Real-time imaging of sodium glucose transporter (SGLT1) trafficking and activity in single cells - PubMed (original) (raw)

Real-time imaging of sodium glucose transporter (SGLT1) trafficking and activity in single cells

Chiara Ghezzi et al. Physiol Rep. 2017 Feb.

Abstract

The processes controlling targeting of glucose transporters to apical and basolateral membranes of polarized cells are complex and not-well understood. We have engineered SGLT1 and GLUT4 constructs linked to fluorescent proteins to highlight the differences in transporter expression and trafficking, in real time, in different cell types. Activity was assessed in parallel using a FRET glucose sensor. In COS cells and HEK cells, SGLT1 was distributed between the plasma membrane and intracellular compartments, but there was little expression in CHO cells. Trafficking was investigated using the lysosome inhibitors NH4Cl (10 mmol/L) and chloroquine (150 _μ_mol/L) and the proteasome inhibitors MG-262 (1 _μ_mol/L) and lactacystin (5 _μ_mol/L). Lysosome inhibitors caused SGLT1 accumulation into intracellular bodies, whereas proteasome inhibitors induced SGLT1 accumulation in the plasma membrane, even in CHO cells. Our data suggest that a fraction of SGLT1 is rapidly degraded by lysosomes and never reached the plasma membrane; another fraction reaches the membrane and is subsequently degraded by lysosomes following internalization. The latter process is regulated by the ubiquitin/proteasome pathway, acting at a late stage of the lysosomal pathway. Using the cholesterol inhibitor M_β_CD (3 mmol/L), a dominant negative dynamin (K44A) and caveolin, we showed that SGLT1 internalization is lipid raft-mediated, but caveolin-independent. In contrast, GLUT4 internalization is dynamin-dependent, but cholesterol-independent. The physiological relevance of these data is discussed in terms of differential membrane compartmentalization of the transporters and expression under stress conditions.

Keywords: Endocytosis; SGLT1; lysosome; proteasome.

© 2017 The Authors. Physiological Reports published by Wiley Periodicals, Inc. on behalf of The Physiological Society and the American Physiological Society.

PubMed Disclaimer

Figures

Figure 1

Figure 1

SGLT

1‐dependent glucose transport in

COS

cells. The differences in

FRET

ratios in Fig 1 illustrate how overexpression of wt

SGLT

1 and fluorescent construct

SS

‐

CFP

‐

SGLT

1 modulate glucose entry in

COS

cells expressing the

FRET

‐ based glucose sensor Flip 600 _μ_mol/L. In this panel, three traces are superimposed that were obtained in three sets of experiments with

COS

cells expressing either wt

SGLT

1,

SS

‐

CFP

‐

SGLT

1 or no

SGLT

1. At the beginning of the traces, glucose entry resulting from the addition of 10 mmol/L glucose to the bath was mediated by both endogenous

GLUT

s and overexpressed

SGLT

1. In all three cases, glucose efflux following bath glucose removal was mediated via endogenous

GLUT

s. In the middle of the traces, 10 _μ_mol/L CytoB was added to the bath to block endogenous

GLUT

s. Under these conditions, addition of 10 mmol/L glucose to the bath evoked glucose entries of similar amplitude with cells expressing wt

SGLT

1 and

SS

‐

CFP

‐

SGLT

1, indicating that

SS

‐

CFP

‐

SGLT

1 is fully functional. The fact that there is no glucose entry in cells that do not overexpress

SGLT

1, suggest that there is no detectable endogenous

SGLT

activity in

COS

cells.

Figure 2

Figure 2

SGLT

1 trafficking uses a classical pathway in

HEK

and

COS

cells. Depending on the experimental conditions, our data show that

SGLT

1‐

YFP

may be found in the trans Golgi network (

TGN

) (1), vesicles (2), the plasma membrane (3), or endosome (4). These images suggest that insertion and retrieval of

SGLT

1 in and out of the membrane follows a classical pathway. This diagram also specifies the modulator of endocytosis and degradation that were used and their potential sites of action. The letter “A” identifies the trafficking of proteins from the

TGN

to lysosome and “B” identifies the trafficking of endocytosed proteins from the plasma membrane to lysosome. The mode of action of the

GTP

ase dynamin is also identified in this diagram.

Figure 3

Figure 3

The cholesterol inhibitor M_β_

CD

enhances insertion of

SGLT

1 in the plasma membrane. In

HEK

cells (A) and

COS

cells (B), and under normal culture conditions (

DMEM

),

SGLT

1‐

YFP

is primarily located inside the cells in small vesicles and in larger structures identified as endosomes. In contrast, there is almost no

SGLT

1 expression in

CHO

cells. An image of a transfected

CHO

cells is shown in (C) for reference, but only 10–12 transfected cells were found in an entire 1.5 cm dish per experiment. Incubation overnight with the cholesterol inhibitor M_β_

CD

(3 mmol/L) facilitates insertion of

SGLT

1 in the membrane in

HEK

cells (D) and

COS

cells (E). In

CHO

cells, incubation with M_β_

CD

had no effect (F) and the number of transfected cells remained very low. In (G), the left panels show fluorescence intensity profiles for each condition. The graph in the right panel shows a quantification of

SGLT

1‐

YFP

insertion in the plasma membrane, using five different intensity profiles for each condition (see Experimental Procedures for details).

Figure 4

Figure 4

In

HEK

cells, the dynamin dominant negative K44A facilitates

GLUT

4 insertion in the plasma membrane, but has no effect on

SGLT

1 trafficking. Panels A (

SGLT

1‐

YFP

expressed in the absence of K44A) and B (co‐expression of

SGLT

1 with K44A) shows that

SGLT

1 remains associated with intracellular compartments when co‐expressed with K44A (B). In contrast,

GLUT

4‐

YFP

, which is also mostly localized to intracellular compartments under normal conditions (C) is directed to the plasma membrane when co‐expressed with K44A (D). Panel E shows the fluorescence intensity profiles for each condition. The graph in the right panel shows quantifications of

SGLT

1‐

YFP

and

GLUT

4‐

GFP

insertion in the plasma membrane.

Figure 5

Figure 5

Caveolin 1 (Cav1) does not regulate

SGLT

1 trafficking in

HEK

and

COS

cells. Lipid raft‐dependent endocytosis may be caveolin‐dependent or independent. Overexpression of Cav1 with

SGLT

1‐

YFP

in

HEK

cells causes accumulation of

SGLT

1 in the plasma membrane (B and E). However, as shown in (F), there is no co‐localization between

SGLT

1‐

YFP

and Cav1‐

CFP

. When expressed alone, Cav1‐

CFP

is targeted to “hot spots” in the plasma membrane, but is also localized to intracellular compartments (C). When co‐expressed with the dynamin‐dominant negative K44A, there is a strong accumulation of Cav1‐

CFP

in the plasma membrane (D). These data contrast with that in Panels A and B of Fig 4 where K44A had no effect on

SGLT

1 trafficking. Panel E shows the fluorescence intensity profiles for (A) and (B). The graph in the right panel shows quantifications of

SGLT

1‐

YFP

insertion in the plasma membrane for (A) and (B).

Figure 6

Figure 6

Effects of metabolic inhibitors on

SGLT

1 trafficking. Vesicle exocytosis and endocytosis are processes that require energy. We tested the effects of inhibitors of

ATP

production by glycolysis (

IAA

) and

TCA

cycle (Na

CN

) on

SGLT

1 trafficking. Incubation with 150 _μ_mol/L

IAA

caused

SGLT

1 insertion in the plasma membrane within 1–2 h in

HEK

cells (B) and

COS

cells (E). In contrast, incubation with Na

CN

(1.5 mmol/L) had no effect on

SGLT

1 trafficking in either

HEK

(C) and

COS

cells (F). The inhibitor of

PI

3 kinase, wortmannin causes accumulation of

SGLT

1 in the plasma membrane (G), suggesting that the effect of

ATP

is due in part to activation of

PI

3 kinase. Panel H shows the fluorescence intensity profiles for each of the condition illustrated in Panels (A) to (G). The lower graph shows quantifications of

SGLT

1‐

YFP

insertion in the plasma membrane for each condition from (A) to (G).

Figure 7

Figure 7

Proteosomal‐ and lysosomal‐dependent degradation of

SGLT

‐

YFP

in

CHO

,

HEK

, and

COS

cells. For the series of experiments depicted in this figure, we have used the lysosome inhibitors chloroquine (150 _μ_mol/L) and

NH

4Cl (10 mmol/L) and the proteasome inhibitors

MG

262 (1 _μ_mol/L) and lactacystine (5 _μ_mol/L). Panel A illustrates, as previously shown in Fig 3C, that there is almost no expression of

SGLT

1 in

CHO

cells under normal culture conditions (A). Addition of the lysosomal inhibitor chloroquine for 4–8 h causes

SGLT

1‐

YFP

accumulation in what appears to be the membrane of large vesicles identified as lysosomes (B). Incubation with

NH

4Cl for the same period of time had a similar, but perhaps lesser effect (not shown). Incubation with the proteasome inhibitor

MG

262 also causes expression of

SGLT

1‐

YFP

in

CHO

cells, but in this case, a large fraction of the protein was targeted to the plasma membrane (C). Bar graph data in panel (J) indicate that

SGLT

1 increased insertion of

SGLT

1 in the plasma membrane in response to

MG

262 was not associated with increased glucose uptake. The same experiments carried out in

HEK

and

COS

cells showed similar results. Incubation with the lysosome inhibitor, chloroquine, directs

SGLT

1‐

YFP

to large lysosomal vesicles within 4 h in

HEK

cells (E) and

COS

cells (H). Incubation with the proteasome inhibitor

MG

262 enhanced

SGLT

1‐

YFP

insertion in the plasma membrane after 4–6 h in

HEK

cells (F) and

COS

cells (I). However, in the cases of

HEK

and

COS

cells, most of

SGLT

1‐

YFP

was targeted to the plasma membrane (and the Golgi) and there was little fluorescence associated with intracellular compartments. Incubation with the proteasome inhibitor lactacystin had the same effects.

Figure 8

Figure 8

Does

SGLT

1 traffic via the plasma membrane prior to degradation by lysosomes? To test whether

SGLT

1 traffics through the plasma membrane prior to be degraded by lysosomes, we first incubated overnight

HEK

cells expressing

SGLT

1‐

YFP

with the cholesterol inhibitor (M_β_

CD

) (A) and added after that the lysosome inhibitor chloroquine to the incubation medium (B). After incubation with M_β_

CD

, a large fraction of

SGLT

1‐

YFP

is targeted to the plasma membrane (A), but after addition of the lysosome inhibitor, the newly synthesized

SGLT

1 is directed to lysosome, with no additional insertion in the plasma membrane (B). These data suggest that a large fraction of the newly synthesized

SGLT

1 is targeted to lysosomes. We did the same experiments using the proteasome inhibitor. In this case,

HEK

cells expressing

SGLT

1 were first incubated with

MG

262 for 4–6 h to target the transporter to the plasma membrane (C), then the lysosome inhibitor choloroquine was added (D). Data in (D) show that, following addition of the inhibitor,

SGLT

1‐

YFP

was not targeted to lysosomes, but instead

SGLT

1‐

YFP

insertion in the plasma membrane was sustained, while labeling of the Golgi increased. Western blot in (E) shows an increase of 34.5% (n = 2) of

SGLT

1‐

YFP

protein level after 8 h incubation in the presence of the proteasome inhibitor

MG

262 (1 _μ_mol/L). Actin (lower band) was used to normalize the protein levels obtained with and without incubation with

MG

Figure 9

Figure 9

Model for

SGLT

1 and

GLUT

4 trafficking in polarized epithelial cells. Based on our data as well as those of others and the model of Rodriguez‐Boulan (Rodriguez‐Boulan et al. 2005; Lakkaraju and Rodriguez‐Boulan 2007), we propose that

SGLT

1 traffics from the Golgi (1) to the basolateral membrane (3) before being internalized to endosome (4) and finally inserted into the apical membrane. This trafficking pathway has been denoted the transcytotic pathway, in opposition to the direct pathway that takes the proteins from the Golgi to the apical membrane. In this model, we assume that the basolateral to apical membrane trafficking is lipid‐raft mediated. It follows based on this simple assumption that

SGLT

1, which internalization is cholesterol‐dependent, would translocate to the apical membrane, while

GLUT

4, which internalization is not cholesterol‐mediated, would remain in the basolateral membrane.

Similar articles

Cited by

References

    1. Achiriloaie, M. , Barylko B., and Albanesi J. P.. 1999. Essential role of the dynamin pleckstrin homology domain in receptor‐mediated endocytosis. Mol. Cell. Biol. 19:1410–1415. - PMC - PubMed
    1. Al‐Hasani, H. , Hinck C. S., and Cushman S. W.. 1998. Endocytosis of the glucose transporter GLUT4 is mediated by the GTPase dynamin. J. Biol. Chem. 273:17504–17510. - PubMed
    1. Alwan, H. A. , van Zoelen E. J., and van Leeuwen J. E.. 2003. Ligand‐induced lysosomal epidermal growth factor receptor (EGFR) degradation is preceded by proteasome‐dependent EGFR deubiquitination. J. Biol. Chem. 278:35781–35790. - PubMed
    1. Banerjee, S. K. , McGaffin K. R., Pastor‐Soler N. M., and Ahmad F.. 2009. SGLT1 is a novel cardiac glucose transporter that is perturbed in disease states. Cardiovasc. Res. 84:111–118. - PMC - PubMed
    1. Calmettes, G. , Drummond G. B., and Vowler S. L.. 2012. Making do with what we have: use your bootstraps. J. Physiol. 590(Pt 15):3403–3406. - PMC - PubMed

MeSH terms

Substances

LinkOut - more resources