| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Chemistry and Biochemistry and the Program in Biochemistry and Molecular Biology, University of California, Santa Barbara, California 93106, USA
1Correspondence: Department of Chemistry and Biochemistry, University of California, Santa Barbara, CA 93106, USA. E-mail: parsons{at}chem.ucsb.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT BACKGROUNDRECENT RESULTSMATTERS OF CONTROVERSY AND...PROSPECTS AND PREDICTIONSREFERENCES |
|---|
Key Words: vesicular acetylcholine transporter • vesicular monoamine transporter • neurotransmitter transport
|
BACKGROUND |
|---|
|
TOPABSTRACTBACKGROUNDRECENT RESULTSMATTERS OF CONTROVERSY AND...PROSPECTS AND PREDICTIONSREFERENCES |
|---|
VMATs arising from different genes (endocrine VMAT1 and neuronal VMAT2)
and VAChT have been cloned from species ranging from
Caenorhabditis elegans and Drosophila to humans
(1
2
3
4
5)
. The predicted amino acid sequences are consistent
with 12 transmembrane domains (TMDs, Fig. 1
). Although the review assumes that 12 TMDs are present, little direct
work testing the model has been conducted. Many residues in or near
every putative TMD are conserved across VMAT1, VMAT2, and VAChT,
suggesting similar structures for TMDs and mechanisms of transport.
VMAT isoforms appear to be similar enough to each other in transport
mechanisms that they will not be distinguished except when important,
and the terms vesicle and granule are used interchangeably.
|
The physical chemistry of uptake and storage
This topic is well reviewed (6
7
8
9
10
11)
, and important
points are summarized without primary references. The values of
transmembrane pH and electrical gradients (
pH and 
,
respectively) are about -1.4 pH units and +39 mV, respectively, for
filled cholinergic and monoaminergic secretory vesicles in
situ. The effects of different pH and values on
equilibrium gradients of monoamines in vitro demonstrate
that VMAT exchanges two protons for each monoamine of plus-1 charge
taken up. The estimated driving force for monoamine uptake in
vivo is 630-fold from pH and 4.8-fold from for a total
3000-fold in equilibrium activity gradient. The exchange stoichiometry
and resulting driving force for ACh uptake were not explicitly
determined until recently (see below).
The concentration of cytoplasmic monoamines in nerve terminals is in
the low µM range. The concentration in secretory vesicles is
10,000-fold higher. Because the concentration gradient is larger
than the predicted activity gradient, monoamines in vivo
historically were assumed to be in equilibrium with vesicular pH and
. The excess concentration gradient is due to stabilization of
stored monoamines by intravesicular association, which lowers activity.
However, the extent of stabilization is not known accurately and recent
evidence suggests that equilibrium often is not attained (see below).
Pharmacology
A large family of inhibitory compounds is known for VAChT
(12
, 13)
. The lead compound vesamicol binds to the
cytoplasmic side (14)
with an equilibrium dissociation
constant (KD) value of 5 nM
(15)
. Binding sometimes exhibits positive cooperativity,
but this is a variable observation (16
, 17)
. Vesamicol
analogs inhibit transport noncompetitively (16
, 18)
, but
they apparently need bind only a fraction of VAChT in isolated vesicles
to inhibit fully (19)
. Addition of vesamicol to vesicles
after uptake of [3H]ACh stimulates release of
much of the recently transported ACh but not the endogenous ACh
(20)
. The significance of these complexities in the
effects of vesamicol is unknown. The vesamicol dissociation constant is
insensitive to ‘energization’ of vesicles (establishment of pH or
). Comparison of their structure–activity relationships suggests
that vesamicol and ACh bind to different sites in VAChT, even though
the molecules compete with each other for binding at equilibrium
(16)
.
Three major inhibitory compounds are known for VMAT. Reserpine binds to
the cytoplasmic side (21)
with a
KD value of 25–340 nM. This decreases
to 30 pM when vesicles are energized (22
23
24
25)
.
Monoamines inhibit the increase in reserpine affinity with
IC50 values (in the presence of ATP) similar to
their transport Michaelis constant (Km) values
(22
, 25
, 26)
. The similarity suggests that substrates and
reserpine compete for the same form of VMAT, probably because reserpine
is a dead-end inhibitor in the first translocation step of the
transport cycle (see below). High-affinity reserpine binding to mutants
indicates that the first translocation step is intact, even if the
mutation blocks steady-state transport. Ketanserin and tetrabenazine
also are potent inhibitors of VMAT. Energization has no affect on their
dissociation constants (27
, 28)
.
Relative values for substrate KD,
Km, and Vmax are
distinctive
Binding and initial velocity transport of substrates exhibit
hyperbolic saturation. VAChT binds ACh with a
KD value 10- to 100-fold larger than
the Km value of 0.3–1 mM (16
, 29
, 30)
.
VMAT likewise binds monoamines with KD
values 10- to 100-fold larger than the corresponding
Km values of 1–10 µM (24
, 31)
.
Different values of KD and
Km can arise from a single binding site when at
least one catalytic rate constant is a major contributor to
Km. The Km will be smaller than
KD when uptake of the occupied binding site for
substrate is fast relative to reappearance of the unoccupied binding
site during transport (see below).
VAChT transports ACh and large synthetic analogs of ACh with the same
Vmax values (32
, 33)
. It
does not significantly transport choline and monoamines
(34)
. VMAT likewise transports different monoamines and
synthetic substrates with the same
Vmax values (35
36
37
38
39)
.
Because the amount of damage to vesicles caused by isolation differs in
different preparations, the same preparation must be used to compare
Vmax values for different substrates
(33
, 40)
. Substrate-independent Vmax
suggests that a substrate-independent step in the transport cycle is
rate-limiting. This probably is reappearance of the unoccupied binding
site for substrate in the last step of the cycle, which is consistent
with a small Km/KD
ratio. A report that Vmax for dopamine
exceeds that of serotonin by three- to fivefold in cloned VMAT is not
reconciled with this conclusion (41)
.
Effects of pH reveal important titratable residues
When the amount of binding or transport by inhibitors and
substrates is plotted vs. pH, bell-shaped curves often are obtained
that reflect two protonation events. The events are of particular
interest because they might be related to proton antiport sites. The
data tell us more than usually noted. When external pH is varied under
standard transport conditions, internal pH varies relatively little due
to ‘clamping’ by vacuolar ATPase activity, which is nearly constant
over the pH range typically studied (42)
. When a large
effect of external pH on transport occurs, it likely arises from
protonation of an external transporter site. In an equilibrium binding
measurement, one does not know whether a pH effect arises from the
inside or outside because pH is the same on both sides of the membrane.
However, because all of the equilibrium effects are present in
transport measurements, all pH effects reviewed in this section
probably occur from the outside.
Transport of subsaturating ACh requires deprotonation with
pKa 7.3 and protonation with
pKa 9 (16)
. Because a
change in either Vmax or
Km alters transport rate in subsaturating substrate
(as rate 
[substrateo]Vmax/Km),
these data do not determine whether
Vmax or Km is
affected by pH. Equilibrium binding of vesamicol or a high-affinity
synthetic analog of ACh requires deprotonation with
pKa 7.1 and 7.4, respectively
(43
44
45)
. Vesamicol binding also requires protonation with
pKa > 9. However, protonated vesamicol
itself (which is a tertiary amine) has a similar
pKa value. Whether an additional
protonation in VAChT is required to bind vesamicol is undetermined.
Transport of subsaturating serotonin requires VMAT deprotonation with
pKa 7.1 and protonation with
pKa 9 (46)
. The
Vmax is approximately constant from pH
6.5 to 8.5 and then decreases at higher pH with
pKa 9.0 to 9.5. Km
decreases as pH increases with pKa
7.1–7.4 in most experiments (35
, 36
, 42
, 47
, 48)
.
Equilibrium binding of [3H]dihydrotetrabenazine
or norepinephrine requires VMAT deprotonation with
pKa 7.3 (31)
. Finally,
high-affinity binding of reserpine requires deprotonation with
pKa 7.3 and protonation with
pKa 9.5 (46)
.
Some of the pH studies are fraught with large experimental error, changed conditions between experiments, and change in the protonation state of the probing radiolabeled substrate or inhibitor at different pH values. Critical evaluation of the data using Occam’s Razor leads to the following conclusions that incorporate recent findings.
1) A single external residue of pKa 7.1–7.4 in each transporter type might account for all deprotonation requirements for both inhibitor and substrate binding. This suggests that deprotonation controls the KD term in the mathematical expression for substrate Km (below). A pKa value of 7.1–7.4 is consistent with histidine and aspartate or glutamate in a hydrophobic environment. Whether the residue is conserved between the transporter types is not known.
2) Two pathways for proton translocation exist in each transporter (see Recent Results for further evidence on VAChT). By analogy to lactose permease (below), an acidic amino acid in each pathway probably moves between aqueous luminal and cytoplasmic exposures to bind and release a proton, respectively. Because protons would move preferentially through much of the transporter in a likely aqueous transport channel or channels, a proton ‘wire’ passing protons from amino acid to amino acid probably does not operate.
3) If it reorients, the residue of pKa 7.1–7.4 could translocate the second proton. Alternatively, it could interact (for example, as an ion pair) with a separate translocation site for the second proton. This residue is unlikely to be the first proton translocation site, because that site probably is inwardly oriented when substrate binds (below).
4) With the possible exception of the residue of pKa 7.1–7.4, proton translocation sites have pKa < 6.5 in the outward orientation. This is because other proton translocation sites do not reveal themselves in these pH studies, which implies that they are essentially fully deprotonated above pH 6.5.
5) An unidentified external residue with
pKa 
9 controls proton
translocation, but it does so indirectly. It cannot itself translocate
a proton because the required protonation state is not compatible with
antiport.
|
RECENT RESULTS |
|---|
|
TOPABSTRACTBACKGROUNDRECENT RESULTSMATTERS OF CONTROVERSY AND...PROSPECTS AND PREDICTIONSREFERENCES |
|---|
pH collapses
(49)
. The time courses of uptake at different
pHo, initial pHi and
[ACho] contain information about microscopic
rate and equilibrium dissociation constants. Microscopic constants
apply to elementary steps in a kinetic mechanism, whereas the
macroscopic constants Vmax and
Km apply to net steady-state behavior composed of
many microscopic steps. A microscopic kinetic mechanism restricts the
range of allowed chemical mechanisms much more than the macroscopic
kinetic mechanism does.
Numerically integrated differential equations describing the two-proton
exchange mechanism shown in Fig. 2
fit the data much better than similar equations for one-proton
mechanisms (43)
. A more positive generated with
Ko+-valinomycin at the same
initial pHi stimulates pH jump uptake, and
permeant anions like chloride, which convert to pH, stimulate
ATPase-driven uptake (43
, 50
51
52)
. Both observations imply
that at least two protons are exchanged per molecule of ACh. The
distinctive macroscopic kinetics reviewed in the Background section
(Km << KD,
substrate-independent Vmax, and
similar pKa values) imply that the
microscopic kinetic mechanisms for VAChT and VMAT are very similar to
each other. Thus, all direct and indirect evidence supports a
two-proton exchange mechanism for VAChT.
|
Characteristics of the microscopic constants
Estimated values of the microscopic constants for VAChT are listed
in Table 1
. Their important characteristics are 1) the
k2 step is rate-limiting,
2) AChi binds weakly, and
3) pKa1 controlling the
k1 step is lower than
pKa2 controlling the
k2 step. The value of
pKa1 is low enough that it would not
have been observed in previous experiments.
|
The mathematical relationships between macroscopic and microscopic
kinetic constants have been derived (16
, 43)
. They
simplify to Vmax
k2'Bmax,
Km
KDo'k2'/k1',
and KD
KDo', where
Bmax is the concentration of
transporter determined with saturating radiolabeled inhibitor and the
primes indicate that the constants are pH dependent. For example,
k2' = k210 -
pHi/(10 - pHi + 10
-pKa2), where 10 - pHi =
[Hi+] and 10
-pKa2 =
Ka2. The
k2' value estimated from the
parameters in Table 1
and a frequently reported
pHi value of 5.8 is 140
min-1. It should be similar to the steady-state
turnover rate
(Vmax/Bmax).
The turnover rate for human VAChT expressed in PC12 cells is 65
min-1 (29)
and the turnover rate
for bovine VMAT in chromaffin granules is 140
min-1 (40)
. The consistency of
values obtained by different methods supports their validity and the
microscopic analysis.
The VMAT transport cycle also has been modeled as in Fig. 2
(10
, 53)
, but the values for most of the corresponding microscopic
constants have not been estimated. They probably are similar to those
in Table 1
because of the extensive similarities in the macroscopic
kinetics of VMAT and VAChT. Also, Km for VMAT
increases when pHi increases (42)
.
This indicates that k1' decreases more
than k2' does at higher
pHi, which happens when
pKa1 <
pKa2. Thus, VMAT has the same rank
order in pKa1 and
pKa2 as VAChT does.
Computer simulation indicates that greater than 98% of the substrate
binding site in VAChT is outwardly oriented and distributed among the
top species in Fig. 2
when no proton gradient is present
(43)
. That is why macroscopic
KD is essentially equal to microscopic
KDo'. In contrast, the substrate
binding site in resting VMAT sometimes has been thought to be inwardly
oriented (25)
. However, the similarity in VAChT and VMAT
kinetics suggests that it, too, prefers the outward orientation. This
in turn implies that the energization-dependent increase in reserpine
affinity monitors translocation of the first proton as reserpine forms
a dead-end complex.
Microscopic constants are very important to structure–function
analysis
Because
k2'/k1'
is 0.01 to 0.1 (depending on animal species and pH), VAChT exists
mostly as Ti.Hi+ during
active transport with half-saturating ACho (Fig. 2)
. This pseudo-trapping of an intermediate derived from ACh-bound
transporter is what causes Km to be 10- to 100-fold
smaller than KD. Moreover, slow
k2 has the following additional
consequence. A mutation might decrease
k1 by 10-fold. This would increase
Km by 10-fold with no significant effect on
Vmax, even though the mutation has
nothing to do with the ACh binding site! Perturbation of
Km by a mutation does not necessarily mean that the
ACh binding site has been affected. VMAT probably exhibits the same
type of behavior. Some published data suffer this ambiguity.
The predicted macroscopic consequences of a change in the value of each
microscopic constant are given in Table 2
. Functional mutants (below) can be analyzed using the table. The input
information is the effect of a mutation on 1) macroscopic
KD (estimated by substrate competition
against binding of subsaturating radiolabeled inhibitor in the absence
of ATP), 2) macroscopic Km and
Vmax (estimated from initial
velocity steady-state transport kinetics using radiolabeled substrate,
with Vmax normalized to the amount of
mutant by Bmax or Western blot), and
3) pKa2 (which is both
macroscopic and microscopic and estimated from pH effects on binding of
subsaturating radiolabeled inhibitor in the absence of ATP). The output
deduction is the microscopic step affected by the mutation. For
example, a similar fold decrease in Km and
Vmax with no change in
KD and
pKa2 likely is due to the same fold
decrease in k2. All possible
combinations of mutant effects lead to a unique microscopic deduction,
except when Km alone changes. This result leaves
k1 and
pKa1 entangled, but pH jump studies
can separate them. Microscopic analysis of mutant effects is important
if specific amino acids are to be reliably correlated with specific
functions.
|
Microscopic analysis provides another important caution. If a mutation moderately affecting k2 was screened for transport using a subsaturating concentration of radiolabeled substrate, the mutation would appear to be a null due to cancellation of effects on Vmax and Km (above). Some published data suffer this ambiguity. Screening of mutants should be done with substrate concentrations that are low and high relative to the wild-type Km to avoid missing effects.
Thermodynamics constrains regulation of storage
Physiological observations demonstrating that storage of both ACh
and monoamines is regulated at the level of the secretory vesicle are
reviewed elsewhere (54
55
56
57
58)
. Fundamental thermodynamic
constraints to such regulation will be discussed here. The ACh
concentration in nerve terminal cytoplasm is low mM and consistent with
the Km value for vesicular transport. ACh in
synaptic vesicles is 100-fold more concentrated (7)
.
Because pH and values and the exchange stoichiometries are
the same for VAChT and VMAT, the driving forces for uptake are the
same. Thus, the vesicular gradient of ACh concentration in
vivo is 30-fold smaller than the driving force of 3000-fold. Why
might this be? One possible reason is that equilibrium would produce an
impossible, grossly hyperosmotic concentration of vesicular ACh (over 3
M). This limitation does not exist for VMAT, as the concentrations of
cytoplasmic monoamines are 100-fold lower. These observations lead to
the hypothesis that down-regulation terminates ACh storage well short
of theoretical equilibrium. The consequent excess in driving force
would buffer storage against a drop in the concentration of cytoplasmic
ACh and allow vesicles to fill at nearly the initial velocity
throughout the storage process (43)
.
Regulation can affect the rate and final amount of storage differently.
Because they make serotonin and histamine and express VMAT2, mast cells
store serotonin and histamine in the same secretory granules. Mutant
cells isolated from knockout heterozygotes expressing one-half of the
normal amount of VMAT2 and wild-type cells incubated in tetrabenazine
secrete (and thus store) substantially less serotonin but nearly normal
or increased amounts of histamine compared to untreated wild-type cells
(59)
. Yet the storage rates for serotonin and histamine
presumably were ‘regulated’ identically by the manipulations.
Different rates of neurotransmitter leakage explain the results.
Intragranular histamine leaks slowly due to protonation of the
imidazole ring, and the histamine gradient thus reaches equilibrium
with the driving force. As equilibria are not affected by changes in
catalyst activity, regulation of VMAT2 has little effect on histamine
storage. In contrast, intragranular serotonin leaks rapidly, as do the
rest of the monoamines except for histamine (38
, 39
, 60)
.
The serotonin gradient thus reaches only a steady state that responds
to changes in the rate of uptake.
The results of the mast cell study have the following implications for ACh storage. The existence of vesicular regulation confirms the absence of equilibrium with the initial velocity driving force. As no evidence exists that intravesicular ACh leaks fast enough to limit storage to 1/30th of equilibrium, cholinergic vesicles probably regulate storage by a novel mechanism not found in monoaminergic vesicles.
Structure–function analysis
Because VAChT and VMAT have similar initial velocity transport
kinetics and probably have similar TMD packing, available
structure–function data for them will be merged in an effort to
construct a preliminary model for the chemical mechanism of transport.
The certitude of the result is subject to two reservations. Alteration
of transporter function by chemical reaction or mutation could arise
from propagated conformational change, and macroscopic kinetics can be
misleading when not interpreted microscopically (above). The one-letter
abbreviations for amino acids will be used to identify specific
residues.
Protein chemistry
Organomercurials and methylmethanethiosulfonate react with
Torpedo VAChT to reveal an ACh-protected cysteine that is
readily accessible after modification. Another reactive cysteine is not
protected by ACh. Both residues are protected by vesamicol, suggesting
that vesamicol induces widespread conformational change
(61)
. Identification of the reactive cysteines should help
localize the ACh binding site.
Vesamicol binding to Torpedo VAChT is inhibited rapidly by
histidine-selective diethylpyrocarbonate, and the rate of the reaction
is slowed by vesamicol but not ACh (45)
. The conformation
of the reactive histidine was proposed to change on protonation.
Photoaffinity labeling
Photoaffinity analogs of ketanserin and tetrabenazine covalently
label the NH2 terminus of VMAT just before TMD I
and in loop X/XI (62
, 63)
. The results suggest that these
regions of the sequence are close to each other in inhibited VMAT.
Mutations affecting apparent ligand binding or energy coupling
Recombinant mutants of VAChT and VMAT have been expressed and
characterized primarily in vacuolar membranes of mammalian cell lines.
Only those mutants exhibiting perturbations of function larger than
errors and that inform binding and transport mechanisms will be
discussed. A K residue in TMD II and four D residues in TMDs I, VI, X,
and XI are conserved in all wild-type VAChTs and VMATs (Fig. 1)
. Their
presence in hydrophobic environments suggests that these normally
charged residues have special functions and thus have been subjected to
special scrutiny.
Chimeras
VMAT1 exhibits higher Km values for monoamines,
particularly histamine, and a higher
KD value for tetrabenazine than VMAT2
does (64)
. Km and
KD values for chimeras of VMAT1 and
VMAT2 suggest that multiple regions affect apparent affinities
(65
, 66)
. A chimera composed of human VMAT2 from the
NH2 terminus up to the beginning of TMD II and
human VAChT in the rest of the sequence exhibited sevenfold higher
Km for ACh but retained normal vesamicol binding.
Thus, part of the ACh binding site might be in a conserved
region before TMD II, namely, in TMD I, and the vesamicol binding site
lies beyond the beginning of TMD II (67)
.
Single-site substitutions
Perturbing mutations that express are summarized in Table 3
. Most of the mutations affecting apparent substrate binding lie in the
regions of TMDs I-III, XI, and XII. The exceptions are P in TMD V and A
and K in loop VII/VIII. Many of the mutations in VMAT affect
Km values (or equivalently the transport
IC50 values) for some substrates and not others.
The results suggest different binding sites for different VMAT
substrates. If a mutant exhibits an unchanged Km,
the values of the underlying microscopic constants probably are
unchanged. If the same mutant exhibits changed Km
for a different substrate, the KDo
value (and thus the binding site) for that substrate probably is
changed, as a mutation affecting a microscopic rate constant or
pKa value likely would affect
Km for all substrates.
|
Mutations affecting inhibitor binding lie throughout TMD IV to TMD XII. Although their locations are not well determined, the vesamicol and ACh binding sites clearly have been resolved from each other. Most discussion of substrate and inhibitor binding sites is deferred to a later section.
D in TMD X is implicated as the translocation site for the first proton
(Fig. 1)
. The S and C mutations in VMAT destroy transport and coupling
of reserpine binding to energization. The E mutation transports, and it
is the only mutation in VAChT and VMAT for which pH effects have been
determined. The results illustrate the power of pH studies to reveal
important features of mutant behavior not otherwise observable. The pH
activity curve for transport of subsaturating monoamine is very steep
in the pH 7.3–7.6 range, which means that deprotonation of
two external sites with pKa
values in this pH range is required for transport. The second, newly
observed site (in addition to the wild-type site of
pKa 7.1–7.4) must be either
unimportant or not present in wild type. A simple explanation is that
the mutant E residue itself accounts for the new
pKa and the
pKa for wild-type D is cryptic at < 6.5. The E mutation also perturbs the
pKa for required external protonation,
with the new value being 8. Whether the mutation decreases the
pKa of the group in wild type that
is 9 or unmasks a different group is unknown (46)
.
D in TMD X of VAChT has similar properties (30)
. The
residue might be modified by dicyclohexylcarbodiimide (75
, 76)
.
A dispensable ion pair between K in TMD II and D in TMD XI is present
in VMAT, because double mutants substituting another ion pair or charge
removal can transport whereas mutants retaining only one of the charges
cannot (46
, 69)
. In VAChT, a positively charged residue is
required at the position of K in TMD II for transport, consistent with
an ion pair, but charge-removing and charge-exchanged double mutants do
not transport (30
, 70)
. This behavior is consistent with
but does not require a proton translocation role for D in TMD XI of
VAChT. However, if proton translocation sites are conserved between
VAChT and VMAT, then this residue cannot be one (see below). Possibly
the K-D ion pair is more critical to structure in VAChT than it is in
VMAT. The observations indicate that TMDs II and XI contact each other.
D or E at the position of D in TMD I of VMAT is critical to serotonin
binding, and the E mutant transports subsaturating serotonin about
one-third as well as wild type does. The transport-incompetent N mutant
retains coupling of reserpine binding to energization, indicating that
translocation of the first proton is intact (68)
. The
residue is not critical to transport by VAChT, which is consistent with
binding of serotonin and ACh to separate sites (30)
.
Again, if proton translocation sites are conserved between VAChT and
VMAT, then this residue cannot be one (see below). All other aspartates
in cytoplasmic domains and TMDs are individually not critical to
transport, including completely conserved D in TMD VI (30
, 69)
.
An H in loop X/XI is the only completely conserved H in VMAT (Fig. 1)
.
Mutation to R or C destroys transport and coupling of reserpine binding
to energization without affecting low-affinity reserpine binding
(74)
. The latter observation indicates that the tested
mutations disrupt translocation of the first proton. Modification of
histidine or arginine groups in VMAT by classical protein chemistry has
similar effects (77)
, and a conserved R also is in loop
X/XI (Fig. 1)
. However, mutation of this H in rVAChT does not affect
transport, although it affects vesamicol binding (70)
. The
residue in Torpedo VAChT is Y. If it is conserved between
VAChT and VMAT, the other proton translocation site cannot be this
residue either (see below).
An H in TMD VIII is the only completely conserved H in VAChT (Fig. 1)
.
Mutation to A or K does not block transport (although C or R does) and
a conserved Y is present in VMAT (70)
. This H cannot be a
proton translocation site per se. It might form an ion pair
with D in TMD X, because the double mutation reversing these residues
restores vesamicol binding lost in a double mutation removing the
charges. It has been speculated to be the site of
pKa 7.1 in VAChT (45)
.
None of the remaining histidines in rVAChT apparently affect
transport (70)
.
Vectorial relationships in the major facilitator superfamily
VAChT and VMAT are members of the major facilitator superfamily
defined by sequence homology (78)
. The superfamily
includes proton antiporters and symporters. How can small changes in
amino acid sequence produce apparently fundamental change in the
vectorial relationship of protons and substrate? A transporter
architecture containing 1) two truncated channels leading to
opposite sides of the membrane and 2) a rotating domain that
binds and moves a proton and substrate molecule between the channels
can explain this (79)
. Binding of the proton and substrate
on the opposite or same sides of the rotating domain generates antiport
and symport, respectively (Fig. 3
).
|
Mapping VAChT and VMAT into the lactose permease paradigm
Lactose permease from Escherichia coli is the most
studied member of the major facilitator superfamily. It contains 12
helical TMDs for which packing is substantially described from analysis
of near neighbors. Most of the TMDs pack circumferentially around TMD
VII to form a central transport region (80)
. Can lactose
permease inform structure–function analysis of VAChT and VMAT, even
though it is a proton symporter? Alignment of VAChT and VMAT sequences
with the lactose permease sequence by the BLAST computer program
reveals homology (26% identity in the best case) from the beginning of
TMD II to the center of TMD V. Is additional structural and mechanistic
homology also present? To address this question, the assumed TMDs of
VAChT and VMAT were packed similarly to the current model of lactose
permease, except that helix tilt was ignored for clarity (Fig. 4
). Tilt brings some helices closer together than apparent in Fig. 4
, for
example, VIII and X (81)
and II and XI (82)
.
Use of lactose permease as a folding template is consistent with the
principle that secondary and super-secondary structures in a protein
family are more highly conserved than primary sequence is.
|
Only six residues in lactose permease are irreplaceable for transport,
and they all occur in TMDs. They are E126 (TMD
IV) and R144 (TMD V) in the lactose binding site;
E269 (TMD VIII), R302 (TMD
IX) and H322 (TMD X) probably in a net-neutral
triad; and E325 (TMD X), which has
pKa > 10, possibly because it faces
the low dielectric interior of the membrane. To begin a transport
cycle, a periplasmic proton binds to E325.
Periplasmic lactose then binds. A conformational change propagated from
the occupied lactose binding site disrupts the triad and releases
R302. This in turn promotes rotation of TMD X so
that E325 can form an ion pair with
R302. The isomerized ternary complex releases
lactose and the proton to cytoplasm to complete the first half of the
transport cycle. Empty transporter spontaneously relaxes back to the
original state to finish the cycle (83)
.
Although the amount of data allowing comparison of TMD packing in VAChT
and VMAT with that in lactose permease is limited, the sequence
homology and ion-pairing inferences provide evidence spanning most of
the TMDs (namely, TMDs II-V, VIII, X, and XI) that the packing patterns
are similar. No evidence conflicts. Thus, single-site mutations
affecting the properties of VAChT and VMAT are mapped into Fig. 4
assuming that protons, substrates, and inhibitors bind to a central
transport region. An interesting picture emerges that suggests lactose
permease and the rotating-domain model can inform analyses of VAChT and
VMAT at coarse levels of structure and function.
Components in the chemical mechanism of transport
E325 and D in TMD X are located similarly,
and they each translocate a proton to cytoplasm. If it acts analogously
to E325, D in TMD X translocates the first proton
in the VAChT and VMAT transport cycles. The
pKa of the site is 4.7 (Table 1)
,
which is nearly normal and consistent with the deduction in the
Background that the pKa is cryptic in
standard transport assays. The large difference in
pKa values for
E325 and D in TMD X does not imply different
roles for the residues, as pKa can be
altered by local environment without altering fundamental mechanism.
TMDs I-III, XI, and XII and the adjacent loops probably contain binding
sites for ACh and each monoamine that are separate from but close to
each other, as most of the mutations affecting apparent substrate
binding occur in this region (Fig. 4
and above). Because most of these
mutations are in the half of the putative architecture proximal to
lumen, substrate binding probably occurs deep in the transport channel.
Also, the putative neurotransmitter binding sites lie largely away from
the lactose binding site toward the opposite side of TMD VII. Such a
lateral change in location of substrate binding coincident with a
switch from symport to antiport evokes the rotating-domain model. If
TMDs VII and X act in concert as a rotating domain, a lactose
permease-like conformational change controlled by protonation of D in
TMD X might mediate the first translocation step in the VAChT and VMAT
transport cycles. The values of the rate constants for the
‘k1’ steps in VAChT and lactose
permease are similar (84)
.
VAChT and VMAT must use microscopic steps that are different from those
of lactose permease in the second half of the transport cycle to
translocate another proton. The final steps must be difficult, as VAChT
turns over 10-fold more slowly than lactose permease does
(84)
. No candidate for another proton translocation site
conserved between VAChT and VMAT has been identified. Possibly the
second site is not conserved.
As discussed above, H in loop X/XI and D in TMD I of VMAT as well as D
in TMD XI of VAChT each have an attribute expected of a proton
translocation site. However, they also suffer problems in this role.
The H in VMAT appears to be linked to translocation of the first
proton, but this proton was assigned to D in TMD X and the assignment
seems well supported. Moreover, the residue probably is not in contact
with luminal protons (Fig. 4)
. Possibly it modulates and does not
mediate proton translocation. All three of the residues have homologues
apparently not important to function in the other transporter type.
This odd behavior implies that these residues have undiscovered
differentiating interactions. As no well-supported candidate for the
second proton translocation site has been identified for either VAChT
or VMAT, discussion of chemical mechanism for the remainder of the
transport cycle is premature.
Finally, sites linked to binding of vesamicol, ketanserin, and
tetrabenazine by photoaffinity labeling or mutation are distributed
widely (Fig. 4
and above). The dispersal suggests that binding of
inhibitors causes widespread conformational change. Possibly binding of
protons to the site of pKa 7.1–7.4
causes similar widespread conformational change.
|
MATTERS OF CONTROVERSY AND UNANSWERED QUESTIONS |
|---|
|
TOPABSTRACTBACKGROUNDRECENT RESULTSMATTERS OF CONTROVERSY AND...PROSPECTS AND PREDICTIONSREFERENCES |
|---|
The largest unknowns surely are the tertiary structures of VAChT and VMAT. Unfortunately, even a low-resolution structure is not forthcoming. In the absence of one, it would be useful to test further whether 12 TMDs indeed are present and packed similarly to those in lactose permease. Until more substantial structural information is available, the components of chemical mechanism hypothesized here must be regarded as speculative. The comparison of VAChT and VMAT with lactose permease is heuristic. It is intended to aid conceptual organization of a large body of data, and it will change as more data accumulate.
Evidence for regulation in the storage of classical neurotransmitters is accumulating. However, the important difference between regulation of rates and final amounts, and the role that thermodynamics plays in regulation, have not been sufficiently recognized and characterized. The hypothesis that the final amount of ACh storage is regulated by a novel mechanism appears robust because it is based on large discrepancies with monoamine storage and the driving force. However, the existence of regulation in ACh storage is controversial because only a few workers have obtained direct evidence consistent with the idea. New physiological tests will be welcomed.
Progress in understanding the VAChT and VMAT transport mechanisms would be facilitated by microscopic interpretation of the effects of mutations and many more pH studies. Particularly needed are assignments of observed pKa values to specific residues acting in specific steps and identification of the second proton translocation site. Also needed are better localization of substrate binding sites, descriptions of the conformational steps returning VAChT and VMAT to the beginning of their respective transport cycles, and comprehensive models for regulation of VAChT and VMAT activities. A lot of work remains!
|
PROSPECTS AND PREDICTIONS |
|---|
|
TOPABSTRACTBACKGROUNDRECENT RESULTSMATTERS OF CONTROVERSY AND...PROSPECTS AND PREDICTIONSREFERENCES |
|---|
pH or by the amount required to explain a 30-fold storage
shortfall. The mechanistic and regulatory insights gained could lead to
new treatments for neurological disorders by control of
neurotransmitter storage.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTBACKGROUNDRECENT RESULTSMATTERS OF CONTROVERSY AND...PROSPECTS AND PREDICTIONSREFERENCES |
|---|
