VIEWS: 219 PAGES: 37

									                                        EPITHELIAL STRUCTURE AND FUNCTION


           All external and internal surfaces of the body are lined by epithelial cells which serve to define
compartments that are controlled by the specialized functions of the epithelium. The main function of the epithelia
is to serve as a barrier separating the inside and outside fluids. These tissues are generally the sites at which the
exchange of water, minerals and organic molecules takes place with the environment. Therefore, epithelial cells
are specialized to perform a wide variety of vectorial functions. Transporting epithelia such as those found in the
renal tubule, absorptive epithelia of the intestine, and secretory epithelia in the liver and pancreas are examples of
epithelia that create and maintain concentration gradients between the compartments they separate. The vectorial
functions of the epithelia are a direct consequence of the polar organization of the epithelial cell. In this regard,
epithelial cells are highly organized with distinct sets of cell surface components in separate plasma membrane

Epithelial Organization

           Absorptive epithelia can be in sheets or tubular formations that are monolayered as the renal tubule,
stomach, small and large intestine and gall bladder or multilayered as the skin, esophagus and lower urinary tract.
Epithelial sheets or tubular formations generally serve two main important functions: Fluid and salt transport and
as a defense against invasion from pathogens and chemicals.

                                                                         Epithelial Cell Junctions
     micro vill i

                                                                                The ability of the epithelium to behave as a
  tigh t ju nction
zon ula o cclud ens
                                                                         barrier depends on the integrity of the junctions
be lt d esmoso me
zon ula a dhe rens
                                                                         between the individual epithelial cells. These junctions,
                                                          ju nction al
                                                          comp lex
                                                                         shown in Figure 1, can be grouped into three
spo t de smosome
macu la ad here ns                                                       functional categories: (1) limiting junctions, which
                                                                         form the barrier to ions and molecules through the
        kera ti n
      fila ments
                                                                         paracellular space; (2) adhering junctions which
                                                                         mechanically     hold    cells   together;    and     (3)
   ga p jun cti on                                                       communicating       junctions,    which   mediate    the
                                                                         passage of small molecules from one cell to another.

                                                                         Limiting Junctions
                      he mide smosome    ba sal la mina

Figure 1. Junctional Complexes                                                  The zonula occludens (ZO) or tight junction
                                                                         (TJ) is one of the major hallmarks of epithelia.       It
contributes to the transepithelial permeability barrier by controlling the passage of molecules through the
paracellular pathway. The TJ is a narrow belt-like structure on the plasma membrane that circumferentially wraps
the cell near the apical pole and forms contacts between strands of membrane proteins contained in the outer
leaflets of the two adjacent cell membranes (Figure 2). Because the TJ is a barrier to the lateral diffusion of lipids
and membrane proteins, it forms the boundary between the structurally and functionally distinct apical and
basolateral plasma membrane domains.

           Please remember that these membranes are often called by different names depending on the epithelia.
The apical membrane is also known as the brush border, the mucosal, or the lumenal membrane; and the

basolateral as the serosal or peritubular membrane. The apical membrane of epithelial cells is often comprised of
microvilli or cilia. The microvilli can increase the absorptive surface area of the epithelia over 20-fold. The
basolateral membrane of epithelial cells is in contact with the basal lamina, a continuous mat of extracellular
matrix proteins that separates the epithelial cells from the underlying connective tissue. In some types of epithelia
such as the kidney glomerulus and lung alveolus the basal lamina lies between two different cell sheets (capillary
endothelial cells).

                                                          Although the TJ appears to be the rate-limiting barrier for
             i nteracti ng                         transport through the paracellular pathway, it is a relative, rather
        pl asma membran es
                                                   than absolute barrier. The TJ ion permeability from various
                                                   epithelia can differ by several orders of magnitude. Epithelia with
                                                   TJs with a low permeability to ions are considered "tight"
                                                   whereas epithelia with a high permeability to ions are called
 i ntra cel l ul ar                                "leaky".     Generally these differences are a reflection of the
                                          0.6 µM
 space                                             differences in the junctional permeability.

                                                          Recent results suggest that the flow of ions across the TJ
 Occl udi n                                        is mediated by multiple, discrete transjunctional "pores" in the TJ.
Cl au di n(s)
                                                   These TJ channels exhibit specific size and charge selectivity
                                                   that help determine the character of the epithelium. In addition,
                                                   as will be discussed later, different signaling proteins and
                                                   transduction pathways regulate junction function and assembly.
                                     cytopl asmic
                                     ha lf of          Thus, the TJ is a dynamic structure with properties that can be
                                     l ip id b il ayer pharmacologically and physiologically modified. It appears that

                                                   subtle changes in junctional permeability can be induced by
Figure 2. Limiting Junction
                                                   modification of individual proteins within the junctional complex.
It is conceivable that regulatable channels or pores may exist within the tight junction itself to modulate
transjunctional fluid and electrolyte movement.

          Significant advances in understanding the structure and regulation of the vertebrate tight junction have
been recently provided by the identification and isolation of several junctional proteins.       Occludin, a 65-kDa
transmembrane phosphoprotein, was originally thought to be the main sealing component of the TJ (Figure 2,3).
This protein consists of approximately 500 amino acids with four transmembrane domains. The amino acid
composition of the extracellular loops suggested that the extracellular domains created a paracellular barrier
through hydrophobic contacts with occludin domains from adjacent cells to form the primary seal of the junction.
However, embryonic stem cells lacking occludin show normal polarity, TJ function and organization and mice
lacking occludin only exhibit subtle histological abnormalities. Thus, occludin appears to be more of a regulator
than an essential component of TJ. The variability in the transjunctional permeability properties of the different
tight junctional complexes appears to be related to the expression of a family of proteins called claudins (“to close”
Lat.). Claudins were originally identified biochemically through their cofractionation with occludin.      At present
roughly 24 claudin isoforms have been identified in the nucleotide databases. The predicted proteins range in size
from 211 to over 260 amino acids. The different claudins share sequence identities of 12 to 70% and appear to
group into subfamilies. They share a similar structure of four transmembrane domains with small extra and
intracellular loops. It now appears that while occludin may have an important role in determining TJ functions, the
claudins probably have the primary role in directing the permeability properties of the junction. Evidence for the
importance of the claudins in conferring the paracellular selectivity properties of the TJ comes from the recent
observation that a mutation in claudin-16 is responsible for a rare magnesium-wasting syndrome called renal

                                                                              hypo-magnesemia with hypercalciuria
                                                                              and     nephrocalcinosis.           Affected
                                                                              individuals lose magnesium in the
                                                                              urine leading to hypomagnesemia and
                                                                              seizures at an early age. Apparently
                                                                              claudin-16 creates         a magnesium-
                                                                              permissive channel within the tight
                                                                              junction of cells along the thick
                                                                              ascending     limb    of   Henle.     When
                                                                              functioning properly magnesium is
                                                                              reabsorbed from the tubular fluid into
                                                                              the blood.     With the mutation in
                                                                              claudin-16,    massive       amounts      of
                                                                              magnesium are lost in the urine.
                                                                              Mutations of other claudin isoforms
                                                                              have been linked to other human
                                                                              diseases. For example, mutations in
                                                                              claudin-5     cause        velocardiofacial
                                                                              syndrome, mutations in claudin-3 and
Figure 3. Schematic diagram of the protein interactions at the ZO. ZO-1 -4 are found in patients affected by
interacts with the transmembrane proteins occludin and claudin(s) as well as the Williams-Beuren     syndrome   and
cytoplasmic proteins ZO-2, ZO-3, actin, AF-6, the kinase ZAK, the transcriptional
                                                                                  claudin-14 mutations are found in
regulator ZO-1-associated nucleic acid binding protein (ZONAB), and cingulin.
                                                                              some forms of hereditary deafness.
These inherited diseases, the diverse tissue distributions of the claudins and observations using knockout mice,
strongly suggests that the claudins are the primary proteins responsible for the physiological and structural
properties of the paracellular pathway.

       In addition to occludin and claudin, other transmembrane proteins are found in the TJ. JAM (Junctional
Adhesion Molecule) is a glycosylated 43 kDa single membrane span protein that belongs to the immunoglobulin
superfamily. The protein consists of three distinct structural domains: an extracellular region of 215 amino acids
that contains two Ig-like domains which mediate homophylic adhesion, a transmembrane domain, and a short
cytoplasmic domain that contains a PDZ         (PSD-95, discs large, ZO-1) binding motif. The carboxy-terminal
cytoplasmic domain of JAM interacts with PDZ domains of several TJ cytoplasmic proteins including ZO-1, AF-6,
and PAR-3. In addition, JAM interacts with the amino terminal globular head of cingulin. Recently, two JAM
isoforms have been identified.     Although the exact function of JAM is unknown, the protein may mediate a
cascade of signaling events initiated by its homophylic adhesion to adjacent or migrating cells. In endothelial cells,
JAM may regulate leukocyte transmigration. The TJ of endothelial cells forms the barrier that prevents leukocyte
passage. However, during inflammation leukocytes readily exit the microvasculature; it appears that JAM plays a
role in regulating this process. Another protein with an Ig-like domain, CAR (Coxsackievirus and Adenovirus
Receptor) is also found in the TJ. CAR, like JAM, is a single membrane spanning protein with two extracellular Ig-
like domains that appear to function in homophylic adhesion. In transfected fibroblasts, CAR recruits ZO-1 to sites
of cell-cell contact. In native epithelial cells CAR coimmunoprecipitates with ZO-1 suggesting that the CAR
cytoplasmic PDZ binding domain associates with ZO-1.

       The cytoplasmic plaque of the TJ consists of several different types of proteins. This includes proteins that
connect the TJ transmembrane proteins to other cytoplasmic proteins and the cytoskeleton, along with several

types of regulatory and/or signaling proteins. ZO-1, ZO-2, ZO-3 and cingulin are peripheral membrane proteins
that are involved in the scaffolding of proteins at the TJ. ZO-1, ZO-2 and ZO-3 are members of the membrane-
associated guanylate kinase (MAGUK) protein family. These proteins are characterized by three PDZ domains, a
src homology (SH-3) domain and an enzymatically inactive guanylate kinase-like (GUK) domain. All three of these
domains have been identified as protein-binding regions. ZO-1 is a 210-255 kDa protein that associates with the
C-terminus of the claudins through its PDZ-1 domain. The ZO-1 PDZ-2 and -3 domains interact with JAM and the
GUK domain binds to occludin. ZO-2 and ZO-3 independently associate with ZO-1 through their PDZ-2 domains.
In addition, ZO-1 also binds actin, actin binding protein 4.1, AF-6 and cingulin (Figure 3). Several of these proteins
may function in tyrosine kinase pathways that regulate junctional properties. Moreover, ZO-1 can associate with
the adhering junctional protein catenin and to the gap junctional protein connexin 43. To add to this complexity
several splice variants of ZO-1 have been identified. These isoforms exhibit a tissue-specific and developmental
pattern of expression that may be important in the maintenance and regulation of the TJ.

       Interestingly, it has been suggested that ZO-1 can shuttle between the cytoplasm and nucleus depending
on the proliferative state of the cells. Although the role of ZO-1 in nuclear processes is poorly understood and
somewhat controversial, it is clear that ZO-1 can influence transcription through its association with the
transcription factor ZONAB (ZO-1-associated Nucleic Acid Binding). ZONAB is a Y-box transcription factor that
localizes to either the nucleus where it influences gene expression or to the TJ where it binds to ZO-1. When cells
are at a high density with established TJs, ZONAB is found associated with ZO-1 at the junctions. When cells are
at low density ZONAB accumulates in the nucleus. Thus ZO-1 and ZONAB influence gene expression in a cell-
density dependent fashion.

       ZO-2 and ZO-3, 160 kDa and 130 kDa proteins respectively, were originally identified through
coimmunoprecipitation with ZO-1. Like ZO-1, ZO-2 appears to bind to claudin, occludin, cingulin, catenin, actin
and actin binding protein 4.1, and ZO-3 associates with claudin, occludin and cingulin. Both ZO-2 and ZO-3 are
phosphorylated, although the exact role of their phosphorylation in TJ assembly and regulation is not well
characterized. Another cytoplasmic protein associated with the TJ is cingulin. Cingulin, a 140- to 160 kDa
phosphoprotein, consists of globular head and tail domains and a central -helical rod region. As mentioned,
cingulin interacts with ZO-1, ZO-2, ZO-3, JAM, AF-6, actin and myosin. Cingulin through is association with actin
and myosin provides another link from the TJ complex to the cytoskeleton. Several other proteins have been
shown to associate with the TJ. Some of these proteins may be involved in the architecture of the junction while
others may play a role in vesicle transport to the junction (rab13, rab3B, VAP), in signaling (G proteins, PKC,
PP2A), or in transcriptional regulation (ZONAB, symplekin). Additional studies will be required to fully define the
functions of the individual TJ-associated proteins.

       The assembly of tight junctions is a prerequisite for epithelial function and is therefore a carefully regulated
process. Tight junction formation is initiated by establishment of cell-cell contacts mediated by the calcium-
dependent adhesion molecule E-cadherin. In cultured epithelial cells that are not polarized, ZO-1 is diffusely
organized within the cytoplasm.     However, after cell-cell contact and cadherin-mediated cell adhesion, ZO-1
becomes localized to the plasma membrane at the sites of contact. Assembly of tight junctions may also be
regulated by protein phosphorylation. Inhibitors of protein kinases block the assembly of tight junctions as well as
their dissociation by calcium chelators. It also seems that diverse intracellular activation signals can alter TJs. For
example, transepithelial resistance of amphibian gallbladder epithelium substantially increases after treatment with
cAMP analogues. Since passive ion flow occurs predominantly across the TJ in these cells, these results suggest
that elevated cAMP levels can influence TJ permeability. Also, Ca           and activation of protein kinase C by phorbol
esters can affect TJ permeability. It appears that intracellular Ca        decreases permeability, whereas activation of
protein kinase C enhances TJ permeability. In addition, extracellular proteolytic enzymes and chelation of Ca

                                                                 and Mg      can disrupt TJ function.
actin filaments      LUMEN
inside microvillus                   micro vill i extend ing
                                     from api cal surface
                                                                        In addition to regulators of protein kinases, the uncoupler
                                                                 of     oxidative        phosphorylation,        carbonylcyanide      p-
                                                                 trifluoromethoyphenylhydrazone         (FCCP)     induces   a     sharp
                                                                 decrease in transepithelial resistance in MDCK cells (within 20
                                            tigh t ju nction
                                                                 seconds). When FCCP is withdrawn from the cells the resistance
                                                                 recovers in less than 2 hours. The changes in resistance are
       ad hesi on
          be lt
                                            bu ndl e of          associated with the diffusing of ZO-1 away from junctional
                                            actin fil amen ts
                                                                 complex, whereas recovery is associated with reconcentration to
                                                                 the peripheral membrane.        During the short period of FCCP
                                           la te ral pl asma
                                           memb rane s o f
                                           ad jace nt            treatment the cytoplasmic ATP concentration does not change,
                                           ep ithel ial ce lls
                                                                 however, intracellular pH decreases. These results suggest that
                                            ba sal surface       tight junction permeability is influenced by proton gradients and/or
Figure 4. Belt Desmosomes
                                                                 the intracellular pH.    Understanding the cellular pathways that
affect the assembly and functional properties of the TJ will require a more detailed understanding of the
constituents and molecular structure of the junctional complex. Undoubtedly, the characterization of the TJ
specific polypeptides will afford insights into the assembly, structure and regulation of the TJ.

            In summary three main functions can be ascribed to limiting junctions: (1) they contribute to holding the
epithelial cells together;        (2) they form a selectively permeable barrier between the fluid compartments
separated by the epithelium;         (3) they separate the apical membrane from the basolateral membrane
effectively limiting transport proteins to specific membrane domains. This segregation of transport proteins is an
essential property of epithelial cells, allowing transcellular transport to occur in the absence of a net transepithelial
driving force.

Adhering Junctions

            Adhering junctions (AJ) or desmosomes mechanically keep the epithelial cells together and anchored to
the basil lamina.       These adhesive elements stably connect cells together playing an essential role in tissue
organization and function. They are most abundant in tissues that are subject to mechanical stress, such as
cardiac muscle and skin. Although these junctions are relatively stable in providing mechanical integrity to the
epithelial sheet they should not be regarded as static structures. It is clear that they can be extremely dynamic
structures that respond with exquisite sensitivity to environmental agents to allow tissue remodeling during
development, differentiation, wound repair and invasion.                        In addition to being modulated in response to their
environment, cell junction molecules themselves play active roles in signal cascades initiated by extracellular
matrix ligands and growth factors.             Thus these junctions have important roles in integrating the changes in
morphology and gene expression during tissue and organ development and remodeling. The adhering junctions
consist of three different distinct structures: belt desmosomes, spot desmosomes, and hemidesmosomes.

            Belt desmosomes (also called zonula adherens, ZA) consist of bundles of actin located under the plasma
membrane that form a belt around the cell and connect the junctional complex to the actin network of the cell
(Figure 4). They are positioned at the apical end of the cell below the TJ. The adhesive interaction with adjacent
cells is mainly mediated by the cadherin family of adhesion molecules. Cadherins are transmembrane calcium-
dependent, homophilic adhesion receptors that play important roles in cell recognition and sorting during
development. Cadherins belong to a large superfamily of proteins that can be divided into four groups based on
their structure. These four subfamilies include the classical cadherins, desmosomal cadherins, protocadherins
and cadherin-like proteins (Table 1). As shown, there are many proteins that are members of the classic cadherin

                                            QuickTime™ and a
                                        TIFF (LZW) decompressor
                                      are neede d to see this picture.

                                              QuickTime™ and a
                                          TIFF (LZW) decompressor
                                       are needed to see this picture.
family, but the most important in belt desmosome formation is E-cadherin (also known as uvomorulin). E-cadherin
is a type-1 transmembrane glycoprotein (single transmembrane spanning domain, N-terminus extracellular, C-
terminus intracellular) that has a large extracellular domain that contains a homophilic binding surface that leads to
interdigitation of adhesive elements from the two adjacent cell surfaces (Figure 5, see Figure 9). In addition to E-
cadherin the plasma membrane of the junction also contains the proteins nectin-2 and vezatin. Nectins comprise
a sub-family of the immunoglobulin superfamily.        Unlike cadherins they promote cell adhesion in a calcium-
independent fashion.      Nectin-2 is composed of three extracellular immunoglobulin domains, a transmembrane
region and a small cytoplasmic domain. Like cadherin, nectin-2 forms dimers that interact with other dimers on
adjacent cells to form homophilic attachments between cells. Vezatin is a plasma membrane component of the
ZA that has a short extracellular domain, a membrane spanning domain and a large intracellular domain. The
intracellular domain binds to myosin VIIA and may interact with the cytoplasmic domain of E-cadherin.

        At the cytoplasmic surface E-cadherin interacts with cytoplasmic proteins called catenins. The cytoplasmic
tail of E-cadherin interacts with ß-catenin that in turn interacts with -catenin (Figure 6). -catenin binds to actin to
link E-cadherin to the actin cytoskeleton. Normally, ß-catenin is required for adhesion because it joins -catenin to
E-cadherin. However, experimentally ß-catenin is dispensable for rudimentary cell adhesion as long as -catenin
is fused directly (using recombinant DNA techniques) to the cadherin cytoplasmic tail. This finding, along with
                                                                            observations            that            tyrosine
                                                                            phosphorylation of ß-catenin correlates
                                                                            with diminished adhesion in response to
  5           5           5
  4           4           4
                                                                            growth factors and cell transformation
  3           3           3                                                 has suggested that ß-catenin acts as a
  2           2           2                                                 regulatory component of the complex. ß-
                                EC1                                         catenin    also     participates   in    signal
         2          2           2                                           transduction        and        developmental
         3          3           3                                           patterning, suggesting that it may also
         4          4           4                                           couple physical adhesion to signaling
         5          5           5
                                                                            events     during      morphogenesis         or
                                                                            development. These roles may also be
                                                                            influenced by the unique ability of ß-
Figure 5. A. Cadherin Zipper model for the structure of Cadherin-mediated catenin to behave as a transcriptional
Adherens Junctions. B. Plasma membrane proteins of the Adherens cofactor in canonical Wnt signaling

                                                 (Figure 7). Cytoplasmic ß-catenin can translocate into the nucleus and
                                                 transcriptionally activate Lef1/Tcf target genes.

                                                        In their capacity to maintain the overall state of adhesion
                                                 between epithelial cells, E-cadherin and the catenins may act as
                                                 important suppressors of epithelial tumor cell invasiveness and
                                                 metastasis.    For example, a diminution of E-cadherin expression or
                                                 function leads to enhanced cell invasiveness in cell culture, and E-
                                                 cadherin deficiencies or mutations correlate with the invasiveness and
                                                 metastasis of certain human tumors. This can be demonstrated in the
    ß-cateni n
                    -cate nin
                                                 fact that the E-cadherin gene is mutated in 50% of diffuse type gastric
                                                 carcinomas.      In addition, because the function of E-cadherin is
                                                 dependent upon the functional activity of the catenins, some tumors
                                                 have reduced or absent - and/or ß- catenin expression. These findings
Figure 6. Catenin complexes associated
with belt desmosomes.
                                                 are not that surprising given the fundamental role of these proteins in
                                                 the establishment and integrity of epithelial tissues.

                                                        In addition to the role of the belt desmosome in epithelial
                                                 structure, it is thought that the oriented contraction of the actin filaments
                                                 of the belt desmosomes causes the epithelial folds which lead to the
                                                 formation of villi in the intestine or the rolling of the epithelial sheet to
                                                 form a tube.        Thus, these bundles probably mediate one of the
                                                 fundamental processes in animal morphogenesis- the formation of an
                                                 epithelial tube from a sheet of epithelial cells.

                                                        The formation of the adhering and tight junctional complexes are
                                                 often coordinated events. In subconfluent cells TJ and AJ cytoplasmic
                                                 proteins are not associated with cell-cell contacts.        Cytoplasmic ß-
                                                 catenin is mostly targeted to the APC, glycogen synthase kinase 3ß
                                                 (GSK-3), axin complex for degradation. Upon activation of the Wnt
                                                 pathway or in tumor cells with an APC mutation, ß-catenin activates
                                                 nuclear TCF4-mediated transcription. Because cellular ZO-1 levels are
                                                 relatively low, ZONAB is free to interact with Cdk4 to repress
                                                 transcription of specific target genes. Upon the formation of the AJs, E-
                                                 cadherin mediated contacts result in the recruitment of several proteins
                                                 including ß-catenin and P13K to the AJ. Activation of P13K leads to the
                                                 activation of several protein kinases which results in the transcriptional
                                                 activation of several genes associated with cell differentiation. As the
                                                 TJ forms ZO-1 is targeted to the junction sequestering ZONAB away
                                                 from the nucleus. In a similar manner ß-catenin is sequestered at the
Figure 7. Core elements of the Wnt/ß-            AJ. This concerted regulation leads to the formation of the junctional
catenin pathway are shown, depicting how
activation of the Frizzled receptor by the       complexes and the transcriptional activation of genes required for
Wnt ligand leads to activation of the function   cellular differentiation. The events associated with the formation of the
of ß-catenin. This activates gene expression
leading to diverse cellular responses in both    AJ and TJ complexes are outlined in Figure 8.
embryonic development and in adults. Other
pathways, such as integrin-linked kinase
and p53, also regulate ß-catenin.

                                       de smogl ein s                                kera ti n fi lame nts   spo t de smosome
                                                                                                                                             Spot        desmosomes              are
                                                    cytopl asmic
                                                    pl aqu e made o f                                                               analogous to rivets and hold cells
                                                    de smopl akin s

                                                                                                                                    together at specific punctate locations
                                                                                                                                    (Figure 9).     They are prominent in
                                                                                                                                    tissues such as stratified squamous
                                                                                                                                    epithelia that are subject to great
                                                                                                                                    mechanical stress. These junctions
                                                                                                                                    consist   of    a     membrane         domain
                                                                                                                                    consisting of glycoproteins, and a

         kera ti n fi lame nts
                                                                                                                                    dense cytoplasmic plaque that is
         an chore d to                                     in te rcell ula r
         cytopl asmic pl aqu e                             spa ce
                                                QuickTime™ and a
                                                                                                                                    anchored        to     keratin       filaments
                                 TIFF (Uncompressed) decompre ssor
                                         in te racting to see this picture.
                                    are neede drane s
                                    pl asma memb
                                                                                                                                    (intermediate         filaments)       of    the
                                                                               ba sal la mina                     he mide smosome
                                          0.3 µm                                                                                    cytoskeleton. Besides their function
         Figure 9. Spot desmosomes and hemidesmosomes
                                                                                                                                    in cell coupling, desmosomes provide
                                                                                                                                    anchoring points for the intermediate
                                                                                                                          filament network contributing to the cell and
                                                                                                                          tissue organization. Because filaments join
                                                                                                                          spot desmosomes from adjacent cells, the
                                                                                                                          filament networks between adjacent cells are
                                                                                                                          connected indirectly through these junctions
                                                                                                                          to form a continuous network throughout the
                                                                                                                          entire epithelial sheet.         In the last several
                                                                                                                          years,     some     of    the     major      constitutive
Figure 8. Representative signaling pathways during the formation of
junctions. (A) In isolated cells, cytoplasmic proteins from AJs and TJs are                                               desmosomal          components            have        been
not associated with cell-cell contacts. ß-catenin (ß-cat) is mostly targeted                                              identified. The spot desmosome consists of
to the APC–GSK3-Axin complex. Upon activation of the Wnt pathway ß-
cat can activate TCF4-mediated transcription in the nucleus (green                                                        two types of transmembrane glycoproteins,
circle). Because [ZO-1] is very low, ZONAB is free to interact with Cdk4                                                  desmoglein (Dsg) and desmocollin (Dsc).
and to repress transcription of target genes. (B) Formation of AJs [yellow
boxes in (B) and (C)]. When E-cadherin–mediated contacts are engaged,                                                     Dsg and Dsc belong to the cadherin family of
several proteins, including ß-cat and PI3K, are recruited to AJs.
                                                                                                                          proteins (Figure 10). Like E-cadherin these
Activation of PI3K leads to activation of the Akt and p38 MAPK protein
kinases resulting in the transcription of genes associated with cell                                                      desmosomal cadherins are transmembrane
differentiation. Amounts of ZO-1 increase gradually, but ZONAB still
represses transcription. (C) Formation of TJs (red box). The polarization
                                                                                                                          proteins with conserved regions of homology
is stabilized by the segregation of apical and basolateral membranes.                                                     in the extracellular domain that are involved
The expression of proteins associated with polarization, such as the
brush border enzymes intestinal alkaline phosphatase (IAP) and sucrase-                                                   in calcium binding and adhesion.                      The
isomaltase (Suc), is increased. ZO-1 is targeted to the TJs and                                                           cytoplasmic region of the molecules is
sequesters ZONAB away from the nucleus, and ß-cat is similarly located
at AJs. Some PI3K is located in the apical membrane, probably within                                                      required     for    attachment       to    cytoplasmic
lipid rafts, but can be mobilized to participate in junction dissociation by                                              plaque      proteins.          Three      genotypically
external signals. Unexpectedly, PI3K can promote not only cell
polarization (ABC), but also TJ and AJ dissociation (CBA).                                                                different desmogleins (Dsg1, Dsg2, Dsg3)
                                                                                                                          and desmocollins (Dsc1, Dsc2, Dsc3) have
been identified. In addition, each Dsc isoform exhibits alternative splicing of the cytoplasmic domain resulting in a
longer “a” form and a shorter “b” form. These proteins are expressed in a tissue- and stratification-specific pattern
that may be important in determining distinct epithelial functions.

       The cytoplasmic plaque of desmosomes is complex and exhibits tissue-specific differences in both
structure and composition. The plaque contains the ß-catenin related proteins plakoglobin and plakophilin, and
desmoplakin (Figure 11). Desmoplakin is the most abundant. Other unidentified proteins may also be present in
the cytoplasmic plaque. Plakoglobin and plakophilin bind tightly to the cytoplasmic domains of both desmosomal

cadherins (desmocollin and desmogleins) possibly serving as the molecular link between the outer and inner
portions of the desmosomal plaque.         Plakoglobin is not restricted to spot desmosomes and may be associated
with other cadherins. In addition, plakoglobin, like ß-catenin may be involved in regulatory cascades that are
important in cell growth control, development and differentiation.
     As mentioned, the most abundant protein of the cytoplasmic plaque is desmoplakin. Desmoplakin is a
dumbbell shaped molecule with a central -helical coiled rod domain flanked by two globular end domains. Two
alternatively spliced isoforms exhibiting tissue-specific expression have been reported adding to the complexity. It
appears that desmoplakin directly interacts with intermediate filaments of the cytoskeleton and the desmosomal
plaque. These results suggest that desmoplakin functions as a link to anchor the intermediate filaments at the
desmosome. This interaction between desmoplakin and the filaments may be regulated by PKA phosphorylation
of desmoplakin at a serine residue.            Thus, phosphorylation of desmoplakin may prevent its non-specific
association with the filaments until its recruitment into desmosomes.

       Hemidesmosomes, as the name implies, are half-desmosomes that instead of connecting adjacent cells,
join the basal surface of the cell to the underlying basil lamina. While desmosomes and hemidesmosomes exhibit
similar structural characteristics, most of the proteins comprising these junctions, although related, are distinct.
While the transmembrane domains of desmosomes are composed of cadherins, hemidesmosomes utilize
integrins.    The integrins are a large family of adhesion receptors involved in many physiological functions.
Integrins mediate cell-cell as well as cell-extracellular matrix interactions. A distinctive feature of integrins is their
variable adhesive competence that is reversibly modified depending on the state of cell differentiation and/or
activation or in response to environmental signals. Integrins are heterodimeric proteins composed of an  and a ß
subunit. In epithelial cells it appears that integrin consisting of the 6 and ß4 isoforms is concentrated in the
hemidesmosomes. Other proteins found in hemidesmosomes include a 180 kDa transmembrane protein called
BP(180) (see Clinical Highlights, below) that appears to interact with the 6 integrin subunit and several plaque
proteins including desmoplakin. At the extracellular surface, it appears that the hemidesmosome interacts with
                                                                  several   cell    matrix   proteins.    Specifically,   the
         E-cadherin            Desmocollin-1a                     extracellular matrix protein laminin 5, along with other
                    Desmoglein-1              Desmocollin-1b
                                                                  proteins related to laminins appear to bind to the
        E1                      E1             E1                 hemidesmosome integrins. Thus, the hemidesmosome
        E2                      E2             E2                 is a complex and distinct structure that serves to attach
                                                                  the epithelial cell layer to the underlying basement
        E3                      E3             E3
                    E3                                            membrane         and   stroma.     Together    with     the
                                E4             E4                 desmosomes, the        hemidesmosomes distribute any
        EA                      EA             EA
                                                                  shearing forces through the epithelial sheet. This action
                                                                  helps in maintaining the epithelium as a single intact
                                                                  structure. A schematic diagram outlining the structure
             IA       IA           IA            IA
                                        DSI           DSI         of both spot and hemidesmosomes is shown in Figure
         ICS                     ICS
                     ICS                       ICS                12. A summary of the major proteins associated with
                     IPL                                          the adhering junctions is shown in Table 2.



Figure 10. Domains of E-cadherin, desmoglein and
desmocollin. E1-E4, extracellular repeating domains; EA,
extracellular anchoring domain; IA, intracellular anchoring
domain; ICS, intracellular cadherin-typical sequence; DSI,
desmocollin-specific insertion; IPL, intracellular proline-rich
linker; RUD, repeating elements; DTD, desmoglein-specific
terminal domain.

                                                                                300kD IFAP

                                                                 Dsc a

                                                                                                  300kD IFAP

                                                                                              Lam 5
      Figure 12. Schematic showing the major components of desmosomes and hemidesmosomes. Both the desmosome and
      hemidesmosome are connected to the intermediate filament (IF) cytoskeleton which shows interaction with the surface of the
      nucleus. Dsg, desmoglein; Dsc, desmocollin; Pg, plakoglobin; DP, desmoplakin; Lam 5, laminin 5.

Table 2. Adhering Junctions

  Structure             Intracellular             Intracellular               Transmembrane            Extracellular
                        Plaque                    Cytoskeletal                Link Protein             Ligand
                        Proteins                  Attachment

  Belt                  catenins                  actin filaments              E-cadherin              E-cadherin
  Desmosome                                                                    nectin-2                nectin-2 in
                                                                                                       adjacent cell

  Spot                  desmoplakin               Intermediate                 cadherins:              cadherins in
  Desmosome             plakoglobin               filaments                    desmoglein              adjacent cell
                        plakophilin                                            desmocollin

  Hemi-                 desmoplakin               Intermediate                 integrin                laminin & other
  desmosome             BP230,                    filaments                    BP180                   matrix proteins

Clinical Highlights:

        Epidermolysis bullosa (EB) is a group of rare genetic disorders characterized by noninflammatory blistering
lesions of the skin occurring after minor mechanical trauma. Generally tissue separation occurs at variable depths
in the skin and/or mucosa depending on the specific EB type. Characteristically, the epidermis detaches from the
basement membrane between the basal cells and the lamina lucida due to reduced numbers of
hemidesmosomes. Defects in genes coding for the structural proteins of the basement membrane zone have
been defined in some EB subtypes and abnormal expression of the structural proteins in others. Other blistering

disorders appear to be associated with autoimmune reactions. For these disorders the use of human
autoantibodies has helped to identify biologically important adhesion molecules of the skin. Some examples

Junctional EB- Apparently a result of a mutation in the ß4 integrin gene that leads to premature termination of
message transcription.

Generalized Atrophic Benign EB- Null mutation in the gene that codes for the 180 kDa hemidesmosomal
protein (BP180).

Bullous Pemphigoid- Autoantibodies circulating in some afflicted patients recognize BP180, or another
hemidesmosomal plaque protein with a molecular weight of 230 kDa termed BP230.

In addition several other inherited diseases have been linked to desmosomal mutations. Some of these are
outlined below: For more information visit: www.debRA.org

Communication Junctions

          Communicating junctions in vertebrate epithelia are exclusively gap junctions. Gap junctions are protein
channels that allow for direct communication between the interiors of cells in the epithelial sheet. The gap junction
consists of a bipartite protein structure composed of six subunits called a connexon. Two connexons join across
the intracellular gap to form a channel. Each cell in the junction contributes one connexon. The channel diameter
                                                         in mammalian cells is 20-30 Å which is large enough to
                                                         allow passage of the smaller cytoplasmic molecules (ions,
                                                         metabolites and chemical messengers up to ≈ 1 kDa)
                                                         while   limiting   macromolecules.      Consequently, gap
                                                         junctions keep epithelial cells coupled electrically and
      Ions                                       cAMP    metabolically. It is thought that the main function of the
                                                         gap junction is to provide coordination of the epithelial cell
                                                         layer as a whole, in particular during responses to agents
                                                         that induce changes in function.

Protein                                                          The connexon (or hemichannel) is made up of six
                                                         subunits called connexins or hemichannels.               The
                                                         connexin is a polypeptide with a molecular weight of
Figure 13. Gap Junction                                  approximately 30-50 kDa that crosses the lipid bilayer as
                                                         four  helices. Expression studies have shown that the
connexin polypeptides are sufficient for functional activity of gap junctions.      In humans the connexin family
consists of 21 different homologous members. The channels formed by the different connexins have different
permeabilities and properties towards metabolites and second messengers.            This selectivity may be greatly
increased by the association of different connexin isoforms into connexons. For example, as many as 14 different
connexons can form when two connexins intermix (Figure 14). In addition, some connexins are phosphorylated by
tyrosine kinases. This phosphorylation may result in conformational changes that influence the ion conductance
or permeability properties of the junction. Thus, the gap junction channel may be rapidly regulated in response to
various cellular events.

          While gap junctions allow cells in tissues to be in direct communication, they also have the potential to

                                                       make the tissue vulnerable to damage. If a cell dies or is
                                                       damaged it must be rapidly uncoupled from its neighboring
                                                       cells or the cytoplasm of the attached cells will be
                                                       compromised.      Therefore the channel has a closure
                                                       mechanism that is activated by an increase in intracellular
                                                          2+                            2+
                                                       Ca .     The increase in Ca , caused either by influx
                                                       through a damaged plasma membrane or the inability of
                                                       the cell to remove Ca , initiates the closing of the channel,
                                                       effectively segregating the cell from its neighbors. In
                                                       addition to responding to Ca , gap junctions can close in
                                                       response to cytoplasmic acidification, and in some cells to
                                                       a change in membrane potential. Presumably, Ca , or
                                                       phosphorylation may result in a connexin conformational
                                                       change that leads to occlusion of the channel pore (Figure
                                                               Since   several     connexins     are    phosphorylated,
                                                       protein kinases and phosphatases interact with members
                                                       of the connexin family. Several kinases that interact with

Figure 14. Connexin Oligomerization                    connexins include v- and c-src kinase, protein kinase C,
                                                       MAPK (mitogen-activated protein kinase), cdc2 kinase
(also known as cyclin-dependent kinase), casein kinase 1 proteins associated with the other junctional complexes
have been shown to associate with connexins. For example, ZO-1 binds to several members of the connexin
family, and ZO-2 has been shown to bind to the connexin, Cx43. Binding of these MAGUK family members to
connexins, may regulate gap junction assembly, stability and size. In addition to MAGUK family members, it also
appears that -catenin coimmunoprecipitates and co-localizes with Cx43. This association, whether direct or
indirect, suggests that -catenin may play a role in gap junctional regulation through the Wnt signalling pathway.
Gap junctional connexins may also be associated with microtubules and microfiliments.

                                                                       Several human diseases have been linked
                                                               to mutations in connexin family members. These
                                                               diseases range from the relatively common non-
                                                               syndromic         sensorineural         deafness      and
                                                               hyperproliferative skin disorders to the extremely
                                                               rare oculodentodigital dysplasia.

                                                               Lateral Intercellular Spaces

                                                                       Along their lateral surfaces, below the
                                                               limiting junctions, cells are usually separated by
                                                               convoluted intercellular spaces.         In epithelia that
                                                               transport salt and water, the width of these spaces
                                                               varies directly with the rate of transepithelial
Figure 15. Gating of connexons.                                transport. This suggests that transport occurs, not
                                                               only across the basal membrane but across the
lateral membrane as well. This view is supported by the demonstration that basolateral membrane transporters
are distributed throughout the basolateral membrane.

  Surface Amplification

            Epithelial surface area is often amplified several orders of magnitude by folds in the epithelial sheet itself
  and in the cell plasma membrane.           As mentioned above, epithelial folds generally are created by regional
  contractions of the actin filaments of the belt desmosomes. In many epithelial cells the apical membrane surface
  contains thin, finger-like projections (1-3 µm length, 0.1 µm diameter) called, microvilli. Microvilli contain about
  40 actin filaments that run in a parallel bundle along its length. At the tip the actin filaments are anchored to
  membrane proteins and at the base they extend into a perpendicular network composed primarily of actin
  filaments called the terminal web. The terminal web also contains myosin, and part of its function may be to keep
  the microvilli upright.

                                                                    The basolateral membrane surface area is also
                                                           increased by invaginations though it does not have microvilli.
                                                           The capillary bed surrounding transporting epithelia is
                                                           usually very rich and in close proximity to the epithelial layer.
                                                           The large flow rate of the capillaries allows rapid exchange
                                                           and/or removal of solutes, water and ions with the epithelia.

                                                           Classification of Epithelia

                                                                    Epithelia can be classified by several criteria. Often
                                                           epithelia are classified as "tight" (low junctional permeability)
                                                           or "leaky" (high junctional permeability) though this offers
                                                           little in understanding the physiological importance of the
                                                           tissue. Perhaps the most useful classification is one that
                                                           takes into account the primary transported substrates.
Figure 16. CX-43 binding proteins.
                                                           Examples of several are outlined below.

  Electrogenic, Ion-Transporting Epithelia

            In these epithelia an ion is the species transported by a process that is electrogenic. This means that
  transepithelial current or voltage is generated.      All of these tissues generate sizable transepithelial electrical
  potential differences when characterized with identical solutions on both sides. In general these tissues have
  lower overall ionic permeability and water permeability than salt-transporting epithelia. The ion transport proteins
  that mediate the transport of these ions are outlined at the end of this handout. Based on the main transported
  species they can be characterized as follows:

  1. Na -Transporting Epithelia

            Na is the sole or main species transported. These epithelia usually have a low overall ionic permeability
  that creates a high electrical resistance across the epithelia. The transport of Na across the epithelia generates a
  sizable transepithelial potential difference with the side from which Na is transported becoming negative.

            Examples of Na -transporting epithelia include the distal segments (distal tubule and cortical collecting
  tubule) of the renal tubule, colon, amphibian skin, and amphibian and mammalian urinary bladder.

  2. Cl- Transporting Epithelia

            Cl- is the sole or main species transported. The electrical resistance is usually lower than Na -transporting

epithelia. As expected the transepithelial potential difference is positive from the side which Cl- is transported.

      Examples include: Regions involved in Cl- absorption such as the thick segments of the loop of Henle in
the mammalian kidney and the diluting segment of amphibian renal tubule and tissues involve in Cl - secretion
such as the trachea, corneal epithelium and the rectal gland of some fishes.

3. H -Transporting Epithelia

        As the name implies the predominant function of this epithelia is to secrete H . Generally this type of
epithelia has a low ionic permeability. Transport of other ions is observed, and depending on the mechanism of H
transport can be directly coupled (H,K-ATPase) or independent (H-ATPase) from H secretion. A transepithelial
potential difference is generated such that the side from which the H is transported is negative.

        Examples include: gastric epithelium, medullary renal collecting tubule, and reptilian urinary bladder.

4. K -Transporting Epithelia

        Transport of K is the predominant function.        Large gradients are often established and maintained
indicating a low ionic permeability. The side from which K is transported is negative.

        Examples include: stria vascularis epithelium of the inner ear that transports K into the endolymph and the
insect midgut that secretes K into the midgut lumen.

Non-Electrogenic, Salt-Transporting Epithelia

        These epithelia transport NaCl and sometimes also NaHCO3, by an electrically neutral process. As would
be expected the transepithelial electrical resistance is low indicating a high overall ionic permeability. Water
permeability is also high so salt transport is accompanied by water transport in near isosmotic proportions.

        This list is by no means complete. The classification of the epithelium was based on its main transport
function; however, realize that in most cases more than one species is subject to net transport by a particular

                 Common Membrane Properties of Epithelia

                         Although epithelia transport a wide spectrum of substances there are a few basic
  Frog skin      properties that they all tend to share. These include: (1) generally the Na,K pump (Na,K-ATPase)
                                                                                +                     +               +
is located exclusively on the basolateral membrane. Here it expels three Na in exchange for 2 K ions. (2) K is
accumulated intracellularly by the Na,K pump, and the basolateral membrane is predominantly K permeable;
                                                                +                        +
therefore the membrane potential is typically close to the K diffusion potential. (3) Na activity is much lower in the
cell than in the extracellular fluid (Na,K pump). In addition to the approximate 10 fold concentration ratio, the cell
                                                                                    +                         +
negative membrane potential provides an additional driving force for Na entry.               Therefore, Na , using its
electrochemical gradient, can drive the accumulation of an uncharged solute, producing up to a 100 fold
concentration ratio.

        The properties of the apical membrane of epithelial cells are very diverse and depend on the specific
transport functions of the tissue, consequently no generalizations are warranted.

Models of Epithelial Transport

  A                                                                   B

                        +                                             Na+
                      Na                                                                 glucose

                                  +                        +
  Figure 1. A) Transepithelial Na transport. Entry of Na is passive, through a channel, exit is through a pump. The
  electrochemical potential difference favors Na entry. B) Transepithelial glucose transport. Entry is secondarily active,
  coupled to Na (cotransport). Exit is through a facilitated diffusion transporter.

       Modern work on epithelial ion transport began in the mid 1930's with observations on NaCl absorption
across frog skin. At this time it was shown that ions were transferred across the skin against their electrochemical
gradients and that this transport required metabolic energy. This work has served as a foundation of most current
epithelial transport work; the advances since then, including short circuiting or voltage clamping, flux ratio criterion
for active transporting, and the use of radioactive isotopes for the quantitation of unidirectional fluxes, have relied
heavily on the frog skin model.

Basic Models

       Transport across electrogenic, ion-transporting epithelia can be reduced to one of two simple models.
Transport is transcellular and therefore must include two translocations: entry into the cell and exit from the cell.
When the destination of the transported species is the extracellular fluid the process is called absorption and
when the destination is the lumen the process is secretion. In this simple model ion transport is an uphill process
at one side of the cell and a passive process at the other. The active step can be primary, that is mediated by a
pump, or secondary, utilizing the thermodynamic gradient of one ion to power the uphill transport of another. For
                                                               +                                      +
example, Figure 1A shows a model for transepithelial Na transport. In this model entry of Na at one side of the
cell is passive, as through a channel, with exit at the other side by a pump. Alternatively, the active transport step
can be at the apical membrane. This model of transepithelial transport is shown in Figure 1B. In this example
                                                                                               +                             +
describing transepithelial glucose transport, entry of glucose is active, coupled to Na and driven by the Na
gradient (Na /glucose cotransporter).      The glucose chemical potential is high in the cell, driving passive exit
through facilitated diffusion ((glucose transporter, GLUT)).          For this system to operate, intracellular Na
concentrations must be kept low. This is accomplished by the Na,K pump which maintains the low intracellular
   +                                                                                                               +
Na concentration. Transepithelial transport of this type can occur for organic substrates, phosphate and K ; entry
into the cell is active and exit is passive with the ions being above their electrochemical equilibrium in the cell.

Sodium-Transporting Epithelia. Ussing's Two-Membrane Hypothesis
       For several reasons frog skin is probably the best studied Na -transporting epithelia. This is because it is
readily available, it permits easy manipulation of the media bathing both surfaces of the tissue, and the difference
in electrical potential can be changed reasonably uniformly through the entire epithelial sheet. Therefore it is easy
to understand why frog skin has been used for over 135 years to study salt and water transport.

     Figure 2. The Ussing chamber.
          As a frog sits in pond water ([Na ] < 1 mM) the normal tendency is for it to lose salt through its skin.
However, the skin of the frog is capable of transporting NaCl from the low concentration in pond water to the high
concentration in the body. The salient features of frog skin can be demonstrated, in vitro, in a special chamber
called a Ussing chamber (named after Hans Ussing). This chamber, illustrated in Figure 2 allows separation of
the solutions bathing the external and internal surfaces of the epithelium. Using the Ussing chamber the following
features can be demonstrated:
1.        When the solutions on both sides of the frog skin are identical in composition (and similar to the frog's
          extracellular fluid), the skin develops an electrical potential difference (outside negative) of 50 to 100 mV.

2.        If an external current is applied across the tissue the difference in voltage between the two aqueous
          phases can be reduced to zero. Under these conditions the preparation is termed to be short-circuited and
          the current necessary to voltage clamp the preparation to 0 mV is called the short circuit current (Isc). In
                                                                               +               +
          the short-circuited state we would expect no net movement of Na since the Na electrochemical gradient
          is zero (remember the Na concentration on both sides of the membrane is equal and V = 0). However,
                                                  +                                                            +
          measuring unidirectional fluxes of Na with radioactive tracers indicates net movement of Na from the
          outside surface to the inside. Furthermore, under a wide range of experimental conditions it can be shown
                          +                                                                                                +
          that the net Na flux from outside to inside is equivalent to the Isc. These observations indicate that Na
          transport is active and can be simply measured by recording the Isc.

3.        Isc is abolished if Na is omitted from the outside solution.
4.        Isc and net Na flux drop if O2 is omitted, metabolic inhibitors are used, cardiac steroids (Na,K-ATPase
          inhibitors) are added to the inside solution, or K is omitted from the inside solution.
5.        The normal open-circuit transepithelial potential is a steep function of [Na]out (Na concentration in the
          external solution) and [K]in (K concentration in the internal bathing solution).

     These observations lead to a useful working definition of active transport; it is a process
dependent on metabolic energy that can produce net movement of a solute across a barrier even in the
absence of external driving forces.

        From these and other experimental observations, Ussing and his colleagues proposed the model illustrated
in Figure 3 describing transepithelial Na transport in frog skin. In this model the outer membrane is a passive
   +                                      +
Na -selective barrier, across which Na enters the cell by electrodiffusion down its electrochemical potential
                                            +                                                          +
gradient. Depending on the epithelium Na entry at the apical membrane can be mediated by Na channels,
                                    +   +                                            +
Na/substrate cotransporters or Na /H exchangers. At the inner membrane, Na is extruded actively by the
                                                               +           +
operation of the Na,K pump. The Na,K pump transports 3 Na out and 2 K in for each ATP hydrolyzed generating
an electrical potential. K leaks from the cell back to the inner medium along its electrochemical gradient, across
      +                                                          +
the K -selective inner barrier. The apical membrane is K impermeable. In the physiological open-circuit
                                                                     condition,   Cl- flows passively from the
                                                                                   outside solution to the inside solution. The
                     outer                               inner                     movement of Cl- is driven by the electrical
                    barrier                             barrier                    potential difference and probably occurs
                                                                                   across the junctions.      Therefore, under
                                                                                   normal conditions, the net process is NaCl
                                               Na                           Na     transport. In the short-circuited condition,
           Na                                                                              +
                                                                                   only Na is transported across the epithelium.
                                                K                           K
                                                                                             Ussing's model can be generalized to
                                                                                   a        number     of    epithelia        that   exhibit
                                                                                   predominant or exclusive transport of a
                 Na entry is by                 Na extrusion is by a single ion. These epithelia have been listed
                   diffusion                   Na,K exchange pump above. It is important to remember that the
                          Na out                                Kcell              passive step in ion-transporting epithelia is
                 Vo = f                             Vi = f                         conductive and mediated by a channel or
                          Na cell                               Kout
                                                                                   carrier and the active step can be directly
           Figure 3
                                                                                   coupled to a metabolic energy source (pump)
or can be secondarily linked through utilization of a pre-existing gradient (co or countertransport).

Paracellular Transport

        The presence of a high conductance intercellular pathway has been demonstrated in several NaCl-
transporting epithelia. An elegant demonstration of the high conductance of the paracellular pathway is shown in

                          A                   ²V
                                                                            X    XX XX X X             X X X       X     X
                                                                            J    J J CC J C            J    C C    J     C
             -                                                         0
                                                             ²V (mV)

                          A                                            -1

                                                                            0          50        100         150             200

                                                                                   Position of electrode (µm)

Figure 4. Example of voltage scanning Necturus gall bladder epithelium. To detect high conductance pathways for
transepithelial current flow, a microelectrode was moved along the lumenal surface of the gall bladder while current was
passed transepithelially between electrodes marked A in the inset. In the sketch above the microelectrode path is indicated
by a heavy line, with respect to the cell borders indicated by a dashed line, and the approximate position of the microelectr ode
tip during each voltage measurement is marked by an X. The letters indicate whether each position is over a junction (J) or
over a cell (C). Note that current is funneled through the junctions.

Figure 4. As can be seen, the high conductance pathways correspond to the junctional regions. The functional
effects of the paracellular pathway can be summarized as follows:

1.      The paracellular pathway is a "leak" pathway. If net solute transport results in a concentration gradient
(basolateral > lumenal) some of the solute will leak back by diffusion (or electrodiffusion if appropriate). Therefore,
the steady-state transepithelial concentration gradient will be small as compared to epithelia with a low paracellular
permeability. However, a more complicated situation arises when the paracellular pathway is selectively
permeable to only one ion. For instance, if the Cl- permeability is high, and if only Na is actively transported, Cl-

can follow passively through the paracellular pathway. This can prevent the generation of a large transepithelial
electrical potential that would normally oppose the transport of more Na . The trade-off in this mechanism of
                 +                           +
augmenting Na transport is that some Na will invariably diffuse back to the lumenal solution through the leaky
limiting junction.

2.      If the paracellular pathway is significantly water permeable, the rate of salt transport can be increased
considerably by the mechanism known as solvent drag. Ions and solutes transported from the cell into the lateral
intercellular spaces will tend to increase the local osmolarity. This increase in osmolarity will cause water flow
from the lumenal (apical) medium. If the solutes present in the lumenal solution are permeant across the limiting
junction they will be dragged by the water flow according to the following relationship:

       Ji = Jv Ci (1-i)

        Where Ji is the flux of the substance i, Jv is the volume flow, Ci is the concentration of i, and i is its
reflection coefficient (i is equal to 1 if i is impermeant, and equal to 0 if freely permeant).

3.      The electrical consequence of the presence of a shunt pathway is to couple electrically the two cell
membranes, and to change the electrochemical potential gradient responsible for ion transport across the cellular

Water Transport Across Epithelia

                                                            "Leaky" and "tight" epithelia differ radically in their water
                                                    permeability.    Tight epithelia under control conditions are
                                                    generally water impermeable. Under the action of antidiuretic
                                                    hormone (ADH), some tight epithelia increase their water
                                                    permeability (cortical and medullary collecting tubules, and
                                                    amphibian skin and urinary bladder). The site of the increase in
                                                    water permeability is the apical membrane, which in the
                                                    absence of ADH, has a low water permeability. The basolateral
                                                    membrane has a high water permeability that is independent of
                                                    ADH. The water permeability of the plasma membrane is
                                                    dependent on water channels called, aquaporins (AQP).
                                                    Aquaporins are six-span membrane proteins that mediate
                                                    water transport across the membrane. At least 10 isoforms
                                                    have been identified. AQP3 is constituitively expressed in the
                                                    basolateral membrane. In contrast, AQP2 resides in a sub-
                                                    apical membrane vesicles. The binding of ADH to its receptor
                                                    results in a signaling cascade that drives the exocytotic fusion
Figure 5. The mechanism of ADH-stimulated water
                                                    of the AQP2 containing vesicles with the apical membrane

                                                                     (Figure 5). Net water transport across ADH-
                                                                     sensitive tight epithelia is driven by osmotic
   salt                                                              gradients generated by salt transport.     Other
                                                 sweeping away       tight epithelia (the thick ascending segment of
                                                 diffusion           the loop of Henle and the mammalian lower
          H2O                                                        urinary tract) tend to be unresponsive to ADH.

                                                                            In leaky epithelia such as the renal
                                                                     proximal tubule and the small intestine, water is
                                                                     transported in isosmotic proportions with solute
                                                                     transport even though lumenal and extracellular
                                                                     fluid omolarities do not differ experimentally.
                                                                     Under certain experimental conditions water
                   Length                                            moves apparently in an uphill fashion with no
                                                                     detectable osmotic gradient across the epithelia.
  Figure 6. The standing osmotic gradient hypothesis.
                                                                     The apparent absorption of water without the
requisite driving forces had been a puzzle to physiologists for several years. However, experimental results using
amphibian gallbladder suggest that the water permeability of the cell membranes is so large that a trivial osmotic
gradient might suffice to explain the normal rates of water transport.          The water permeability of the apical
membrane has been estimated from changes in cell volume and from rapid changes in the intracellular activity of
an impermeant ion used as a volume marker. These results suggest that osmolarity differences of 1 mOsm or
less across the cell membranes (normal osmolarity of the extracellular fluid is ≈ 300 mOsm) could account for the
normal rates of fluid transport. Therefore, the necessary gradient may not be detectably different from isosmotic.
Hence, it appears that water absorption in the amphibian gallbladder, and probably in renal proximal tubule and
small intestine as well, is transcellular and driven by very small differences in osmolarity.

         In some epithelia water may flow down an osmotic gradient generated across a restricted intraepithelial
compartment. This compartment would be rendered hyperosmotic by salt transport, water would be dragged into
it building up hydrostatic pressure, resulting in bulk flow to the serosal space. The appropriate location for such a
compartment is just beneath the basolateral cell membrane. In addition, to account for the maintenance of
permanent hyperosmolarity, the compartment has to be effectively unstirred. This model is illustrated in Figure 6.
As shown, the pumps are presumed to be located at the apical ends of the lateral intercellular spaces. Salt
transport into these spaces renders the solution hyperosmotic which draws water from the cell. The water flows
towards the serosal end of the channel, until by the time the fluid emerges from the intercellular space, it has
approximately reached full osmotic equilibration.

         The model relies on the establishment of local osmotic gradients through active solute transport. The
specific epithelial ultrastructure involving confined spaces and unstirred layers help to explain the movement of
water driven by small osmotic gradients. However, recently it has been shown that water transport apparently
occurs through ion cotransporters. For example, in addition to transporting Na ions and glucose, the Na ,glucose
cotransporter mediates the movement of water molecules. This movement of water may be another mechanism
to explain the apparent isosmotic transport of water across epithelia.

Basic Models of Net Absorbing and Net Secreting Fluid Transport Epithelium

         By extending Ussing’s basic model the mechanism for absorbing and secreting epithelia can be described.
This requires an assessment of the specific transport proteins present in the apical and basolateral plasma

Figure 7

  membrane domains. This is typically accomplished through careful transport studies using relatively specific
  inhibitors for the transport proteins in question. Figure 7A shows          a       generalized mechanism for absorptive
  epithelium (such as found in the gut or nephron).     The basolateral Na,K-ATPase maintains a relatively large
  negative basolateral membrane potential and a low intracellular Na concentration (5-10 mM).                This generates a
  large electrochemical gradient for Na entry at the apical membrane. Depending on the specific epithelial cell, Na
                                   +              +                               +    +
  entry can be through apical Na channels, Na cotransporters or Na /H exchangers. Electrical neutrality is
  maintained through the movement of Cl- through channels (transcellular) or the tight junctions (paracellular).
  Similarly, water movement is through aquaporins (transcellular) or the tight junctions (paracellular).      Taken
  together this model explains the net transepithelial movement of a roughly isosmotic fluid containing Na and Cl-.

         An example of a secretory epithelium is shown in Figure 7B. In this case basolateral Na/K/2Cl
  cotransporters (NKCC1) increase intracellular [Cl-] which exits the cell down its electrochemical gradient through
  apical Cl- channels. At the basolateral surface the Na gradient supplied by the Na,K-ATPase drives the uptake of

  HCO3 through the Na /HCO3 cotransporter (NBCn1). HCO3 exit at the apical membrane is through a member of
  the anion exchanger family.    Na movement into the luminal space is down its electrochemical gradient and
  generally through paracellular pathways. Water movement is mediated by apical and basolateral AQPs or through
  the TJ.

                                                                Regulation of Epithelial Transport

                                                                          The essential function of transporting epithelial
                                                                is to regulate the volume and composition of body
                                                                fluids.       This is accomplished by adjusting the
                                                                transepithelial transport rates according to the
                                                                requirements of the organism as a whole.                     For
                                                                instance, if a subject or experimental animal were to
                                                                undergo a net loss of sodium, the compensation
                                                                would be to lose less sodium in the urine, which
                                                                means at least in part, to reabsorb more sodium in
                                                                the renal tubule. Similar considerations can be made
                                                                for a number of substances when their total amount
    Figure 8
                                                                in   the      body         undergoes   primary   increases    or
  decreases. Most of the adaptive responses of transporting epithelia under these conditions are mediated by
  hormonal first messengers that bind to receptors in the epithelial cell. Activation of enzymes linked to these
  receptors causes changes in levels of intracellular second messengers. The second messengers in turn affect

the number and/or properties of the transporters which determines the final stimulation or inhibition of transport.

       Changes in transepithelial transport rate tend to cause changes in epithelial cell volume and composition,
which can perturb its function. Therefore, transporting epithelia are faced with two regulatory problems: (1) to
change the rate at which transepithelial transport takes place; and (2) to maintain their volume and composition
nearly constant. To achieve both objectives, the transport rates of both membranes often must change in a
synchronized fashion.

       The rate of transport of ions, nonelectrolytes or water across an epithelial cell can be altered by changes in
the rate of entry, the rate of exit, or both. For the simple case in which only one of these rates is primarily altered,
the site involved can be identified by measuring both the transepithelial transport rate and the intracellular content
(or if appropriate, the concentration) of the substrate. The possible combinations are shown in Figure 8.

       There are two general mechanisms by which changes in transport rate such as those shown in Figure 9
can occur: (1) change in the number of transport sites such as the incorporation or removal of transporters from
the membrane; (2) change in the kinetic properties of the transporters. Some examples of both mechanisms are
presented below.

Regulation of Transepithelial Transport by Changes in the Number of Transport Sites

Apical Membrane
(1)    Electrogenic, Na -transporting epithelia are characteristically responsive to the steroid hormone
aldosterone, which is secreted by the adrenal cortex in response to reductions in total amount of sodium, and
hence volume, of the extracellular fluid. The effect of aldosterone is to increase the number of Na channels in the
                                                    +                                                                     +
apical membrane. This results in enhanced Na entry which stimulates the Na,K pump to transport more Na out
of the cell. This leads to an increase in transepithelial Na transport. The mechanism of action of aldosterone is
similar to that of other steroid hormones. It involves entry of the hormone into the cell, binding to a cytoplasmic
                                                                       receptor, entry of the hormone-receptor complex
                   Acti ve Rece ptor Acti ve compl ex                  into the nucleus, followed by the production of
 HORMONE                                                               specific mRNA and increased biosynthesis of
           H        H          R      H     R           H     R                                                   +
                                                                       polypeptides,      in   this   case   Na       channels.
                                                                   A   Because of the biochemical steps involved, the
                                                                       response to aldosterone is slow, often on the
           A         A         R'      A R'                 mRNA       order of hours to days. The essential aspects of
                     Inacti ve          Inacti ve                      this mechanism are illustrated in Figure 9.
                     Re cep to r        co mpl ex       Protein
                                                                       (2)      Electrogenic, H -secreting epithelia such

Figure 9                                                               as gastric epithelia and distal segments of the
                                                                       renal tubule can increase the rate of H pumping
at the apical membrane by increasing the number of pumps in the membrane. In these epithelia, pumps are
present in cytoplasmic vesicles that are located under the apical membrane. In the gastric mucosa, fusion of
these vesicles with the apical membrane occurs upon stimulation of mucosa with histamine, causing an increase
in HCl secretion. In the distal renal tubule, insertion of pumps is stimulated by cytoplasmic acidification, probably
by an indirect mechanism that includes an elevation of cytosolic Ca            activity caused by the fall of intracellular pH.
Both responses involve insertion of pre-formed transporter and therefore, are rapid (seconds to minutes).

(3)    Epithelia that are responsive to antidiuretic hormone (collecting segments of the renal tubule) undergo an
                                                         increase in apical membrane water permeability that is
       0 addition of glucose
            or alanine                                   brought about by insertion of vesicles into the plasma
                                                         membrane. During the resting state these vesicles are near
  V                                                      the apical membrane. In the vesicular membrane are water
(mV)                                                     channels, that upon ADH stimulation can be seen in clusters
                                                         in freeze fracture images of the apical membrane. These
                                                         water channels greatly increase the permeability of the
                                                         apical membrane to water.
      -60                                                         The effects of ADH are mediated by increased
                                                         intracellular levels of cAMP. This mechanism of action is
Figure 10
                                                         different from that of aldosterone. Cyclic AMP is also the
second messenger of other hormones that exert effects on epithelial transport, such as parathyroid hormone
(PTH), calcitonin and catecholamines.

       ADH binds to a receptor in the basolateral membrane and activates a regulatory protein (G protein) which
upon activation binds GTP and in turn activates adenylate cyclase. The activation of the latter causes an elevation
of intracellular cAMP levels. The distal effect of cAMP involves protein phosphorylation by cAMP-dependent
protein kinases. The precise mechanism of insertion of the pore-containing vesicles has not been clarified, it is
clear, however, that intracellular Ca        is involved and that the vesicles fuse with the plasma membrane. The
response to ADH, as others mediated by cAMP, is rapid, on the order of minutes.

Basolateral Membrane

       The best documented cases of transport regulation by increased number of transporters in the basolateral
membrane involve secondary adaptation to increased entry at the apical membrane.

(1)    As previously discussed, in the small intestine, a portion of Na            entry at the apical surface is via
Na /organic solute cotransporters. This process carries net positive charge into the cell. If glucose or alanine, in
the presence of Na , is added to the lumenal surface of an in vitro small intestine preparation, a rapid
depolarization of the apical and basolateral membranes is observed. The depolarization of the apical membrane
is caused by the current flow produced by Na /organic solute cotransport (positive net charge enters the cell); the
depolarization of the basolateral membrane is caused by current flow from apical to basolateral membrane via the
low-resistance limiting junctions. As diagrammed in Figure 10, the depolarization is followed by a spontaneous
repolarization. It is likely that this repolarization is caused by an increase in the number of K channels in the
basolateral membrane. The insertion of K channels increases the total conductance of the membrane and its
potassium permeability, bringing the voltage closer to EK. It has also been suggested that insertion of K channels
is accompanied by insertion of Na,K pumps because the intracellular Na+ concentration changes very little during
this process.   Direct evidence for this hypothesis is still lacking.     Similar changes in basolateral membrane
properties can be produced by swelling the cells osmotically, suggesting that a change in cell volume (which is
certain to occur during Na entry) may trigger the events described.

(2)    Changes in the basolateral membrane in response to increased Na entry have been well studied in renal
cortical collecting tubules. During aldosterone treatment, as explained above, the number of apical membrane
  +                              +
Na channels and the rate of Na entry increase. In addition to this effect, after a delay ranging from less than an
hour to a few days, an increase in activity of the Na,K pump can be demonstrated. Interestingly, if during the

                                         +                                      +
period of aldosterone treatment, Na entry is blocked with a Na channel blocker, the increase in pump activity
does not occur even though synthesis of pump subunits is stimulated.                       These results suggest that
mineralocorticoids such as aldosterone stimulate synthesis of new Na,K pumps but their insertion into the
basolateral membrane require an elevation of intracellular Na .

         In addition to these effects on pump activity, prolonged mineralocorticoid treatment also produces an
increase in surface area and total conductance of the basolateral membrane, results consistent with insertion of
lipids and K+ channels as well.

Regulation of Transepithelial Transport by Modification of Pre-Existing Sites

         A number of chemical agents whose intracellular concentrations can be subject to modulation can regulate
the kinetic parameters of specific transporters involved in transepithelial transport. The three principal agents of
these effects are: pH, Ca        activity and cAMP levels. The effects of these agents and their interrelationships vary
among different epithelia. A few illustrative examples are described below.

Effects of changes of intracellular pH

         In electrogenic Na -absorbing epithelia, cell acidification has been shown to inhibit the Na,K pump,
                                 +                                                                                  +    +
reducing transepithelial Na transport. In NaCl-absorbing epithelia, intracellular acidification stimulates Na /H
exchange by both a mass action effect and an allosteric effect on the transporter. Because intracellular pH and
Ca      activity influence each other, complicated effects on transport parameters are frequently observed.

Effects of elevation of intracellular Ca2+ activity

                                                    2+                                 +                                2+
         In most epithelia, high intracellular Ca        activity causes an increase in K permeability by opening Ca -
              +                                                                                 +
activated K channels at the basolateral membrane, which hyperpolarizes the cells. In Na -transporting epithelia,
a reduction in Na permeability at the apical membrane has been demonstrated as well. Both effects are caused
by small changes, near physiologic levels, in the intracellular Ca .

Effects of elevation of intracellular cAMP levels

      In addition to its effect on the number of transporters as described above, cAMP can initiate the
phosphorylation of transporters. In Cl--secreting epithelia, cAMP causes an increase in apical membrane Cl-
permeability that stimulates secretion. In NaCl- absorbing epithelia, cAMP also increases apical Cl- permeability,
so that the Cl- that enters the cell via Cl-/HCO3- exchange is recycled back to the lumen via the Cl- channel.
                                                            +   +
Therefore net absorption is reduced. In addition, Na /H exchange is also inhibited by cAMP in these cells. Both
effects reduce NaCl entry and hence NaCl absorption. These effects are presumably mediated through cAMP-
dependent phosphorylation of the transporters/channels.

Epithelial Secretion by Exocytosis

         Many epithelial tissues and organs export large amounts of proteins. These substances are generally
transported from the intracellular to the extracellular compartment by exocytosis.            Such substances can in
principle be secreted to one of two compartments:                (1) a compartment external to the body (lumen of the
gastrointestinal tract), or (2) the extracellular fluid. If secreted into the extracellular space they can become part
of the extracellular matrix, stay in solution and exert effects locally, either on the same cell (autocrine action) or on
neighboring cells (paracrine action), or enter the blood and exert their effects on distant cells (endocrine action).
The mechanisms of exocytosis have been covered elsewhere and will not be repeated here. However, remember

that in addition to maintaining electrolyte homeostasis, most epithelia secrete protein products (enzymes,
hormones, etc.), that play an important role in their function.

                                                       CELL POLARITY


       The plasma membrane of polarized epithelial cells is divided into distinct apical and basolateral domains.
This functional polarity is largely determined by the presence of different lipids and plasma membrane proteins in
the two cell surface domains.          The apical domain has a high concentration of sphingolipids (mainly
glycosphingolipids) that may serve to protect the cell against harsh environmental influences. In contrast, the lipid
content of the basolateral membrane resembles the plasma membrane of nonpolarized cells being enriched in
phosphatidylcholine. Table 1 shows a list of the proteins that have been used as markers for the apical and
basolateral membranes. Both membrane and secretory proteins must be recognized as apical or basolateral,
segregated from each other, and delivered to the proper plasma membrane surface. The molecular mechanisms
responsible for the generation and maintenance of cell polarity are not fully understood, however, significant
progress has been made in the last decade in elucidating these processes. Two important advances have helped
in understanding the recognition, sorting, and directional transport of polypeptides to specific plasma membrane
domains; the establishment of epithelial cells in tissue culture and infection of polarized cells with enveloped RNA

Cultured Epithelial Cells

       The availability of polarized epithelial cell lines has been of major importance in the develop-ment of in vitro
systems to study the mechanisms of cell polarity. Some of the difficulties in studying intact epithelia from animals
                                                                            can be avoided by the use of cultured
                                                                            epithelial cells.   Several epithelial cell lines
                                                                            have been derived from both normal and
                                                                            malignant tissues and include cells originating
                                                                            from the kidney, urinary bladder, small and
    influenza virus buds only from   vesicular stomatitis virus buds only   large intestine, thyroid and mammary gland.
    the apical membrane              from the basolateral membrane
                                                                            Probably the best characterized epithelial cell
Figure 1.                                                                   line is the Madin-Darby canine kidney cell
(MDCK) which was isolated over 25 years ago from a normal, cocker spaniel kidney.                       This line has been
extensively used to characterize epithelial cell polarity.

Enveloped RNA Viruses

       In 1978 Rodriquez-Boulan and Sabatini demonstrated that RNA enveloped viruses bud in a polarized
fashion from infected MDCK cells. In these cells, different types of viruses bud exclusively from the apical or
basolateral domains (Figure 1). Viruses that typically bud from the apical membrane are influenza (a myxovirus),
and Sendai and simian virus 5 (paramyxoviruses); whereas vesicular stomatitis virus (VSV), a rhabdovirus, buds
from the basolateral membrane.        These viruses consist of a nucleocapsid core surrounded by a lipoprotein
membrane that is derived from the host plasma membrane.
       The viral envelope contains virally encoded glycoproteins that are synthesized by the infected cell and are
delivered to the plasma membrane in a manner that is similar to the normal delivery of native polypeptides.
Hence, the assembly of the viral particle is preceded by the accumulation of the respective envelope glycoproteins
at the membrane surface from which the budding will take place. Therefore, in addition to viral budding, sorting
can be studied using the viral envelope glycoproteins. A major advantage of using the viral envelope proteins is
that the cDNAs coding for them have been isolated and successfully reintroduced into cultured cells.

Table I. Typical polarity markers found in two or more epithelia.

 Protein                                    Domain                   Tissue

 Adenylate cyclase                              B                   Enterocyte
                                                B                   Hepatocyte
                                                B                   Renal tubule
 Alkaline phosphatase                           A                   Enterocyte
                                                B                   Hepatocyte
                                                A                   Placenta
                                                A                   Pig kidney line
                                                A                   Renal tubule

 Amiloride sensitive                            A                   Colon
  sodium channel                                A                   Frog skin
                                                A                   Renal tubule
                                                A                   Salivary gland
                                                A                   Urinary bladder

 ATP-dependent                                  B                   Enterocyte
 calcium uptake                                 B                   Renal tubule
 Dipeptidyl peptidase                           A                   Enterocyte
                                                A                   Renal tubule

 Furosemide-sensitive                           B                   Cornea
 Na,K, 2Cl                                      B                   MDCK
 cotransport                                    A                   Renal tubule
                                                A                   Shark rectal gland
 -glutamyl transferase                         A                   Enterocyte
                                                A                   Hepatocyte
                                                A                   Renal tubule
                                                A                   Pig kidney line
 H-2 antigens                                   B                   Enterocyte
                                                B                   Gall bladder
                                                B                   Hepatocyte
                                                B                   Trachea
                                                B                   Uterus
 Insulin Receptor                               B                   Hepatocyte
                                                B                   Renal tubule
 Leucine aminopeptidase                         A                   Hepatocyte
  (aminopeptidase N)                            A                   MDCK
                                                A                   Renal tubule
                                                A                   Thyroid Follicle
                                                A                   Enterocyte

 Maltase                                        A                   Enterocyte
                                                A                   Renal tubule
 Magnesium ATPase                               A                   Hepatocyte
                                                A                   Renal tubule
 Na,K-ATPase                                    B                   Avian salt gland
                                                B                   Choroid plexus
                                                B                   Enterocyte
                                                B                   Frog skin
                                                B                   Gall bladder
                                                B                   Hepatocyte
                                                B                   MDCK
                                                B                   Pancreas
                                                B                   Renal tubule
                                                B                   Salivary gland
                                                B                   Shark rectal gland
                                                B                   Sweat gland
                                                B                   Urinary bladder
 Neutral Endopeptidase                          A                   Enterocyte
                                                A                   Renal tubule
 5’ Nucleotidase                                A                   Enterocyte
                                                A                   Hepatocyte
                                                A                   Renal tubule

 A = apical, B = basolateral
 Organism is mammalian unless indicated

Biogenesis and Sorting of Apical and Basolateral Proteins

       You will recall that similar to secretory proteins, plasma membrane glycoproteins are synthesized by bound
polysomes in the endoplasmic reticulum (ER). It is at this point that they acquire their initial glycosylation (core
glycosylation) and their asymmetric orientation with respect to the membrane. Studies in fibroblasts infected with
viruses have shown that from the ER, viral envelope glycoproteins are transported to Golgi vesicles. In the Golgi
the glycoproteins are concentrated, terminally glycosylated, and shipped via vesicles to the cell surface. Studies
using MDCK cells infected with viruses have demonstrated that the viral envelope G protein from VSV, and
hemagglutinin (HA) of influenza, are sorted intracellularly and delivered in a polarized fashion to the appropriate
plasma membrane domain.           In MDCK cells coinfected with influenza and VSV, each virus preserved its
characteristic polarity of budding.   Using immunoelectron microscopy both the apically directed HA, and the
basolateral G protein were detected in the same Golgi apparatus. Both envelope glycoproteins exhibited similar
kinetics of migration to the Golgi when intracellular transport of HA and G was synchronized using temperature-
sensitive mutants of both viruses. Biochemical data indicate that G protein and the apical glycoprotein, influenza
neuraminidase, share the synthetic pathway at least as far the terminal cisternae of the Golgi apparatus where
sialic acid is added to G protein. These results suggest that the pathways of apical and basolateral membrane
proteins are identical prior to passage from the Golgi apparatus. In transfection experiments using eucaryotic
expression plasmids that contain the cDNAs for either the G or HA glycoproteins, both proteins are sorted to the
proper plasma membrane domain demonstrating that sorting can occur even in the absence of other viral
components.     These and other experiments have made it clear that the polarized sorting of G and HA
glycoproteins takes place soon after their passage through the Golgi apparatus since both glycoproteins are found
in the same Golgi cisternae. Thus, upon arrival in the trans-Golgi network (TGN), viral glycoproteins are sorted
from one another and packaged into distinct post-Golgi transport vesicles. The analysis of many other different
exogenous and endogenous apical and basolateral proteins is consistent with this observation. Consequently, it
appears that in MDCK cells sorting mainly takes place on the exocytic route upon exit from the TGN.

       Studies in other types of epithelial cells have shown that the TGN is not the only site involved in the sorting
of newly synthesized plasma membrane proteins. In hepatocytes it appears that all apical membrane proteins are
first delivered to the basolateral surface where they are sorted from resident basolateral proteins and transported
to the apical surface by transcytosis. Transcytosis involves the delivery of membrane proteins and/or ligands via
endocytic vesicles to the opposite side of the cell. Each hepatocyte has several apical poles that line the bile
canaliculi and are separated from the intervening basolateral domains by tight junctions. In these cells both apical
and basolateral membrane proteins are delivered from the TGN only to the basolateral domain. Subsequently, the
apical proteins are endocytosed and transported via the transcytotic route to the apical membrane. In hepatocytes
the large quantities of albumin and other secretory proteins produced are secreted solely from the basolateral
membrane.     Consequently, the secretory proteins present in the bile are derived from the blood by either
transcytosis across the liver cell or passage through the intercellular space. Thus it appears that hepatocytes lack
the direct sorting mechanism from the TGN to the apical surface and that sorting occurs only through the
endocytic/transcytotic pathway.

       Interestingly, in contrast to the sorting in MDCK cells and hepatocytes, it appears that in intestinal epithelial
cells newly synthesized apical proteins make use of both a direct and endocytic/transcytotic pathway for delivery.
In the cultured intestinal epithelial cell line, Caco-2, as well as primary intestinal epithelial cells, it appears that
apical membrane proteins are sorted in the TGN and routed directly to the apical surface. However, some apical
membrane proteins, in addition to being delivered directly to the apical surface are routed to the basolateral
domain. At the basolateral surface the apical proteins are transcytosed to the apical surface. Basolateral proteins
seem to be always directly transported to the basolateral membrane domain.

       Taken together these results suggest that, depending on the epithelium, different routes for the
establishment and maintenance of cell polarity are present and/or emphasized. Thus, vesicular sorting can occur
at two major sites: the TGN and at the plasma membrane through endocytosis. Different epithelial cell types rely
on these different sorting sites. MDCK cells sort nearly all their apical and basolateral proteins in the TGN and
deliver these proteins directly to the appropriate membrane domain. In contrast, in hepatocytes proteins are
initially targeted from the TGN to the basolateral membrane. After endocytosis into basolateral early endosomes,
apical proteins are segregated and directed to transcytotic vesicles. Intestinal cells utilize both routes to a roughly
equal extent to sort apical and basolateral proteins.

Determinates of Polarized Protein Sorting

       Among the extensive vesicular trafficking of epithelial cells, what are the signals that direct proteins into a
particular pathway?    It has been well known that certain features of a protein can influence its inclusion or
exclusion into specific organelles.    However, the identification of specific signals that direct the sorting of
polypeptides to either the apical or basolateral domains have remained more elusive. Nevertheless, in the last
several years strides have been made in identifying signals that are important in targeting proteins to their
respective plasma membrane domains.

Apical Sorting Signals

       The first epithelial sorting signal identified was a result of the observation that proteins anchored to the cell
surface through a specific, covalent linkage to the glycolipid glycosylphosphatidyl inositol (GPI), are almost
exclusively confined to the apical surface.      As diagrammed in Figure 2, this linkage involves the covalent
attachment of the polypeptide carboxyl-terminus via ethanolamine to an oligosaccharide chain, which in turn is
linked to the inositol ring of phosphatidylinositol. After synthesis in the ER the C-terminus of the anchored protein
is cleaved and the covalent linkage added. As mentioned, most GPI linked proteins are found solely at the apical
surface. Moreover, the addition of GPI to exogenous secretory proteins or to the ectodomain of VSV-G protein
results in the apical delivery of the modified polypeptides. Although a specific protein sequence is not a signal for
GPI addition, it appears that the anchored proteins have a similar "motif" at the C-terminus. Studies of decay-
accelerating factor (DAF), a GPI anchored regulatory polypeptide, have indicated that the signal for GPI addition
lies in the 37 C-terminal amino acids. During GPI addition these amino acids are cleaved. The region consists of
a 17 amino acid hydrophobic portion at the C-terminus with a 20 amino acid adjacent region. Both regions are
required for GPI addition; in addition random hydrophobic amino acids can substitute for the hydrophobic C-
terminal tail without affecting anchoring. These results suggest that the signaling mechanism for GPI addition
must recognize domains at the C-terminus.

       As mentioned, as a result of sorting and vectorial transport the apical membrane is enriched in
glycosphingolipids. Glycosphingolipids have been shown to self-segregate from other membrane constituents by
their inherent ability to form intramolecular hydrogen bonds.             It appears that GPI-anchored proteins,
glycosphingolipids and a few transmembrane proteins (HA) form hydrogen-bonded “rafts” or microdomains within
the TGN. The rafts also contain the src-substrate caveolin, a component of the plasma membrane caveolae.
These differentiated lipid microdomains, with their selectively partitioned proteins, may be selectively configured
for apical transport. Thus, apical sorting may involve the selective clustering of proteins into glycolipid- and
cholesterol containing membrane rafts in the exoplasmic leaflet of the Golgi. However, apical transport of many
other proteins appears to occur independently of rafts. Therefore, apical sorting is likely to involve more than one
mechanism, but the details of these mechanisms are unclear. The exact mechanism and recognition signals that
direct glycosphingolipids and GPI-anchored proteins to the apical surface will have to await further study.

                                                                       The search for apical sorting signals
                 O                                              directed by the primary structure has been less
                 C-NH                                           productive.   Early results demonstrating that a
                 CH 2       ethanolamine                        truncated HA lacking the transmembrane and
                 CH 2                                           cytoplasmic domains is secreted into the apical
                                                                medium, and a chimeric glycoprotein with the HA
                                                                ectodomain and the membrane spanning and
                 GLYCAN-GlcNH 2
                                                                cytoplasmic    sequences          of        G     protein   is
                                                 O   O   preferentially sorted to the apical surface,
                                                 C=O C=O suggested that the apical sorting signal was
                                                                contained in the lumenal domain.                 Moreover, it
                                                                appears that apical sorting information is localized
                                                                to the lumenal domain of human aminopeptidase
Figure 2. GPI anchor                                            N.   However, because of the lack of sequence
homology among apical proteins, it has been proposed that the sorting information is not formed by a linear
sequence of specific amino acids but may be a subtle structural motif. Alternatively, apical sorting information
could be contained in widely distributed post-translation modifications, such as specific carbohydrate moieties.
One possible mechanism involves recognition of carbohydrate moieties by sorting lectins, such as VIP36, that
partition into microdomains of the TGN along with other apical membrane proteins and lipids.                     It has been
suggested that N-glycans play a role; however, apical sorting of many proteins occurs independently of N-
glycosylation. The exact mechanisms of carbohydrate-mediated apical sorting are unknown.

Basolateral Sorting Signals

       In the last few years substantial progress has been made in identifying signals important in basolateral
transport. In contrast to apical protein sorting which seems to be more dependent upon signals within the lumenal
domain, protein sorting to the basolateral domain is dependent upon the cytoplasmic domain (Figure 3). Distinct
signals have now been identified in the cytoplasmic domains of membrane proteins that direct their sorting to the
basolateral cell surface. Upon removal or inactivation of these signals the basolateral proteins are transported to
the apical membrane. In addition, if these signals are transferred to the cytoplasmic domain of an apical protein it
                                                                         is   redirected         to     the       basolateral
                                                                         membrane.       Thus these sorting signals
                                                                         are sufficient to direct polypeptides to the
                                                                         basolateral domain.

                                                                                   To date three types of basolateral
                                                                         sorting signals have been identified. The
                                                                         first type of targeting signal is superficially
                                                                         related to clathrin coated pit endocytosis
                                                                         signals      (usually        4-6       amino   acids
                                                                         containing a tyrosine). This type of signal
                                                                         has been found in a variety of a
                                                                         basolateral proteins and generally rely on
                                                                         the same tyrosine residue that is required
                                                                    for endocytosis. The best characterized
Figure 3. Sorting Signals and mechanisms for basolateral and apical
transport.                                                          basolateral determinant of this type is

                                                                            found on the low density lipoprotein
                             PROXIMAL                        DISTAL
                                                                            receptor (LDL-receptor).      This receptor
                              SIGNA L                        SIGNAL         actually    contains    two    independent
       KNWRLKNINSINFDNPVYQKTTEDEVHICHNQDGYSYPSRQMVSLEDDVA                   basolateral targeting signals within its 50
                       PIT SIGNAL                                           residue cytoplasmic tail (Figure 4). In the
                                                                            membrane-proximal position there is a
                                                                     clathrin-coated pit localization signal while
Figure 4.    Arrangement of basolateral-targeting signals in the
cytoplasmic domain of the human LDL receptor. The locations of the in the distal position is a signal unrelated
two basolateral-targeting signals and the internalization signal are to endocytosis determinants. Another
                                                                     basolateral signal that is non-tyrosine
based, exists within the cytoplasmic tail of the polymeric immunoglobulin receptor (pIg-R). This receptor mediates
the transport of immunoglobulins (Ig) across epithelium. At the basolateral membrane, the Fc portion of the
immunoglobulins IgA and IgM bind to pIg-R, where they are internalized through receptor mediated endocyt osis.
Receptor and ligand are then transported to the apical membrane where the extracellular portion of the receptor,
termed the secretory component (SC), along with its ligand, are released into the extracellular medium.

       Studies of the sorting and transcytosis of the pIg-R transfected into MDCK cells has provided some insights
into the signals involved in the basolateral delivery and transcytosis of the receptor. From these studies it has
been shown that the 17 cytoplasmic amino acids next to the membrane serve as a signal to direct the receptor
from the TGN to the basolateral membrane. Removal of this segment results in a receptor that is mistargeted
from the TGN to the apical surface. Furthermore, this sequence is sufficient to redirect the normally apical protein,
alkaline phosphatase, to the basolateral membrane. It appears that this sorting signal is similar to other signals, in
that it contains a type I ß-turn. It has been proposed that this type of signal, based on the exact amino acid
sequence, can serve as a basolateral targeting signal, internalization signal, TGN retrieval signal, lysosomal
targeting signal or combinations of these functions. This hypothesis suggests that these various ß-turn signals are
recognized by similar proteins that compose part of the sorting machinery of the cell. In support of this notion, it
appears that changes in the basolateral signal of the pIg-R influences its ability to be transcytosed.

       It has been previously shown that a serine within the pIg-R basolateral targeting signal is phosphorylated
(Ser-664). If this serine is changed to alanine (a residue that is not phosphorylated), the receptor is still targeted to
the basolateral surface, however, it trans-cytoses very slowly to the apical domain. The change in the receptor's
ability to undergo transcytosis does not reflect its inability to enter the endocytic pathway since it rapidly cycles
through endosomes. Conversely, conversion of serine to aspartic acid (an amino acid that mimics the negative
charge of phosphate) results in rapid transcytosis. Thus, it appears that the basolateral targeting signal also acts
as a basolateral retrieval signal. When the protein is internalized the signal directs it back to the basolateral
membrane. In the case of the pIg-R, it appears that phosphorylation inactivates the retrieval signal allowing the
protein to be transcytosed to the apical surface. Other putative basolateral sorting signals are outlined in Figure 5.

       Another completely different type of basolateral targeting signal has recently been identified. This signal,
consisting of the sequence LeuLeu, was found in the mouse Fc receptor for IgG. Originally the signal was
implicated in targeting proteins from the TGN to endosomes and lysosomes, however, the di-leu (or di-
hydrophobic) signal has also been shown to function in basolateral targeting and endocytosis.              There is no
structural data on the conformation of the LeuLeu motif, however, it has been suggested that it may be contained
in a ß-turn.
       All these basolateral targeting motifs may direct the selective binding of proteins to clathrin adaptor proteins
(AP-1, AP-2, AP-3) that are selectively packaged in the TGN into basolateral sorting vesicles.        The role of other
proteins, the formation of specific sorting vesicles and their mechanism of transport to the plasma membrane

awaits further study.

Endocytic and Transcytotic Pathways for the Sorting of Proteins in Polarized Cells

       The sorting of proteins at the basolateral domain suggests a difference between apical and basolateral
postendocytic cycling pathways.     Support for this possibility comes from characterization of the apical and
basolateral endocytic pathways in MDCK cells. Using fluorescent fluid phase markers to analyze the organization
of apical and basolateral endosomes, and biochemical methods to quantitate recycling, transcytosis and transport
to later endocytic compartments, it has been demonstrated that the early apical and basolateral endosomes are
                                                                            functionally and topologically distinct.
  Tyr osine -De pe nde nt Basolate r al Sorting Signals                     Although the endosomes from the

      LDLR              9NSINFDNPVYQKTTEDEVHICHN                            different plasma membrane domains
                                                                            are initially separate, they converge at
      LAP                   R
                        4 0 5 MQAQPPGYRHVADGEDHA
                                                                            later stages in the endocytic pathway.
      HA Y5 43                  N
                            5 3 8 GSLQYRICI
                                                                            Analysis of the kinetics of fluid phase
      ASGPR H1                  1MTKEYQDLQML
                                                                            marker internalization, recycling and
      Igp120               1RKRSHAGYQTI
                                                                            transcytosis    revealed      fundamental
      TGN38             5VTRRPKASDYQRLNLKL
                                                                            differences    in   the     trafficking    of
  Di-Hydrophobic Basolate r al Sorting Signals                              endosomes derived from the apical and

    FcRII-B2          22NTITYSLLKH                                          basolateral domains. Although fluid is
                                                                            internalized    from      either    plasma
                                                                            membrane domains with the same rate
  Figure 5. Sequence of basolateral sorting signals in the cytoplasmic      per membrane area, the majority of the
  domains of several proteins. Amino acids important for sorting are in     marker     molecules      (73%)    in     the
  bold. Abbreviations: LDLR, low density lipoprotein receptor; LAP,
  lysosomal acid phosphatase; HA Y543, hemagglutinin T for C                basolateral early endosomes are routed
  substitution at 543; ASGPR, asialoglycoprotein; lgp 120, lysosomal        to the late endocytic compartments.
  protein 120; TGN38, TGN marker protein; FcRII-B2, Fc receptor II-B2       Only 13% of the internalized markers
  isoform; MHC II/Ii, major histocompatability class II invariant chain.
                                                                            are recycled back to the basolateral
surface while 14% are transcytosed. These results are in contrast to the previously described studies in which a
large portion of the pIg-R was recycled back to the basolateral surface and to earlier studies characterizing the
basolaterally located, transferrin receptor. In MDCK cells, the endogenous transferrin receptor, when internalized
at the basolateral surface, recycles solely to the basolateral membrane. Thus as outlined before, it appears that
specific signals within the basolateral receptors can direct their passage into specific endocytic pathways.

       In contrast to the basolateral early endosomes, the apical endosomes direct only 10% of their contents to
the late endocytic compartments. The bulk of the contents in the apically derived endosomes are released equally
from the cell through recycling and transcytosis. These differences in membrane sorting from the early apical and
basolateral endosomes suggest important functional differences in the pathways, though the mechanisms and
significance of these differences are poorly understood.

Differences in the Sorting of Secreted and Plasma Membrane Proteins

       Most of the studies involved with the truncated, secreted glycopeptides have assumed that the polarized
sorting of secreted molecules follows the same pathway as the native membrane glycoprotein. However, this may
not be the case. It has been shown that secretion of the basement membrane components laminin and heparin
sulfate proteoglycan (HSPG), takes place from the basolateral surface. The sorting process which mediates this
polarized secretion requires an acidic intracellular compartment since MDCK cells treated with NH 4Cl (which

raises the pH of intracellular compartments), secrete laminin and HSPG from both the apical and basolateral
membranes. However, NH4Cl treatment does not seem to affect the sorting of apically secreted proteins, and
surprisingly, it does not affect the polarized delivery of membrane proteins to the plasma membrane. In MDCK
cells incubated with NH4Cl, HA was directed to the apical surface and the Na,K-ATPase to the basolateral.
Furthermore, in rat liver, two populations of TGN-to-basolateral vesicles have been isolated. One population of
vesicles is enriched with soluble, secretory proteins, while the other is enriched in membrane proteins. Therefore,
it appears that the sorting mechanisms that operate for the polarized delivery of membrane proteins may be
distinct from those that govern secretory protein sorting. Also, it appears that the sorting pathway of apical and
basolateral secreted proteins must be mechanistically different since an increase in pH of the intracellular
compartments only inhibited basolateral sorting.

Involvement of the Cytoskeleton

       The cytoskeleton may also play an important role in maintaining epithelial polarity. In certain clones of
MDCK cells the Na,K-ATPase is directed to both the apical and basolateral membranes from the TGN. However,
at the apical membrane the enzyme is rapidly internalized at a rate similar to that of fluid-phase endocytosis. In
contrast, at the basolateral membrane the Na,K-ATPase interacts with the cytoskeleton by forming a stable
complex with ankyrin and fodrin.       This interaction stabilizes the enzyme at the basolateral membrane by
preventing its entry into endocytic vesicles. Therefore, while the TGN is the primary location of the polarized
sorting of proteins, in some MDCK cells the selective stabilization of proteins by the cytoskeleton may also
contribute to establishing and maintaining polarity.

       In addition to the role of the cytoskeleton in stabilizing proteins at the plasma membrane, cytoskeletal
elements may participate in the delivery of proteins to the plasma membrane.             The common view is that
microtubules are required for the intracellular transport of vesicles carrying newly synthesized membrane and
secretory proteins. This model would involve the selective association of the intracellular vesicle with the
cytoskeleton, followed by movement and insertion by exocytosis into the appropriate plasma membrane.
However, it has been shown that neither colchicine nor cytochalasin D, two agents that disrupt microtubules, alter
the polarized budding of influenza or VSV. Other results suggest that while movement to the apical domain of
epithelial cells is inhibited, movement to the basolateral domains is not. This result is consistent with the tenfold
higher concentration of microtubules in the apical region than the basolateral one. Thus, it is apparent that the
possible role of the cytoskeleton in the polarized sorting of proteins is not understood and must await further study.


       What sense can be made from all of these observations? Are the sorting signals for apical and
basolateral membrane proteins general for all epithelial cells or do we have to invoke cell specific
mechanisms? For example, will hepatocyte apical proteins be recognized by the sorting machinery in
the TGN of MDCK cells? Evidence against cell specific sorting signals comes from experiments using
aminopeptidase N. Aminopeptidase is an apical protein normally present in Caco-2 cells. When
synthesized in Caco-2 cells it reaches the apical membrane by both the direct and transcytotic routes.
When the enzyme is expressed in MDCK cells it is sorted in the TGN and delivered directly to the
apical surface. In hepatocytes, the endogenous enzyme reaches the apical membrane via transcytosis
from the basolateral membrane. Taken together these results suggest that depending on the
epithelium, different routes for the establishment and maintenance of cell polarity are present and/or
emphasized. For example most strains of MDCK cells efficiently segregate and deliver apical and
basolateral proteins to their respective plasma membranes directly from the TGN. In contrast, in

hepatocytes the segregation of apical proteins from basolateral proteins is absent and sorting must
take place at the basolateral membrane through endocytosis/transcytosis. Finally, Caco-2 cells use
both pathways to sort polypeptides. The exact machinery and mechanism(s) that sorts and directs
apical and basolateral destined vesicles to their respective plasma membranes is largely unknown.
Also, the exact signal(s) that direct proteins into the various vesicle transport pathways, and what
mediates the differences in the delivery of secreted and plasma membrane proteins will have to be
determined. However, some general conclusions can be drawn:

1.   Proteins can be specifically delivered to either the basolateral or apical membrane surface.

2.   Different epithelial cells emphasize different pathways in the sorting mechanism.

3.   Anchoring of polypeptides to the membrane through glycosylphosphatidyl inositol may be an
     important sorting signal to send some polypeptides to the apical surface. In addition, glycosylation
     may play a role in the apical delivery of some proteins.

4.   Protein-based basolateral sorting signals within the cytoplasmic domain of some proteins have
     been identified.

5.   The sorting mechanisms for secreted and membrane proteins may be different.


Facilitated Diffusion: This process, as diffusion, can produce net movement only down concentration gradients.
However, unlike simple diffusion, this transport requires the presence of specific transport proteins. Because of
this transport protein requirement, facilitated diffusion exhibits specificity, saturability and substrate competition.

Primary Active: In this process, molecules are transported against their electrochemical gradient. The energy
for this process comes directly from the splitting of ATP. Primary active transporters include: the Na,K-ATPase
(the most important enzyme known), H,-ATPase, H,K-ATPase, K-ATPase and Ca-ATPase.

Secondary Active Transport: This process involves the movement of two or more substrates by specific
transport proteins. Cotransport involves the movement of all substances in one direction, while countertransport
requires movement of substrates in opposite directions. The important concept in secondary active transport is
that one substance, moving down its chemical gradient (or electrochemical gradient in the case of ions) powers
the transport of the other substrate against its gradient. This process occurs without the direct input of metabolic

Endocytosis/Transcytosis: This process, characterized by the invagination of plasma membrane and the
transport of intracellular vesicles, is an important mechanism for the transport of macromolecules.


Na,K-ATPase                                               Ca-ATPase
                                  out   in                                                out   in

2 major subunits            2K                            Mr = 110 kD
 Mr = 112 kD                               ADP + Pi                                                ADP + Pi

ß Mr - 60 kD                                 ATP          SR Ca-ATPase inhibited                     ATP
                                             3Na          by thapsigargin,                           2Ca
inhibited by ouabain,                                     cyclopiazonic acid
digoxin, digitalis
                                                          Other P-Type Ion Transport ATPases:

H,K-ATPase                        out   in
2 subunits                   K                            Cd-ATPase
 Mr = 114 kD                               ADP + Pi     K-ATPase
ß Mr - 60-85 kD                              ATP
                                             H            FoF1-Type ATPases:
inhibited by Schering
28080,omeprazole,                                         H-ATPase
lansoprazole                                              K-ATPase


Na/H Exchange r                                           Anion Exchange r
                                  out   in                                                out   in
                                                          Cl/HCO3 Exchange r
Mr = 90 kD                  Na                                                       Cl
                                                          Mr - 102 kD
inhibited by amiloride
                                             H            inhibited by DIDS, SITS                    HCO3
                                                          phenyl isothiocyanate

Na/Ca Exchange r
                                  out   in

Mr = 108 kD                3Na

inhibited by dichloro-
benzamil, exchanger                          Ca
inhibitory peptide (XIP)


Na,K,2Cl Cotransporter                                    Na,glucose
                                  out   in                                                out   in
                                                          Cotransporte r
Mr = 120 kD                Na,K                                                     2Na
                                                          Mr = 73 kD
inhibited by bumetanide
furosemide             2Cl                                inhibited by        glucose

Selected Examples of Epithelial Ion Transport

                             GASTRIC PARIETA L CELL

                                       K+                                                "a lkalin e
                                                                                             tid e"
                                                  H+ HCO 3           HCO 3         HCO3

                                                       CA                          Cl
                                            K+      H2O                            Na+
                                            Cl       +
                                                    CO 2             H+


                             Lu me n

                             PANCREATIC DUCTAL CELL

                                                                                          "a ci d
                                                 HCO3     H+          H+                H+

                                                       CA                               Na+


                             SMALL INTESTINAL CELL

                                            H2O                Na+

                                                                                   Na+ K +
                                            Cl                                     2Cl

                                             cAMP                    K+

                             Lu me n


To top