Active Sulforhodamine 101 Uptake into Hippocampal Astrocytes

Sulforhodamine 101 (SR101) is widely used as a marker of astrocytes. In this study we investigated labeling of astrocytes by SR101 in acute slices from the ventrolateral medulla and the hippocampus of transgenic mice expressing EGFP under the control of the astrocyte-specific human GFAP promoter. While SR101 efficiently and specifically labeled EGFP-expressing astrocytes in hippocampus, we found that the same staining procedure failed to label astrocytes efficiently in the ventrolateral medulla. Although carbenoxolone is able to decrease the SR101-labeling of astrocytes in the hippocampus, it is unlikely that SR101 is taken up via gap-junction hemichannels because mefloquine, a blocker for pannexin and connexin hemichannels, was unable to prevent SR101-labeling of hippocampal astrocytes. However, SR101-labeling of the hippocampal astrocytes was significantly reduced by substrates of organic anion transport polypeptides, including estron-3-sulfate and dehydroepiandrosterone sulfate, suggesting that SR101 is actively transported into hippocampal astrocytes.


Introduction
Astrocytes are important players in neuronal networks. They maintain the extracellular milieu by removal of potassium and neurotransmitters like glutamate and modulate synaptic transmission by releasing gliotransmitters [1,2,3]. Unequivocal identification of astrocytes for imaging experiments as well as electrophysiological recordings was facilitated by the availability of transgenic mouse lines expressing fluorescent proteins under the control of glia specific promoters [4,5]. Recently, sulforhodamine 101 (SR101) has been used for specific visualization of astrocytes in cortex and hippocampus [6,7,8,9]. In the spinal cord, however, SR101 has been described as unspecific [10] and it was also reported to label oligodendrocytes in the rabbit retina [11]. Furthermore, during hypoxia SR101 can enter also hippocampal neurons via hemichannels [12]. Currently, it is not known how the specific loading of SR101 into astrocytes is achieved under normoxic conditions. Although the synthetic glycyrrhetinic acid derivative carbenoxolone has been shown to block SR101-labeling of astrocytes [7], it is not yet clear if this is due to its action on gapjunctions and hemichannels or a yet unknown, maybe indirect mechanism. Nevertheless, SR101 is widely used as a marker for astrocytes in different brain regions [13,14,15] even despite the fact that sulforhodamine 101 has been found to trigger epileptic activity in the hippocampus [16].
In this study, we initially aimed to use SR101 for identification of astrocytes in ventrolateral medulla (VLM). Since VLM astrocytes were not labeled by SR101 sufficiently, we tried to improve the staining by pharmacological manipulations of potential mechanisms that might be involved in export of SR101 from medullary astrocytes. The results guided us to a series of experiments in the hippocampus to unveil the potential mechanism and functional role of this regional heterogeneity of astrocytes in ventrolateral medulla and the hippocampus.

Ethics statement
In accordance with the German Protection of Animals Act (Tierschutzgesetz; TierSchG 14 Abs. 3) we did not need formal approval for the post mortem removal of brain tissue. The experiments were communicated to and notified by animal welfare office of University Medical Center Göttingen, Germany (institutional act number: T19.08).

Breeding of mice
Animals were hold and bred in the animal facilities of the University Hospital Göttingen in accordance with guidelines of the German Physiological Society as well as the regulations of the State of Lower Saxony and the Federal Republic of Germany. Experiments were performed on acute brain slice preparations of neonatal (P2-P12), juvenile (P29-33) and adult mice (P98-99) expressing the enhanced green fluorescent protein in astrocytes [Tg(hGFAP-EGFP)GFEC-Fki; [4]]. Additionally we used mice in which glycinergic neurons were labeled [Tg(Slc6a5-EGFP)1Uze; [17]] to unequivocally identify inhibitory neurons.

Slice preparations
Acute transversal slices from brainstem and hippocampus were prepared as described previously [18]. Briefly, animals were decapitated under diethyl-ether anesthesia, brainstem and hippocampus were isolated and placed in ice-cooled, carbogen-saturated (95% O 2 , 5% CO 2 ) artificial cerebrospinal fluid (aCSF) containing 118 mM NaCl, 3 KCl, 1.5 mM CaCl 2 , 1 mM MgCl 2 , 1 mM NaH 2 PO4, 25 mM NaHCO 3 , and 30 mM D-glucose). The osmolarity was 325-335 mosm/l and the pH was adjusted to 7.4. The isolated brain part was glued with cyanoacryl glue (Loctite Deutschland GmbH) to an agar block and mounted in a vibroslicer (VT 1000S or VT 1200S, Leica). Brainstem and hippocampal slices of 250 mm were cut and stored in oxygenated aCSF at room temperature for at least 30 minutes before staining. For the actual experiments, slices were transferred to a recording chamber at the regarding microscope (see below). Slices were kept submerged by a nylon fiber grid and continuously perfused with aCSF at a flow rate of 5-10 ml/min.

Sulforhodamine 101 staining protocol
Sulforhodamine 101 (SR101) labeling was performed with a low concentration of the dye [6,19]. If not stated otherwise, slices were incubated for 20 minutes in carbogen-saturated aCSF containing 1 mM SR101 at 34uC. Thereafter, slices were incubated in aCSF for 10 minutes at 34uC to allow washout of excess dye from the extracellular space. Sodium-reduced solution was prepared by replacing NaCl by cholineCl.

Fluorescence imaging using multifocal 2-photon excitation microscopy
For detection of EGFP-and SR101-fluorescence, we used multifocal (8 or 16 foci) 2-photon excitation. The principle arrangement of the microscope (TriMScope, LaVision BioTec) was described earlier [20]. Two-photon excitation was achieved with a Ti:Sapphire Laser (SpectraPhysics MaiTai BB) at 800 nm. Images were acquired with CCD-cameras (Ixon 885 or Clara, Andor Technology) using either no or 262 binning. Fluorescence signals of hGFAP-EGFP expressing astrocytes were detected through a 531/40 nm band pass emission filter, whereas SR101fluorescence was detected through a BP 645/75 nm band pass emission filter (AHF Analysentechnik AG). Image noise minimization, was achieved either by offline averaging of four consecutive images or by integrating the emitted light of consecutive scans on the CCD camera. To allow quantitative comparison of the SR101-intensity between controls and drug treatments, all image parameters, especially camera settings (exposure time, gain and binning) as well as laser settings (laser power, beam number) were identical for a particular set of experiments. Cell counting was performed in a defined volume that was scanned with 2 mm step zstacks using a piezo-focus (Physik Instrumente). All settings were controlled by ''Imspector'' software (LaVision BioTec, Bielefeld, Germany). Since fluorescence intensity of EGFP in glycinergic neurons was low and sensitive to photo bleaching, we used the dual camera mode of our microscope when performing time-lapse recordings of SR101-loading in slices from Tg(GlyT2-EGFP)mice. In these experiments, emitted light was divided by a dichroic mirror (580 nm, beam splitter 580 DCXR, AHF), thus shorter wavelengths could be passed through a 531/40 nm band pass filter for detection of EGFP-fluorescence with a second CCDcamera (Clara, Andor Technology), while longer wavelengths (SR101-fluorescence) were detected with the other Clara CCDcamera as described above. Overlap of both cameras was adjusted before the experiment using standard fluorescent probes (Chro-ma). This technique allows adjusting the camera settings of the green and red channel independently and, thus, the camera setting for detection of SR101 could remain the same as in the single camera experiments.
For analysis of SR101-fluorescence, ''Imspector''-images were exported to TIFF-format. Deconvolution was performed with Autoquant software (MediaCybernetics) using the theoretical pointspread-function (adaptive PSF, 10 iterations). 3D-volumes were then further analyzed in Imaris (Bitplane). Drift correction was performed for time-lapse recordings. The fluorescence intensity of the SR101 and EGFP channel was determined using the spot objects feature of Imaris. A spherical 3D volume (spot) of 6 mm diameter was manually assigned to EGFP-expressing cells in the ''surpass'' view. In the SR101-channel the ''recenter'' function of Imaris was used to identify cells with SR101-labeling. Only if ''recentering'' was possible a cell was counted as a SR101-positive. To quantify the fluorescence intensity of an individual cell the median intensity of the assigned spot was calculated (in the timelapse experiments for each time point). We used a R script (The R Foundation for Statistical Computing) to extract the number of cells and their intensities out of the comma separated file that was exported from Imaris and to calculate the average SR101fluorescence intensities of the regarding slice.

Electrophysiology
Electrophysiological characterization of SR101-labeled or EGFP-labeled CA1 stratum radiatum cells was performed as described earlier [9,21]. After SR101-staining, using the protocol described above, hippocampal slices were transferred to a custombuilt recording chamber mounted to an upright microscope (Axio Examiner.Z1, Zeiss) and superfused with aCSF at room temperature. Epifluorescence illumination was achieved by a HBO100 mercury lamp (Zeiss). Two filter sets were used to discriminate between SR101-labeled cells that did not express EGFP and EGFP-expressing astrocytes. A dualband GFP/mCherry ET filter set (F56-019; AHF Analysentechnik) allowed the identification SR101-positive cells while the GFP filter set (38; AHF) was used to confirm the expression of EGFP. Fluorescence illumination was stopped to avoid bleaching, and a Dodt-Gradient-Contrast [22] was used to approach the identified cell with the patch-pipette. EGFP-and SR101-fluorescence was documented using a CCDcamera (Sensicam QE; PCO) that was controlled by CamWare software (PCO).
Whole-cell recordings were conducted with a Multiclamp 700A amplifier, Digidata 1440A interface and pClamp10 software (Molecular Devices, Forster City, CA, USA). For characterization of the current-voltage (I-V) relations, SR101-positive and EGFPpositive cells that were voltage-clamped at a holding potential of 280 mV, were exposed to 200 ms voltage-steps that reached from 2160 mV to +60 mV (10 mV increments). Current responses were low-pass filtered at 2 kHz and digitized at 10 kHz. The steady state current at 200 ms was measured to calculate the (I-V) curves. Additionally, the membrane resistance was calculated from the change of the holding current in response to a hyperpolarizing voltage step to 290 mV.

Data analysis
Data are expressed as mean 6 SEM. Statistical comparison was performed with the SigmaPlot software (Systat Software, Inc.). Statistical significance (t-test or Mann-Whitney U test) was expected if p,0.05.

Sulforhodamine 101 is not a selective marker for astrocytes in the ventrolateral medulla
To assess the specificity of SR101-staining in the ventrolateral medulla (VLM) in the brainstem, we adopted the Sulforhodamine 101 (SR101) staining protocol described earlier for acute hippocampal slices [6], to acute brainstem slices from neonatal Tg(hGFAP-EGFP) mice ( Figure 1). It was obvious that labeling of VLM-astrocytes with SR101 was not sufficient for the reliable identification of astrocytes ( Figure 1B). The intensity of SR101staining was so weak that we had to increase the gain of the images to detect SR101-staining in astrocytes at all ( Figure 1B9). In contrast, SR101-staining of hippocampal slices with the same protocol confirmed a good labeling of stratum radiatum astrocytes in the CA1 and CA3 region with SR101 ( Figure 1F).
We quantified the differences in SR101-labeling by measuring the intensity of the red-fluorescence using 2-photon microscopy. In the VLM we were able to identify a faint SR101-fluorescence, which was different from the background, in 59.9612.4% EGFPpositive astrocytes. The SR101-intensity was determined to be 22.3262.9 au (n = 5 slices; mean 6 SEM). In the hippocampus SR101-intensity of astrocytes (EGFP-positive) was five-fold higher (101.72621.29 au) (n = 6, p,0.05, Mann-Whitney U test). The   number of EGFP-positive cells that were stained with SR101 was 76.567.2% (n = 6, n.s., t-test).
In the hippocampus 46.363.1% of SR101-positive cells were lacking EGFP-fluorescence. These cells resembled astrocytes in size and shape and when we recorded from these cells we found no differences in the electrophysiological properties between EGFPnegative SR101-positive cells and EGFP-positive astrocytes ( Figure 2). The resting membrane potential of EGFP-positive astrocytes was 276.6363.74 mV (n = 8) and 274.7362.93 mV (n = 11) for the SR101-positive but EGFP-negative cells. No difference was detected in the membrane resistance (SR101positive but EGFP-negative: 80.1628.8 MV; EGFP-positive: 59.0611.9 MV). Additionally, it appears that SR101-labeled cells that did not express EGFP resembled also mature astrocytes indicated by the almost linear IV-curve ( Figure 2H).
We also performed SR101-labeling in brainstem and hippocampus slices from different developmental stages (Figure 3). While in the VLM SR101 did not allow a reliable identification of astrocytes at any time ( Figure 3A-C), solid SR101-staining of astrocytes was always possible in the stratum radiatum of the CA1 region ( Figure 3D-F).

2-photon time lapse imaging reveals transient labeling of non-astrocytic cells in the VLM
To examine the time course of SR101-labeling of astrocytes, we performed time-lapse 2-photon-imaging during the labeling procedure. After recording the fluorescence background in aCSF for 2 min, SR101 (1 mM) was bath-applied. Then changes of the fluorescence intensity were recorded every 2 min. In the VLM, EGFP-positive astrocytes did not take up much SR101 during the 20 min loading phase (Figure 4 C). However, we found SR101positive cells that were EGFP-negative (Figure 4 C,G). During washout, SR101-fluorescence disappeared rapidly from the latter (Figure 4 F,G).
In the hippocampus most EGFP-positive astrocytes but also EGFP-negative were loaded with SR101 during the 20 min loading phase (Figure 4 C9). Among the EGFP-negative cells most SR101-labeled cell resembled the EGFP-positive astrocytes in size and shape. During washout, SR101-fluorescence of EGFP-positive astrocytes decreased only slightly and allowed identification of astrocytes until the end of the recording (Figure 2 D9,E9). Additionally, a large number of the EGFP-negative cells retained SR101, yet, most of them resembling astrocytes in size and shape, confirming the labeling pattern of the initial staining experiments described above.
In an additional set of 2-photon time lapse imaging experiments, we stepwise increased the SR101-concentration in the bath solution ( Figure 5; n = 3). With a SR101 concentration of 0.1 mM, astrocytes in the CA1 stratum radiatum started to become labeled (Figure 5 B). During the 20 min staining procedure the SR101fluorescence intensity in the astrocytes was increasing above the fluorescence intensity of the SR101-containing aCSF that was measured above the slice ( Figure 5F). At 1 mM, the intracellular SR101-intensity increased further (Figure 5 C,F), while also the SR101 background fluorescence in the slice increased slightly. Both measured fluorescence intensities were larger as compared to the SR101-fluorescence in solution above the slice. Subsequent increase of the SR101-concentration to 10 mM caused an additional increase of the SR101-fluorescence of astrocytes during the 20 min staining procedure. The fluorescence intensity at 20 min did not reach a plateau, suggesting continuing accumulation of SR101 in the astrocytes ( Figure 5D). In all 3 tested conditions the SR101-intensity in astrocytes exceeded the fluorescence intensity of the background in the slice as well as the intensity in the aCSF (See also Movie S1). However the best signal to noise ratio was found in the 1 mM SR101-solution ( Figure 5G).
Since MFQ did not affect SR101-fluorescence intensity, we conclude that hemichannels formed by connexins or pannexins are not the major route for SR101-uptake and that the organic      . Effects of Estrone-3-Sulfate and Na + -free solution on SR101-uptake by astrocytes. A: Overlay of astroglial EGFP-fluorescence (green) and SR101-fluorescence (red) in the hippocampus after a staining in CTRL conditions. B: Overlay of EGFP-and SR101-fluorescence after application Estrone-3-Sulfate (E3S, 100 mM) in the staining solution shows significant reduction of SR101-fluorescence intensity. C: Overlay of EGFPand SR101-fluorescence in the hippocampus after staining in Na + -free solution shows no Na + -dependence of SR101-uptake. Scale bars: 40 mm. D: the statistical comparison confirms significant reduction of SR101-labeling by E3S in the hippocampus, but no effect of Na + -free solution. (E) Statistical comparison of the effects of E3S and Na + -free solution on the percentage of SR101-labeled astrocytes in the hippocampus. doi:10.1371/journal.pone.0049398.g010 Figure 11. Effects of OATP blockade on SR101-uptake in the hippocampus. A: Overlay of astroglial EGFP-fluorescence (green) and SR101fluorescence (red) in the hippocampus after a staining in control conditions. B: Rifampicin in the staining solution decreased SR101-fluorescence intensity. C: While the reduction of SR101-labeling by rifampicin (Rif; 100 mM) was significant, the fraction (D) of SR101-labeled astrocytes remained unchanged by rifampicin. E: astroglial EGFP-fluorescence (green) and SR101-fluorescence (red) in the hippocampus after a staining in control conditions. F: Reduced SR101-labeling of hippocampal astrocytes when the neurosteroid dehydroepiandrosterone sulfate (DHEAS, 100 mM) was present during SR101-staining. G: The statistical comparison reveals a significant reduction of SR101-fluorescence intensity by DHEAS G: Additionally, the percentage of EGFP-positive astrocytes that were also labeled with SR101 was significantly lowered as compared to control. Scale bars: 40 mm. doi:10.1371/journal.pone.0049398.g011 anion carbenoxolone might interfere with another uptake mechanism.

MK-571 reduces SR101-labeling in the hippocampus
Time-lapse imaging suggested that SR101 might be removed from neurons and VLM astrocytes via an active transport mechanism. Possible candidates are multidrug resistance proteins, which can mediate efflux of SR101 from cells [25] and also have been found in astrocytes [26]. Furthermore, blockade of effluxtransporters was shown to improve fluorescent dye-labeling of neurons [27].

Neurons can be labeled by SR101 in the ventrolateral medulla
In the VLM, as shown in Figure 4, SR101 enters EGFPnegative cell when SR101 is applied but leaves these cells once SR101 is removed from the extracellular medium. When MK-571 is applied in the VLM, SR101 was also detectable in EGFPnegative cells (Figure 7), which suggests that extrusion might be depending on a MK-571-sensitive process. Since some of the EGFP-negative SR101-positive cells had the shape and size of neurons, we aimed to test this by using another transgenic mouse line that expresses EGFP under the control of the GlyT2promotor [28]. In a first set of experiments we performed time lapse imaging of the labeling process with 1 mM SR101 in the VLM ( Figure 8A-D). Most of the SR101-fluorescence did not colocalize with cell bodies of glycinergic neurons. We could measure SR101-fluorescence after 20 min of SR101-application in only 2 glycinergic neurons (n = 3 slices). When we labeled VLM-slices (n = 5) from GlyT2-EGFP mice in the presence of MK-5171 (200 mM) we also found SR101 in a small number of glycinergic neurons (Figure 8 E-G). On average 7.362.7% of the SR101positive cells were glycinergic neurons.
Probenecid reduces astroglial SR101-uptake into hippocampal astrocytes Probenecid [29], another MRP-transporter blocker, also did not increase the SR101 labeling in VLM astrocytes ( Figure 9A-D). This indicates that the weak SR101-labeling of astrocytes in the ventrolateral medulla is not due to an active extrusion mechanism but rather result of a poor uptake. Like MK-571, probenecid reduced the SR101-labeling of hippocampal astrocytes. The SR101-intensity of EGFP-positive astrocytes was reduced by probenecid (1 mM) to 41.3615.6 au (n = 4) as compared to 131.8627.8 au in aCSF (n = 4, p,0.05, t-test, Figure 9G). In probenecid, the fraction of SR101-positive astrocytes was smaller (76.564.9% (n = 4)) as compared to aCSF (93.362.3%; n = 4, p,0.05, t-test, Figure 9H). Since it is very unlikely that MRPtransporter (e.g. ABCC1 transporter) mediate an uptake into cells [30], our observation might be explained by an interaction of probenecid and MK-571 with another type of uptake transporters.

SR101 labeling can be reduced by antagonists of organic anion transporting polypeptides
Probenecid is not only known to block MRP-transporters but is also a blocker of organic anion transporters (OATs) [31,32] and some organic anion transporting polypeptides (OATPs) [33]. Since SR101 is an organic anion [34,35], we hypothesized that OATs and OATPs are candidates for the SR101-uptake into hippocampal astrocytes. By simultaneous application of estrone-3-sulfate (E3S; 100 mM), a substrate of both families of organic anion transporters, we observed a strong reduction of SR101-intensity in hippocampus astrocytes (65.566.4 au (n = 5) as compared to 237.3637.7 au in aCSF (n = 6, p,0.05, Mann-Whitney U test,  Figure 10D). The fraction of SR101-positive astrocytes was also reduced (68.169.4% (n = 5) in E3S vs. and 87.362.4% in aCSF (n = 6, p,0.05, Mann-Whitney U test, Figure 10E). Since E3S is also a substrate of sodium-dependent anion transporters of the SLC10 family [36], we tested SR101-uptake in sodium free aCSF. However, no significant reduction of SR101-staining was observed ( Figure 10C-E).
Furthermore, rifampicin (100 mM; Figure 11), which is an inhibitor of OATPs but not of OATs [32] reduced the SR101intensity of hippocampal astrocytes to approximately 50% ( Figure 11C), while leaving the fraction of labeled astrocytes unchanged.

Discussion
SR101 is an excellent tool for the identification of astrocytes for imaging experiments in the hippocampus and cortex of rodents [6,7]. In this study, we show that SR101 is taken up into hippocampal astrocytes via an active transport mechanism. The uptake is likely to be mediated via a yet not described (orphan) organic anion transporting polypeptide (OATP; SLCO family), which can be blocked by probenecid and rifampicin.

SR101 is not a good marker for astrocytes in the ventrolateral medulla
We were unable to stain astrocytes efficiently in the ventrolateral medulla with SR101. SR101-labeling of astrocytes was also poor in the hypoglossal nucleus and in the spinal trigeminal nucleus (data not shown). In the ventrolateral medulla, SR101 can enter neurons ( Figure 8) and other cell types during the staining procedure but escapes from these cells when the slice is rinsed with normal aCSF (Figure 2). However, additional experiments are necessary to clarify the mechanism of SR101-loading in neurons. In this context it is noteworthy that, during hypoxia, hippocampal neurons can be labeled with SR101 via gap-junction hemichannels [12].

SR101 labels astrocytes in the hippocampal stratum radiatum efficiently and selectively
We found that about 46% of the SR101-positive cells did not express EGFP. However, SR101-positive EGFP-negative cells in the hippocampus resemble astrocytes in shape, size and process morphology, corroborating earlier observations by other groups [6,8]. When we performed whole-cell recordings from SR101positive EGFP-negative cells in the stratum radiatum of the hippocampus we only found cells that had electrophysiological properties of astrocytes ( Figure 2). Thus we conclude that SR101 is an efficient and selective marker for hippocampal astrocytes.
Blockers of organic anion uptake prevent SR101-labeling of hippocampal astrocytes In the hippocampus astrocytes, intracellular SR101-concentration reaches a much higher level compared to the VLM (Figure 4). Obviously, the SR101-fluorescence within the astrocytes increased above the fluorescence intensity of the SR101 containing aCSF, suggesting that SR101-fluorescence intensity is not depending on passive diffusion but rather on an active SR101-uptake into hippocampal astrocytes.
Since carbenoxolone [39] and probenecid are organic anions and probenecid is known to block organic anion transport [40], possible transporters for the SR101-uptake are organic anion transport systems including organic anion transporters (OAT) belonging to the SLC22 family or organic anion transporting polypeptides (OATP; SLCO family) and also transporters of the SLC10 family. Estron-3-sulfate (E3S) is the substrate of many members of these transporter families [32] and indeed was able to reduce the SR101-labeling significantly. Interestingly, organic anion transporting polypeptides have also been shown to be inhibited by MK-571 [41] and probenecid [40]. Finally rifampicin (rifampin) is known as an inhibitor of OATPs, i.e. OATP1A2, OATP1B1, OATP1B3, and OATP2B1 but not for OATs [32], strongly suggesting that SR101-uptake is mediated via an organic anion transporting polypeptide.
Although little is known about OATP expression in astrocytes, the pharmacological profile of the SR101-labeling in the hippocampus (figure 12) favors an orphan organic anion transporting polypeptide (OATP) as the transporter for selective loading of SR101 into astrocytes. Unfortunately, most of the pharmacology on OATP has been performed with human transporter clones, and thus species differences cannot be excluded for the mouse orthologues. Nevertheless, we can assume that hippocampal astrocytes have a higher expression of that uptake transporter as compared to VLM astrocytes.
Efflux transporter inhibition does not improve astrocyte labeling in the VLM It appears very unlikely that multidrug resistance proteins (MRPs) of the ABCC-subfamily or p-glycoprotein are actively lowering the SR101-concentration in the VLM astrocytes, since inhibition of those efflux transporters did not improve SR101labeling of VLM astrocytes. Neither MK-571 (MRP1, Mrp2) nor probenecid (MRP2-6) did increase the SR101-fluorescence in VLM astrocytes. However, we observed an increase of SR101staining in EGFP-negative cells, suggesting that MRPs contribute to the de-staining of cells in the VLM.

A potential role of the orphan SR101-transporter in the hippocampus
It is interesting to note that sulfated neurosteroids, such as dehydroepiandrosterone sulfate (DHEAS) [38] or pregnenolone sulfate [42], are known substrates of organic anion transporting polypeptides. Thus, we can speculate that in the hippocampus, astrocytic organic anion transporting polypeptides might be involved in transport and regulation of neurosteroids and thereby might modulate neuronal network activity [37,43] or glial plasticity [44,45]. If this is the case, external application of SR101 could compete with the astrocytic uptake of those neurosteroids and then may cause side effects such as described by Kang et al. [16]. SR101-induced elevation of DHEAS or pregnenolone-sulfate might increase neuronal excitability by blocking GABA A receptors [46,47,48] and activation of NMDAreceptors [43,49].

Conclusion
In conclusion, the selective labeling of astrocytes in the hippocampus with SR101 was confirmed, but this method is not applicable for the identification of astrocytes in the ventrolateral medulla. Additionally, our data strongly suggest that SR101 is taken up by hippocampal astrocytes via a transporter for organic anions, most likely an organic anion transporting polypeptide (OATP), which, however, is missing or differentially regulated in VLM astrocytes. An unequivocal identification of the candidate gene among the cloned OATPs requires new experiments.

Supporting Information
Movie S1 This movie is a two-photon z-stack recording (70 mm, 2 mm, 0.5 min 21 ) of a SR101-staining experiment with 3 different SR101-concentrations. After recording one z-stack without SR101, 0.1 mM SR101 was bath-applied, 20 minutes later the SR101-concentration was increased to 1 mM for 20 minutes. Finally, 10 mM SR101 was applied for 20 minutes before the recording was stopped. The white spots identify EGFPlabeled astrocytes that were selected for analysis of fluorescence intensities ( Figure 5). The blue spot marks the position where fluorescence background within the slice was recorded, while the purple spot was located above the slice to record fluorescence changes of the bath solution (Scale bar 40 mm). (AVI)