peptide transport in enterocytes

was originally described functionally in basolateral mem-brane vesicles from rabbit intestine[6] and later found inthe intestinal cell line Caco-2.[7,8] The basolateral peptidetransporter in Caco-2 cells has been described as beingcapable of transporting dipeptides, tripeptides and somepeptidomimetics, such as the anti-cancer drug bestatin anda number of cephalosporins.[9,10] Basolateral peptide trans-port capacity during Caco-2 cell culture reaches a steadylevel after 12 days of growth, as opposed to apicalhPEPT1-mediated peptide uptake which reaches a steadylevel at day 21.[11] The transporter seems to be under someregulation by the adenosine monophosphate (AMP)-activated protein kinase (AMPK)[12] and cyclic AMP(cAMP).[13] The transport mechanism of the basolateralpeptide transporter has not yet been resolved. The trans-port is described as simple facilitative transport in somestudies,[14,15] and as proton-coupled transport in otherstudies.[6,8]

bs_bs_banner

And PharmacologyJournal of Pharmacy

Research Paper

© 2013 Royal Pharmaceutical Society, Journal of Pharmacy and Pharmacology, 65, pp. 970–979970

mailto:[email protected]

The aim of this study was consequently to resolve thecontroversy regarding the pH dependency of the basolateralpeptide transport process. Furthermore, we wished to char-acterize the transporter in terms of its kinetic parameters atdifferent pH values and to expand the knowledge on basola-teral peptide transporter substrate specificity. Here wepresent data demonstrating that basolateral uptake ofdipeptides is proton-coupled. We also show for the firsttime that arginine, Gly-Asp, Glu-Phe-Tyr, Gly-Glu-Gly, Gly-Phe-Gly, lidocaine and, to a smaller degree, para-aminohippuric acid (PAA) inhibit basolateral peptideuptake.

Materials and Methods

Materials

Caco-2 cells were obtained from the American Type CultureCollection (Manassas, VA, USA). Cell culture media andHanks balanced salt solution (HBSS) were purchased fromLife Technologies (Taastrup, Denmark). Two batches of[3H]glycylsarcosine ([3H]Gly-Sar) with a specific activity of0.7 or 1.0 Ci/mmol, [3H]mannitol with a specific activityof 11.7 Ci/mmol and [14C]mannitol with a specific activityof 56.0 mCi/mmol were obtained from PerkinElmer Life(Boston, MA, USA). [14C]Glycylsarcosine ([14C]Gly-Sar)with a specific activity of 56.0 mCi/mmol was obtainedfrom Moravek Biochemicals (Brea, CA, USA). Glacialacetic acid, 1-butanol, ethanol, ninhydrin, l-arginine andthin-layer chromatography (TLC) plates (silica gel,20 ¥ 20 cm) were purchased from Merck (Darmstadt,Germany). l-Carnosine, Gly-Asp, Gly-Glu-Gly, Gly-Sar-Sar,Trp-Ala and Ala-Pro-Ala were obtained from BachemHolding AG (Bubendorf, Switzerland). 2-(Diethylamino)-N-(2,6-dimethylphenyl)acetamide (lidocaine) was pur-chased from Unikem A/S (Copenhagen, Denmark). Allother amino acids, dipeptides, tripeptides, 2-[(4-aminobenzoyl)amino]acetic acid (PAA), cephalexinhydrate, bovine serum albumin (BSA), 2-(N-morpholino)ethanesulfonic acid (MES) and N-[2-hydroxyethyl]piperazine-N′-[2-ethanosulfonate] (HEPES)were from Sigma-Aldrich Chemie (Steinheim, Germany).Nigericin was obtained from Invitrogen Molecular Probes(Taastrup, Denmark).

Caco-2 cell culture

Caco-2 cells at passage 20 were thawed from cryobatches,seeded in culture flasks and grown in Dulbecco’s ModifiedEagle’s medium (DMEM) supplemented with 10% fetalbovine serum, penicillin/streptomycin (100 U/ml and100 mg/ml, respectively), 1% l-glutamine and 1% non-essential amino acids. When cells reached passage 24–50they were seeded onto tissue-culture treated Transwell

filters (1.1 cm2, 0.4 mm pore size) at a density of 105 cells/cm2. The Transwell filter inserts were placed in T-12 culturetrays with 0.5 ml culture medium in the apical compart-ment and 1 ml culture medium in the basolateral compart-ment. Monolayer cultures were grown in an atmosphere of5% CO2–95% O2 at 37°C. Growth media were replacedevery other day and transepithelial electrical resistance(TEER) was measured during growth in tissue resistancemeasurement chambers (Endohm) with a voltohmmeter(EVOM), both of which were from World Precision Instru-ments (Sarasota, FL, USA). Uptake experiments were per-formed on Caco-2 cells of passages 24–50, day 17–21 afterseeding. TEER was measured before the uptake experi-ments, and only wells with a resistance above 300 W cm2

were used.

Gly-Sar uptake studies

Uptake of radiolabelled Gly-Sar was measured in HBSSsupplemented with 0.05% BSA. Apical media was bufferedwith 10 mm HEPES and adjusted to pH 7.4 if not statedotherwise. Basolateral medium was buffered with 10 mmMES or 10 mm HEPES and the pH was adjusted to pH 5.0,5.5, 6.0, 6.5, 7.0, 7.4 or 8.0, depending on the experiment.The cells were rinsed once in preheated HBSS and placedon a shaking plate (90 rpm), heated to 37°C. After 15 minequilibration, the experiment was initiated by removingthe apical and basolateral fluids by careful suction andadding 0.5 ml apical buffer and 1.0 ml basolateral buffercontaining 1.0 mCi [3H]/[14C]Gly-Sar per well. In someexperiments 0.25 mCi [3H]/[14C]mannitol was added tothe basolateral solution. In experiments performed with[3H]Gly-Sar, [14C]mannitol was used as an extracellularfluid space marker. In experiments with [14C]Gly-Sar,[3H]-labelled mannitol was used. Depending on theexperiment the basolateral buffer contained lidocaine,cephalexin or PAA, glycine, Ala-Gly, Gly-His, Gly-Pro, Gly-Glu-Gly, Gly-Sar-Sar, Trp-Ala, Ala-Pro-Ala, Gly-Sar andmannitol, at the concentrations stated in the figurelegends. In a series of experiments 20 mm of nigericin wasapplied to the basolateral and apical buffer solutions5 min before the start of the experiment. After 10 min,the experiment was terminated and the cell monolayerswere washed four times with ice-cold HBSS and trans-ferred to counting vials. Scintillation fluid was added(Ultima Gold; Packard, Canberra, Australia) and radioac-tivity was counted in a liquid scintillation analyser(Packard Tri-Carb, Canberra, Australia). The pH of theincubation buffers was measured before and after 10 minof incubation with the Caco-2 cell monolayers. A slightalkalinization was observed in the apical (2%) as well asthe basolateral (4%) solution, most likely due to influencefrom atmospheric CO2.

Ragna Berthelsen et al. Basolateral intestinal peptide transport

© 2013 Royal Pharmaceutical Society, Journal of Pharmacy and Pharmacology, 65, pp. 970–979 971

Evaluation of radiolabelled Gly-Sarbatch purity

TLC was used to estimate the content of free radiolabelledamino acids in the Gly-Sar isotope batches. The separationof amino acids and dipeptides was carried out by use ofconventional TLC. Three standards, containing 2 mg/ml ofglycine, sarcosine or Gly-Sar, were prepared by use of stocksolutions, and two test solutions were prepared containingall standards and either [3H]Gly-Sar or [14C]Gly-Sar withactivity 0.9 mCi/ml. The non-labelled Gly-Sar was previ-ously shown to have a glycine content of 0.1%.[16] Fiftymicrolitres of each solution were applied onto a silica plate(20 ¥ 20 cm). The mobile phase consisted of 1-butanol,glacial acetic acid and water in the ratio 4 : 1 : 1. Areas cor-responding to glycine and Gly-Sar were removed after therun, dissolved in 200 ml of methanol and 2 ml of scintilla-tion liquid and counted in a liquid scintillation analyser.

Data analysis and statistics

The uptake rate of Gly-Sar, as a function of basolateral Gly-Sar concentration, was fitted to a Michaelis–Menten typeequation:

VV S

K Sm

= ⋅[ ]+ [ ]

max(1)

where V is the flux (nmol/cm2/min), Vmax is the maximalflux (nmol/cm2/min), Km is the Michaelis constant (mm)and [S] is the concentration of Gly-Sar (mm) in the basola-teral solution. The specific uptake was calculated by sub-tracting the non-specific uptake, which was estimated in thepresence of 30 mm Gly-Pro, from the total uptake. Displace-ment curves were fitted by non-linear least-square regres-sion analysis and a one-site fit:

Y BottomTop Bottom

X LogIC= + −

+ −1 10 50( )(2)

where Y is maximal inhibition (%), Top is the degree ofinhibition in the absence of inhibitor (%), Bottom is themaximal degree of inhibition (%) and X is the concentra-tion of the inhibitor. IC50 values (the concentration ofcompetitor causing a 50% inhibition of the radio liganduptake) were subsequently determined from the fit. TheIC50 is related to the inhibition constant Ki, by the Cheng–Prusoff equation:[17]

IC KS

Ki

m50 1= ⋅ + [ ]⎛

⎝⎜⎞⎠⎟ (3)

where [S] is the concentration of substrate and Km is theMichaelis constant of the substrate.

The uptake rate of Gly-Sar at a given substrate andinhibitor concentration was calculated using the classicalrelationship for competitive inhibition:[18]

VV S

KI

KSm

I

= ⋅[ ]+ [ ]⎛

⎝⎜⎞⎠⎟ + [ ]

max

1(4)

where [I] is the concentration of inhibitor.All experiments were performed on three individual cell

passages (n = 3). Within each cell passage, each treatmentwas measured in triplicate (except for the inhibition datashown in Figure 4, where duplicates where made withineach of the three passages). Data were analysed by one-wayanalysis of variance followed by Dunnett’s test. P < 0.05 wasconsidered significant.

Results

pH dependency of basolateralGly-Sar uptake

To investigate the pH dependency of the basolateral peptidetransport process, a series of experiments was performed inwhich basolateral Gly-Sar uptake as a function of basola-teral pH was determined. Basolateral uptake rate of 18 mmGly-Sar (0.5 mCi/ml) was measured over a period of10 min in Caco-2 cell monolayers at six different basola-teral pH values ranging from 5.0 to 7.4. Simultaneously, theuptake of 18 mm Gly-Sar in the presence of 30 mm Gly-Prowas followed to verify that the uptake was mediated by atransporter. Figure 1 shows the basolateral uptake rate ofGly-Sar, corrected for extracellular isotope using the non-permeable extracellular space marker mannitol. The rela-tionship shows an increase in uptake as pH decreases, with amaximum at pH 5.5. Basolateral Gly-Sar uptake increasedfive-fold when basolateral pH was changed from 7.4 to 5.5.At all pH values the Gly-Sar uptake was inhibited by 30 mmGly-Pro, indicating that the uptake was carrier mediated viaa peptide transport mechanism. The degree of inhibition ofGly-Sar uptake caused by Gly-Pro was fairly constant,ranging from 3.7 to 6.0%. Overall, the Gly-Sar uptakeprofile as a function of pH in the absence or presence ofnon-labelled Gly-Pro indicated the presence of apH-dependent carrier system.

Basolateral Gly-Sar uptake was decreasedsignificantly in the absence of aproton gradient

The observed pH dependency of basolateral Gly-Sar uptakecould be due to an affinity change in the transporter or aproton coupling of the transport process. The nature of thepH dependency was therefore investigated by applying

Ragna Berthelsen et al.Basolateral intestinal peptide transport


20 mm of the electro-neutral ionophore nigericin alongwith 20 mm [3H]Gly-Sar to the basolateral chamber(Figure 2). As evident from Figure 2, the basolateral Gly-Saruptake rate was profoundly diminished in the presence ofnigericin at pH 5.0 and 6.0, but not at pH 7.4 or 8.0. Gly-Sar uptake at pH 6.0 was reduced to 36% of the valueobtained in the absence of nigericin, and to 16% at pH 5.0.The absolute Gly-Sar uptake values in the presence of

nigericin at all pH values were of the same magnitude (witha mean of 0.56 � 0.1 pmol/cm2/s, n = 3) and did not differsignificantly, as tested with a one-way analysis of variance.This would be expected if the concentration gradient fromthe basolateral solution into the cells was the only drivingforce for uptake. The effect of nigericin on the integrityof the Caco-2 cell monolayers was evaluated usingTEER measurements before and after nigericin treatment.There was no significant difference between before(348 � 15 W cm2) and after (340 � 15 W cm2) treatment(n = 6, P = 0.39), which indicates that nigericin did notdisrupt the cell monolayers. The decrease in Gly-Sar uptakewhen the basolateral proton gradient was shunted bynigericin indicated that pH dependency of the basolateralpeptide transporter was due to a proton-coupled transportprocess.

Lowering pH of the basolateral solutionincreased Vmax of basolateral Gly-Sar uptake

The concentration-dependent uptake of Gly-Sar at threedifferent basolateral pH values (5.0, 6.0 and 7.4) was meas-ured, using 18 mm [14C]Gly-Sar (1.0 mCi/ml) and non-labelled Gly-Sar in the concentration range 0.02–10 mm.Basolateral uptake was measured for 10 min in the pres-ence and absence of 30 mm Gly-Pro at 37°C and apicalpH 7.4 (Figure 3a–c). Basolateral uptake of Gly-Sar in thepresence of 30 mm of Gly-Pro was used to determine non-specific uptake, which was present at Gly-Sar concentra-tions higher than 0.5 mm. The non-specific uptake wassubsequently subtracted from the uptake values obtained inthe absence of Gly-Pro and plotted as specific uptake inFigure 3d. The specific Gly-Sar uptake as a function ofbasolateral Gly-Sar concentration was fitted to theMichaelis–Menten equation (Figure 3d). At pH 7.4, the Vmax

was 188 � 19 pmol/cm2/min (n = 3). Lowering the pH to6.0 caused the Vmax to increase to 307 � 25 pmol/cm2/min(n = 3). At pH 5.0, the Vmax was 408 � 71 pmol/cm2/min(n = 3). The Km value was 4.3 � 1 mm, 0.9 � 0.3 mm and1.9 � 1 mm at pH 7.4, 6.0 and 5.0, respectively. The Vmax

values at pH 6.0 and pH 5.0 were significantly higher thanthe Vmax value at pH 7.4 (P < 0.05), but the Vmax at pH 6.0did not differ significantly from the value at pH 5.0. The Km

value was also significantly lower at pH 6.0 than at pH 7.4(P < 0.05), however the Km at pH 5.0 did not differ signifi-cantly from the value at pH 7.4, due to the variance. Nor didthe Km at pH 6.0 and pH 5.0 differ significantly. The Gly-Saruptake as a function of basolateral Gly-Sar concentrationshowed a Michaelis–Menten-like relationship, indicative ofa carrier system. The Vmax of the transport increased signifi-cantly when the pH of the basolateral solution was loweredfrom pH 7.4 to pH 6.0 or pH 5.0, indicative of a proton-driven uptake process.

4 5 6 7 80

2

4

6

8

Basolateral pH

Bas

ola

tera

l G

ly-S

aru

pta

ke r

ate

(pm

ol /

cm2/ m

in)

Figure 1 Basolateral uptake rate of Gly-Sar as a function of pH in thebasolateral solution in Caco-2 cell monolayers grown for 17–22 days.Uptake of 18 mM [14C]Gly-Sar (1.0 mCi/ml) at 37°C for 10 min wasmeasured in the presence (�) or absence (�) of 30 mM Gly-Pro atvarying basolateral pH values (5.0–7.4). The values were corrected forisotope remaining in the extracellular solution by using the extracellularspace marker [3H]mannitol (0.25 mCi/ml). The pH of the apical solutionwas 7.4 throughout the experiment. Error bars represent mean � S.E.of three individual passages.

4 5 6 7 8 90

1

2

3

4***

***

Basolateral pH

Bas

olat

eral

Gly

-Sar

upta

ke r

ate

(pm

ol/c

m2 /m

in)

Figure 2 Basolateral uptake rate of Gly-Sar as a function of a basola-teral pH in the absence and presence of 20 mM of the protonophorenigericin. Caco-2 cell monolayers, grown for 17–22 days, were incu-bated with 20 mM [3H]Gly-Sar (1.0 mCi/ml) for 10 min in the presence(D) or absence (�) of 20 mM nigericin. The values were corrected forisotope remaining in the extracellular solution by using the extracellularspace marker [14C]mannitol (0.25 mCi/ml). Error bars representmean � S.E. of three individual passages. ***P < 0.001, nigericin vs.control (absence of nigericin).



Substrate specificity of the basolateralpeptide transporter

To assess substrate specificity of the basolateral peptidetransporter, we determined uptake of 18 mm [14C]Gly-Sar(0.5 mCi/ml) in the absence and presence of 5 mm ofvarious amino acids, di- and tripeptides with very differentsize and polarity, and drug substances lidocaine, cephalexinand PAA (Figure 4). The basolateral uptake of Gly-Sar wasinhibited to various degrees by all the tested di- and tripep-tides as well as the amino acid arginine, the drug lidocaineand the organic anion transporter substrate PAA (Figure 4).The purity of the radioisotopes was determined by use ofTLC. [3H]Gly-Sar contained 98% of the quantity specifiedon the label in the form of Gly-Sar and 2% as glycine; thesame was true for [14C]Gly-Sar. Basolateral Gly-Sar uptakewas not inhibited by 5 mm glycine and it was concludedthat the small isotope impurity did not affect the inhibitionexperiments. The inhibition of Gly-Sar uptake by Gly-Pro,

Ala-Gly and Gly-His was characterized further (Figures 5and 6) and affinities were calculated from competitionexperiments. All three dipeptides inhibited Gly-Sar uptakein a concentration-dependent manner (Figures 5 and 6),and inhibition constants could be calculated using a one-binding-site fit and least-square regression analysis. TheLog IC50 values for Gly-Pro, Ala-Gly and Gly-Hiswere 2.8 � 0.11 mm, 2.6 � 0.15 mm and 3.2 � 0.14 mm(mean � S.E., n = 3), respectively, corresponding to Ki

values of 2.7 mm, 2.5 mm and 3.1 mm, as calculated usingthe Cheng–Prusoff equation for competitive interactions(Equation 2). Uptake of Gly-Sar in the presence of Gly-Hiswas tested at two different basolateral pH values; affinitieswere found to be identical, indicating that the Ki value ofGly-His was not dependent on the side-chain charge on thehistidine residue (pKa ~6.1). To investigate whether osmoticpressure, caused by the presence of the inhibitors in thebasolateral solution, could influence the basolateral uptakerate of Gly-Sar, we examined the effect of applying varying

Bas

ola

tera

l Gly

-Sar

up

take

rat

e(p

mo

l/cm

2 /min

)

Basolateral [Gly-Sar] (mM)

Bas

ola

tera

l Gly

-Sar

up

take

rat

e(p

mo

l/cm

2 /min

)

Bas

ola

tera

l Gly

-Sar

up

take

rat

e(p

mo

l/cm

2 /min

)

Bas

ola

tera

l Gly

-Sar

up

take

rat

e(p

mo

l/cm

2 /min

)

pH 5.0 pH 6.0

pH 6.0pH 7.4

(c) (d)

(a) (b)

pH 7.4

pH 5.0

0 2 4 6 8 10

Basolateral [Gly-Sar] (mM)0 2 4 6 8 10

500

400

300

200

100

0


500

400

300

200

100

0

500

400

300

200

100

0


500

400

300

200

100

0

Figure 3 Basolateral uptake rate of Gly-Sar as a function of Gly-Sar concentration at basolateral pH 5.0 (a), 6.0 (b) and 7.4 (c) in Caco-2 cell mon-olayers grown for 17–22 days. The cells were incubated with 18 mM [14C]Gly-Sar (1.0 mCi/ml) and unlabelled Gly-Sar in a concentration range of0.02–10 mM for 10 min in the presence (�) or absence (�) of 30 mM Gly-Pro at 37°C and apical pH 7.4. The values were corrected for isotoperemaining in the extracellular solution by using the extracellular space marker [3H]mannitol (0.25 mCi/ml). The peptide transporter component of thebasolateral uptake rate was found as Gly-Sar uptake in the presence of 30 mM Gly-Pro, subtracted from the Gly-Sar uptake in the absence of 30 mM

Gly-Pro (d). The obtained values were fitted to the Michaelis–Menten equation (Equation 1). Error bars represent mean � S.E. of three individualpassages.



concentrations of mannitol to the basolateral solution,together with 20 mm [3H]Gly-Sar. The basolateral Gly-Saruptake rate was not affected by increases in basolateral solu-tion osmolarity (Figure 5).

Estimation of basolateral-to-apical flux ofradiolabelled Gly-Sar during basolateralGly-Sar uptake experiments

Basolateral peptide transport uptake experiments are tech-nically challenging, due to the risk of paracellular passage ofradiolabelled Gly-Sar from the basolateral to the apicalsolution, where uptake via hPEPT1 could influence theresults. To assess the possible contribution of Gly-Saruptake via hPEPT1 we determined the apical uptake of20 mm (1 mCi/ml) [3H]Gly-Sar for 10 min at apical pH 7.4and calculated the apical uptake rate to be 0.2 pmol/cm2/min. Using the same batch and passage of cells we thenmeasured the transepithelial flux of Gly-Sar in the basal-to-apical (B–A) direction over a period of 10 min when20 mm (1 mCi/ml) [3H]Gly-Sar was added to the basola-teral solution at time zero. The concentration of Gly-Sar inthe apical solution was found to be 0.4 mm (0.02 mCi/ml)

after 10 min of transepithelial B–A Gly-Sar transport.Apical uptake of Gly-Sar as a function of apical Gly-Sarconcentration follows the Michaelis–Menten expression.[19]

There is an almost linear relationship between apical Gly-Sar concentration and apical uptake at apical Gly-Sar con-centrations below 20 mm. It follows that if we observe anapical uptake of 0.2 pmol/cm2/min at an apical Gly-Sarconcentration of 20 mm, the slope of this linear relationshipmust be (0.2 pmol/cm2/min)/20 mm = 0.01 pmol/cm2/min/mm. The expression describing apical uptake rate(AUR) as a function of apical Gly-Sar concentration (AGC)in the range 0–20 mm Gly-Sar can therefore be stated to beas follows: AUR (pmol/cm2/min) = AGC (mm) ¥ 0.01(1 pmol/cm2/min/mm).

The AUR during a basolateral uptake experiment couldthus be estimated from the concentration of Gly-Sar in theapical compartment present at the end of the experiment.An estimated AUR at an AGC of 0.4 mm could therefore becalculated to be 0.004 pmol/cm2/min. This is an overesti-mate, since it is assumed in this approach that the concen-tration of Gly-Sar is 0.4 mm throughout the experiment.The basolateral Gly-Sar uptake values in the experimentswere typically 0.5 pmol/cm2/min or higher (see Figures 1

Co

ntr

ol

Arg

inin

e

Gly

cin

e

His

tid

ine

Lysi

ne

Phen

ylal

anin

e

Pro

line

Sarc

osi

ne

Ala

-Gly

Car

no

sin

e

Gly

-Asp

Gly

-His

Gly

-Pro

Gly

-Sar

Trp

-Ala

Ala

-Pro

-Ala

Glu

-Ph

e-Ty

r

Gly

-Glu

-Gly

Gly

-Ph

e-G

ly

Gly

-Sar

-Sar

Cep

hal

exin

Lid

oca

ine

PAA

0

1

2

3

4

***

***

******

******

***

***

******

***

***

***

***

**

Bas

ola

tera

l up

take

rat

e

(p

mo

l/cm

2 /min

)

Figure 4 Effect of amino acids, small peptides and peptide-like compounds on Gly-Sar uptake rate in Caco-2 cells. Caco-2 cell monolayers cul-tured for 17–22 days were incubated with 18 mM [14C]Gly-Sar (1.0 mCi/ml) for 10 min at 37°C at pH 6.0 in the absence (black bar) or presence(hatched bars) of 5 mM inhibitors added to the basolateral side. The values were corrected for isotope remaining in the extracellular solution byusing the extracellular space marker [3H]mannitol (0.25 mCi/ml). The pH of the apical solution was 7.4 in all experiments. Error bars representmean � S.E. of three individual passages. **P < 0.01; ***P < 0.001, test compound vs. control.



and 2) as compared with the corresponding apical uptakevalue of 0.004 pmol/cm2/min. The contribution fromapical uptake via PEPT1 during basolateral uptake experi-ments was therefore considered to be negligible.

Discussion

pH dependency of the basolateralpeptide transporter

The pH dependency of the basolateral peptide transportsystem in Caco-2 cells was investigated. The uptake kineticsshowed a Michaelis–Menten-like relationship, indicative ofcarrier-mediated transport. We observed an increase inbasolateral Gly-Sar uptake when the pH of the basolateralsolution was lowered. The Vmax rose from 188 to 307 pmol/cm2/min when the basolateral pH was changed from 7.4 to5.0. The Km value of the basolateral peptide transporterdecreased somewhat at lower pH values. The Km valuesobtained in this study are comparable with values fromthose found previously.[5,6,15] In accordance with previousstudies of Gly-Pro uptake in vesicles[6] and studies of baso-lateral Gly-Sar uptake in Caco-2 cell monolayers,[7,8] ourexperiments strongly indicated that the basolateral uptakeof Gly-Sar was pH dependent and that the transportprocess of Gly-Sar via the basolateral peptide transportsystem was pH dependent. A pH dependency could intheory be the result of a proton coupling of the substratetransport or a change in substrate affinity of the transporterdue to conformational changes induced by the changed pH.Nigericin abolished the effect of pH on Gly-Sar uptake,

–4 –2 0 2 4 60

50

100

150

log [Ala-Gly (µM)]

0 25 50 75 1000

2

4

6

8

Basolateral [mannitol] (mM)

Bas

oal

ater

alG

ly-S

aru

pta

ke r

ate

–4 –2 0 2 4 60

50

100

150

log [Gly-Pro (µM)]

(a)

(b)

(c)

(p

mo

l/cm

2/m

in1 )

Bas

ola

tera

l Gly

-Sar

up

take

rat

e(%

of

con

tro

l)

Bas

ola

tera

l Gly

-Sar

up

take

rat

e(%

of

con

tro

l)

Figure 5 Basolateral Gly-Sar uptake rate as a function of Gly-Pro,Ala-Gly and mannitol concentrations in Caco-2 cell monolayers cul-tured for 17–22 days. The cells were incubated with varying concentra-tions of Gly-Pro (a), Ala-Gly (b) and mannitol (c), as well as 20 mM

[3H]Gly-Sar (1 mCi/ml) for 10 min at apical pH 7.4 and basolateralpH 6.0. The values were corrected for isotope remaining in the extra-cellular solution by using the extracellular space marker [14C]mannitol(0.25 mCi/ml). The inhibition-Log(concentration) relationships for Gly-Pro and Ala-Gly were fitted to a one-site fit (Equation 2) and the IC50value was obtained. Ki values were determined using the Cheng–Prusoff equation (Equation 3). The dotted lines represent the one-sitefit. a and b. Error bars represent mean � S.D. of three individual deter-minations within a passage. c. Error bars represent � S.E. of three indi-vidual passages.

–4 –2 0 2 4 60

50

100

150 pH 5.5

pH 7.4

log [Gly-His (µM)]

Bas

ola

tera

l Gly

-Sar

up

take

(% o

f co

ntro

l)

Figure 6 Basolateral Gly-Sar uptake rate as a function of Gly-His con-centration in Caco-2 cell monolayers cultured for 17–22 days. The cellswere incubated with varying concentrations of Gly-His, as well as18 mM [14C]Gly-Sar (1 mCi/ml), for 10 min at apical pH 7.4, and basola-teral pH at 5.5 or 7.4. The values were corrected for isotope remainingin the extracellular solution by using the extracellular space marker[3H]mannitol (0.25 mCi/ml). Error bars represent the S.E. of three indi-vidual passages.



indicating that a proton gradient was the driving force forthe basolateral Gly-Sar uptake. A similar proton couplinghas been demonstrated for the di/tripeptide transporterPEPT1 and the amino acid transporter PAT1.[15,20,21]

In contrast, other studies on Caco-2 cell monolayershave reported little or no effect of a basolateral pH gradi-ent.[5,10,15] The most plausible reason for the discrepancy isan apparent difference in the experimental set up used inthe various studies. In the studies where no pH depend-ency was observed, the apical pH was kept at 6.0 duringuptake experiments.[5,10,15] In the present study, and in thestudies performed by Thwaites and co-workers,[7,8] apicalpH was kept at 7.4. In the study by Dyer et al., using baso-lateral membrane vesicles, the pH gradient was fully con-trolled by buffers, with a pH at the cytosolic face of 7.5and a pH at the basolateral side of 5.5.[6] Thwaites et al.showed that apical acidification (pH 6.0) causes a decreasein the intracellular pH of Caco-2 cell monolayers,[22] andthat a change in extracellular pH influences transport andcellular accumulation of Gly-Sar in both directions (A–Band B–A).[7] For example, a change in apical pH from 6.0to 7.4 with basolateral pH 7.4 doubled the cellular accu-mulation from the basolateral compartment.[7] Based onthis observation, the difference in applied apical buffersolution pH seems a probable explanation for the observeddiscrepancies. The apical pH of 6.0 used in studies by Irie,Saito, Terada and co-workers,[10,14,15] might have resulted ina lower proton gradient across the basolateral membrane,due to a more acidic cell interior, as compared with cellsexposed to a pH of 7.4 in the apical solution. However, thediscrepancy between the results on pH dependency mightalso be due to different culture conditions and experimen-tal techniques, as previously suggested.[15] The Vmax and Km

values were measured in the blood-to-cell direction, andmight differ from values that would be observed in thecell-to-blood direction. Therefore the data suggest a carriermechanism with a proton coupling, but the actual cell-to-blood Vmax and Km cannot be estimated from the presentstudy.

The basolateral uptake of Gly-Sar showed a pH depend-ency resembling that of hPEPT1 in the apical mem-brane.[15,23] There are, however, several studies showing thathPEPT1 is exclusively located in the apical membrane ofCaco-2 cells grown on permeable supports, under the sameexperimental conditions as decribed in the present study. Ina previous study we observed apical and cytosolic labellingof hPEPT1 but no basolateral membrane labelling.[19]

Walker et al. showed, using confocal laser scanning micro-scopy (CLSM) and a custom-designed antibody against thehPEPT1 C-terminus, that hPEPT1 was expressed exclusivelyin the apical membrane of Caco-2 cells and in the apicalmembrane of human enterocytes from tissue samples.[24] Byuse of electron microscopy, with a resolution exceeding that

of CLSM by several orders of magnitude, Pieri et al. alsodemonstrated apical and cytosolic labelling but no hPEPT1labelling in the basolateral membrane of filter-grownCaco-2 cells..[12] It might also be argued that the basolateralpeptide transport system could be a cell-culture artefact.However, the transport system was first demonstrated in apreparation of basolateral membrane vesicles from rabbitenterocytes.[6] Furthermore, human intestinal absorption ofintact peptides, such as Gly-Gly-Gly and l-carnosine(b-Ala-His), has been known for decades,[25] indicating thepresence of both an apical and a basolateral transportsystem for di/tripeptides. Transepithelial transport of intactdi/tripeptides and substrates has also been shown in anumber of animal studies and intestinal in-situ models.[25]

The basolateral peptide transporter showedbroad substrate specificity

The basolateral uptake of Gly-Sar was inhibited to variousdegrees by a wide range of di- and tripeptides, which hasalso been reported in other studies.[6,26–28] This indicates thatthe basolateral peptide transporter has broad substrate spe-cificity, not unlike PEPT1, as previously described.[15] To ourknowledge this study is the first in which arginine, Gly-Asp,Glu-Phe-Tyr, Gly-Glu-Gly, Gly-Phe-Gly and lidocaine havebeen investigated as possible substrates for the basolateraltransporter. All these compounds significantly inhibited thebasolateral uptake of Gly-Sar. Gly-His was tested at pHvalues 7.4 and 5.5, and no difference in Ki value wasobserved, indicating that the charge on the histidine residuedid not affect binding. Among the dipeptides, Ala-Gly, Gly-His, Gly-Pro and Trp-Ala were the most potent inhibitors ofbasolateral Gly-Sar uptake. Gly-Asp was also an inhibitor,indicating that the transporter is not sensitive to changes inthe dipeptide amino acid side chains. Lidocaine and, to asmaller degree, PAA also inhibited basolateral Gly-Saruptake. PAA contains an amide bond and a C-terminalcarboxylic acid, and the N-terminal amino acid isp-aminobenzoic acid. The pKa value of the N-terminalaniline is approximately 3, and the amino group wouldtherefore be uncharged at pH 6.0.[29] Overall, dipeptidomi-metics lacking either a charged N- or C-terminal were alsoable to inhibit the basolateral uptake of Gly-Sar. The basola-teral transporter recognized the investigated tripeptidesregardless of the physicochemical properties of the consti-tuting amino acid. Surprisingly, arginine turned out to bean inhibitor of basolateral Gly-Sar uptake. The inhibitorprofile of the basolateral peptide transporter thus differsfrom that of PEPT1 in the apical membrane. Uptake viaPEPT1 is not inhibited by arginine or PAA, whereas both d-and l-cephalexin are PEPT1 substrates. However, furtherstudies are needed to yield a detailed structure–affinity rela-tionship for the basolateral peptide transporter.



Possible physiological roles of thebasolateral peptide transport systemin enterocytes

A proton-coupled peptide transporter placed at the basola-teral cell membrane does not seem efficient from a tele-ological perspective, considering an average intracellular pHof 7.2 and an average pH in the bloodstream of 7.4. A baso-lateral peptide transporter coupled to a proton gradientwould thus not gain much driving force form the protongradient under these conditions. The electrogenicity of thebasolateral peptide transport system has not yet been inves-tigated. If the system mediates net positive charge transfer,as would be expected from the movement of an overallneutral substrate and at least one proton, the negative mem-brane potential would also counteract efflux from the cellinterior to the blood. However, the overall transport processof di- or tripeptides from the lumen into the bloodstreamcould be coupled in such a fashion that PEPT1, when trans-porting peptides and protons into the cell, could create anacidification of the cell along with a depolarization, thusincreasing the driving force for basolateral peptide exit. Theenterocytes express a number of apical proton-coupledtransporters[30] and it is possible that the enterocytes acidifyupon an amino acid and peptide load. Acidification of theepithelial cells during consumption of a meal and, conse-quently, an outward-directed proton gradient over the baso-lateral cell membrane of the epithelial cells could thus intheory be a way of coupling entry and exit of small pep-tides. Alternatively, the physiological role of the basolateralpeptide transporter in the enterocytes may be to mediateclearance of di/tripeptides from the blood. A basolateraldipeptide transporter has been described in the MadinDarby Canine Kidney (MDCK) cell line and in basolateralmembrane vesicles from rat kidney cortex.[26,27] There is asubstantial peritubular uptake of dipeptides in perfused rat

kidney[31] and it has thus been suggested that the physiologi-cal role of the renal basolateral transporter is to participatein the clearance of di/tripeptides from the blood.[26] Theintestinal basolateral peptide transporter might have asimilar role, mediating transport of peptides arising fromendogenous breakdown from the blood into the entero-cytes, where hydrolysis can occur due to the high content ofcytosolic peptidases.

Conclusions

The results of this study show that basolateral peptide trans-port in the intestinal cell line Caco-2 is indeed pH depend-ent and that the pH dependency is due to a proton-coupledsubstrate translocation. Competition studies with radiola-belled Gly-Sar indicated a broad substrate specificity of thebasolateral peptide transporter and demonstrated for thefirst time that arginine, Gly-Asp, Glu-Phe-Tyr, Gly-Glu-Gly,Gly-Phe-Gly, lidocaine and PAA are inhibitors of the basola-teral peptide transporter.

Declarations

Conflict of interest

The Author(s) declare(s) that they have no conflicts ofinterest to disclose.

Funding

This work was supported by the Predicting Drug Absorp-tion Consortium (The Danish Strategic Research Council)and the Carlsberg Foundation (Grant no. 2009-01-0635).

Acknowledgements

The authors acknowledge the expert assistance of techni-cians Birgitte Eltong and Maria Læssøe Pedersen.

References

1. Brodin B et al. Transport of peptido-mimetic drugs by the intestinal Di/tri-peptide transporter, PepT1. PharmacolToxicol 2002; 90: 285–296.

2. Sala-Rabanal M et al. Molecular inter-actions between dipeptides, drugs andthe human intestinal H+ -oligopeptidecotransporter hPEPT1. J Physiol 2006;574: 149–166.

3. Geissler S et al. The bioactive dipap-tide anserine is transported by humanproton-coupled peptide transporters.FEBS J 2010; 277: 790–795.

4. Shepherd EJ et al. Identification of a

candidate membrane protein for thebasolateral peptide transporter of therat small intestine. Biochem BiophysRes Commun 2002; 296: 918–922.

5. Irie M et al. Efflux properties of baso-lateral peptide transporter in humanintestinal cell line Caco-2. PflugersArch 2004; 449: 186–194.

6. Dyer J et al. Glycyl-L-proline trans-port in rabbit enterocyte basolateral-membrane vesicles. Biochem J 1990;269: 565–571.

7. Thwaites DT et al. Transepithelial gly-cylsarcosine transport in intestinalCaco-2 cells mediated by expression ofH+-coupled carriers at both apical

and basal membranes. J Biol Chem1993; 268: 7640–7642.

8. Thwaites DT et al. H+-coupled dipep-tide (glycylsarcosine) transport acrossapical and basal borders of humanintestinal Caco-2 cell monolayersdisplay distinctive characteristics.Biochim Biophys Acta 1993; 1151: 237–245.

9. Inui K et al. Transepithelial trans-port of oral cephalosporins by monol-ayers of intestinal epithelial cell lineCaco-2: specific transport systems inapical and basolateral membranes.J Pharmacol Exp Ther 1992; 261: 195–201.



10. Saito H, Inui K. Dipeptide transport-ers in apical and basolateral mem-branes of the human intestinal cellline Caco-2. Am J Physiol 1993; 265:G289–G294.

11. Bravo SA et al. In-depth evaluation ofGly-Sar transport parameters as afunction of culture time in the Caco-2cell model. Eur J Pharm Sci 2004; 21:77–86.

12. Pieri M et al. The apical (hPepT1) andbasolateral peptide transport systemsof Caco-2 cells are regulated by AMP-activated protein kinase. Am J Physiol2010; 299: G136–G143.

13. Henderson FD et al. cAMP-dependentregulation of basolateral glycyl-sarcosine uptake by human intestinalCaco-2 cell monolayers. J Physiol 2003;549p, PC9.

14. Irie M et al. Recognition and transportcharacteristics of nonpeptidic com-pounds by basolateral peptide trans-porter in Caco-2 cells. J PharmacolExp Ther 2001; 298: 711–717.

15. Terada T et al. Functional characteris-tics of basolateral peptide transporterin the human intestinal cell lineCaco-2. Am J Physiol 1999; 276:G1435–G1441.

16. Frølund S et al. d-aminolevulinic acidis a substrate for the amino acid trans-port SLC36A1 (hPAT1). Br J Pharma-col 2010; 159: 1339–1353.

17. Cheng Y, Prusoff WH. Relationshipbetween the inhibition constant (K1)and the concentration of inhibitor

which causes 50 per cent inhibition(I50) of an enzymatic reaction.Biochem Pharmacol 1973; 22: 3099–3108.

18. Segel IH. Enzyme Kinetics. New York:Wiley Interscience. John Wiley &Sons, Inc, 1975.

19. Nielsen CU et al. Epidermal growthfactor inhibits glycylsarcosine trans-port and hPepT1 expression in ahuman intestinal cell line. Am J Physiol2001; 281: G191–G199.

20. Chen XZ et al. Functional roles of his-tidine and tyrosine residues in theH(+)-peptide transporter PepT1.Biochem Biophys Res Commun 2000;272: 726–730.

21. Metzner L et al. Mutational analysis ofhistidine residues in the humanproton-coupled amino acid trans-porter PAT1. Biochim Biophys Acta2008; 1778: 1042–1050.

22. Thwaites DT et al. Direct assessmentof dipeptide/H+ symport in intacthuman intestinal (Caco-2) epithe-lium: a novel method utilisingcontinuous intracellular pH measure-ment. Biochem Biophys Res Commun1993; 194: 432–438.

23. Matsumoto S et al. Transcellulartransport of oral cephalosporins inhuman intestinal epithelial cells,Caco-2: interaction with dipeptidetransport systems in apical and baso-lateral membranes. J Pharmacol ExpTher 1994; 270: 498–504.

24. Walker D et al. Substrate upregulationof the human intestinal peptide trans-porter, hPepT1. J Physiol 1998; 507:697–706.

25. Matthews DM. Intestinal absorptionof peptides. Physiol Rev 1973; 55: 537–608.

26. Terada T et al. Functional expressionof a novel peptide transporter in renalbasolateral membranes. Am J Physiol2000; 279: F851–F857.

27. Sugarawara M et al. Uptake of dipep-tide and beta-lactam antibiotics bythe basolateral membrane vesiclesprepared from rat kidney. BiochimBiophys Acta 2003; 1609: 39–44.

28. Sawada K et al. Distinct transportcharacteristics of basolateral peptidetransporters between MDCK andCaco-2 cells. Pflugers Arch 2001; 443:31–37.

29. Robinson D et al. The apparent disso-ciation constants of some glucuro-nides, mercapturic acids and relatedcompounds. Biochem J 1953; 55: 151–155.

30. Thwaites DT et al. H+-coupled nutri-ent, micronutrient and drug trans-porters in the mammalian smallintestine. Exp Physiol 2007; 92: 603–619.

31. Lowry M et al. Metabolism of glycine-and hydroxyproline-containing pep-tides by the isolated perfused ratkidney. Biochem J 1985; 229: 545–549.



basolateral glycylsarcosine (gly-sar) transport in caco-2 cell monolayers is ph dependent

Documents