Abstract
Zinc is an essential metabolic heavy metal in all cells. Zinc imbalances in the brain are
considered a cause of neurodegenerative disorders. A member of the TRPM cation channel
family, TRPM7 is a widely-expressed outwardly-rectifying Ca2 and Mg2 channel through
which trace metal ions like Zn“ flow into cells. TRPM7 Ca2 permeation may be essential for
cell viability. While previous characterizations of Zn* permeation involved mM levels of Zinc,
here we show that TRPM7 is the dominant Zn“ pathway in HEK-293 cells at 100 uM Zn2. We
performed confocal microscope imaging experiments with the Zn“-binding fluorophore
FluoZin-3 AM. In the presence of 100 uM Zn“, HEK-293 cells overexpressing murine TRPM7
load Zn“ almost 20% more than normal HEK-293 cells. Data for experiments in the presence of
physiological levels of Ca“ and Mg“ were compared to data for experiments performed in the
absence of Ca“ and Mg“, indicating that the presence of 2 mM Ca2 and 2 mM Mg2 inhibited
Zn? influx.
Introduction
The TRP - transient receptor potential - family consists of ion channel subunits
comprised of six membrane-spanning domains (Clapham et al., 2001). All eukaryotic cells
express some member(s) of the TRP family. The mammalian TRP channels are classified into
three groups: C - short, V - vanilloid, and M - melastatin. Of the TRPM family of cation
channels, TRPM6 and TRPM7 are unique in that they both contain kinase domains. The
function of the TRPM7 kinase domain is not well understood, and this domain's activity has
been characterized only very recently (Ryazanova et al., 2004), though the kinase is known not
to be required for TRPM7 channel gating (Schmitz et al., 2003). The TRPM7 channel is
inactivated by phospholipase C (PLC) activity and resultant phosphatidylinositol 4,5¬
bisphosphate (PIP») hydrolysis (Runnels et al., 2002) and is activated by low intracellular
magnesium (Nadler et al., 2001). The channel conducts divalent cations inwardly and
monovalent cations outwardly and is outwardly-rectifying (Runnels et al., 2001). TRPM7 is
widely expressed in the human body, and likely has a central role in Mg“ homeostasis (Schmitz
et al., 2003). TRPM7 Ca“ may be essential for cell viability (Nadler et al., 2001). In addition to
normal physiological transport of Ca* and Mg*, TRPM7 acts as a trace and nonphysiological
metal divalent cation influx pathway, exhibiting especially high permeation of Zn“ (Monteilh-
Zoller et al., 2003), albeit at concentrations of zinc 1000x above physiological levels (Perveen et
al., 2002).
Zinc is involved in a number of normal physiological functions, ranging from synaptic
release during brain function (Zatta et al., 2003) to immune response (Shankar and Prasad,
1998). However, excessive levels of zinc in neurons can be toxic via nitric oxide-implicated
neuron cell death (Bossy-Wetzel et al., 2004) and elevated glutamate excitotoxicity (Siddiq and
Page 1 of 8
Tsirka, 2003). Zinc is also implicated in the pathogenesis of neurodegenerative conditions like
Alzheimer’s disease (Koh et al., 1996). In the case of the role of zinc in Guamanian amyotrophic
lateral sclerosis (ALS) and parkinsonism-dementia complex (PDC), conflicting results exist.
Significantly decreased Zn“ in frontal cortex gray matter of both ALS and PDC cases was found
in one study (Yasui et al., 1993), while significantly increased Zn“ in gray matter of PDC cases
was found in another (Gellein et al., 2003). It has been suggested that some Guamanians express
mutant forms of TRPM7 (Hermosura, unpublished). In any case, zinc imbalance in the brain is
likely a cause of Guamanian ALS and PDC, and since TRPM7 is a mechanism for zinc influx,
we decided to examine Zn“ movement through TRPMT in vivo with a focus on interactions with
the main physiological divalents, Mg? and Ca?t.
Materials & Methods
Cells and cell culture - Experiments were conducted using HEK-293 cells, as designated by the
American Type Culture Collection (Rockville, MD). Cells were transfected with murine
TRPM7 (Monteilh-Zoller et al., 2003). Cells were grown at 37 °C in T,g tissue culture flasks
containing a mixture of DMD-F12, 10% FBS, 0.1% penicillin, and 0.1% streptomycin in an
atmosphere of 5% CO», 95% O, as per the ClonTech (Palo Alto, CA) tetracycline induction
system. Cells were on a 2-day passage/plating schedule, and were plated onto circular
polylycine-coated glass chips (BLAH in diameter) in 35 mm Petri dishes. Cells were induced
with 1.5 mL tetracycline after approximately 24 hours of growth on chips.
Solutions - Four stocks were produced ahead of time in distilled/deionized water, pH balanced to
7.2, and stored at 4 °C (Table 1). In order to make 100 uM Zn“ solutions, 10 mM ZnCl, was
produced the day of experiments, and 500 uL of this solution was added to 49.5 mL of the
Page 2 of 8
appropriate buffer. Approximately 0.3 g of glucose was added to each 50 mL tube of solution.
In order to make 10 uM pyrithione solutions, 5 uL of 10 mM pyrithione was mixed with 5 mL of
the appropriate buffer (+ 100 uM Zn**) immediately before use. Pyrithione is a selective Zn’
ionophore (Sensi et al., 1997). All solutions were heated to 35 °C.
Dye loading - Chips were placed in clean 35 mM Petri dishes containing -1.5 mL of warm non-
zinc test buffer. 5 uL FluoZin-3 AM (Molecular Probes, Eugene, OR) at 5 uM dilution in
DMSO was added. Dishes were rotated for 40 minutes. Chips were then placed in fresh Petri
dishes containing -1.5 mL (fresh) of the same buffer and rotated for 5 minutes. Chips not being
used were kept in the dark.
Imaging via confocal microscopy - Identical Plexiglas chambers were designed and milled. A
chip was attached with high vacuum grease to the bottom of a chamber containing the starting
solution. A f2 cover slip was greased in place. The chamber was placed on the confocal stage,
which was controlled at 30.0 °C. An Olympus (Melville, NY) UPlanFL 20X lens was used.
Cells were selected based on the following criteria: not too many processes, relatively round in
shape, separated from confluent masses, and if possible, not touching any other cells. Zoom in
FluoView Physiology/TIEMPO/Beta software (Olympus) was 2X. Laser intensity was set to
1.5, and confocal aperture was set to 2 on the Olympus FluoView 300 microscope. For all
experiments, 75 frames were captured at 2.8 sec/frame, at a resolution of 512x512. For frames
0-10, nothing was perfused (baseline). At frame -6, suction was switched on. At frame 10,
either 4 mL warm test solution, or 5 mL pyrithione mix was pipetted by hand onto chamber.
This was usually finished around frames 24-28. Suction was switched off, and recording
Page 3 of 8
continued until 75 frames. Between each experiment, image was refocused. All experiments
took place within 3 hours of start of dye loading procedure.
Experimental design - For each of the four Ca“Mg“ buffer types, induced and uninduced
(control) chips were tested. Three tests were performed on each chip. First, the same non-zinc
solution was perfused. This was a negative control. Second, 100 uM Zn* was added while
keeping [Ca“ and [Mg* constant. Then, 10 uM pyrithione mixed with the 100 uM Zn2
solution was added in order to determine the cells' maximum fluorescence (Fyay).
Data analysis - For each chip, 6 cells were chosen from the visual field that exemplified the
criteria stated earlier when selecting cells for imaging, were clearly visible, and did not move
during perfusion. In Fluo View, distinct areas were drawn around each cell, resulting in a table of
fluorescence values for each cell. The same cells that were used in a zinc addition analysis were
used in the subsequent pyrithione addition analysis. An Excel (Microsoft, Redmond, WA)
macro was created to prepare the data for analysis with Igor Pro (v. 4.09A Carbon)
(WaveMetrics, Lake Öswego, OR). In Igor, the average of the fluorescence values for the last 5
frames of the pyrithione experiment for each cell was used as each cell's Fyax. The fluorescence
over time during the addition of zinc for each cell was divided by the Fyxx, to produce 6 AF/F
graphs. The average of the first 10 values of each of the 6 graphs was calculated; this average
represented the baseline for each cell, since nothing was perfused over the first 10 graphs in any
experiment. The AF/F graphs were divided by their respective averages, and these normalized
graphs were averaged to produce a single graph. Thus, for each chip, the average represents an n
of 6.
Page 4 of 8
Results
Figures 1, 2, and 3 show the raw imaging data, with an artificial lookup table applied.
The results for 2 mM Ca“,0 mM Mg2“, O uM Zn2 (hereafter indicated as 2,0,0) to 2 mM
Ca“, 0 mM Mg“, 100 uM Zn* (hereafter indicated as 2,0,100) and the results for 0,2,0 to
0,2,100 are considered preliminary; in order to balance the images, an unknown nonlinear
conversion (photomultiplier tube voltage change) was applied between the zinc addition
experiment and the pyrithione addition experiment. A different nonlinear PMT voltage change
was used for the induced versus uninduced. This nonlinearity may explain the uncharacteristic
downturn at the ends of the 0,2,0 — 0,2,100 graphs for induced and uninduced (Fig. 5) or the
downturn in the uninduced graph for 2,0,0 - 2,0,100 (Fig. 4). In any case, induced cells
increased in fluorescence more than uninduced cells, which indicates more loading of zinc,
though the nonlinear conversion prevents us from knowing by how much. The nonlinear
conversion also prevents comparison of the 0,2 and 2,0 experiments with each other or any other
experiments, so it is impossible to determine the specific differences in effect between a lack of
Mg“ and a lack of Ca“ on zinc loading.
The 2,2,0 - 2,2,100 (Fig. 6) and 0,0,0 - 0,0,100 (Fig. 7) experimental results are more
useful; because no nonlinear conversion was applied, uninduced data can be compared to
induced data, and the experiments can be compared to each other. Again, induced cells clearly
fluoresced more upon addition of zinc than uninduced cells, and cells in 0,0 fluoresced upon
addition of zinc more than cells in 2,2. Uninduced cells in 0,0 also showed a greater increase in
fluorescence than uninduced cells in 2,2. The average maximum fluorescence value for 2,2 is
1.35 (35% above baseline), in contrast to the max for 0,0, which is 1.53 (53% above baseline).
Page 5 of 8
Discussion
As mentioned before, the results for experiments where only magnesium was present or
where only calcium was present (0,2 and 2,0) are considered preliminary. Both ions,
respectively, seemed to have inhibitory effects on zinc loading, and further experiments should
reveal the relative magnitude of each ion's effect.
The 2,2 and 0,0 data are more useful, especially because induced and uninduced cells
from both buffer types can be compared. Also, experiments were performed at almost the exact
same time after dye loading, so the higher fluorescence increases seen in the 0,0 data are not
likely due to additional dye loading and de-esterification/activation over time. Induced cells
fluoresced an average of almost 20% more in the absence of Mg“ and Ca" than in the presence
of the two ions. A difference in fluorescence levels after Zn“ addition is seen between
uninduced cells in 0,0 and uninduced cells in 2,2 because HEK-293 cells express normal levels
of TRPM7, and because TRPM7 is not the only Zn* influx pathway. Since the only difference
between the 0,0 and 2,2 experiments is in the buffer ion concentrations, and since overexpression
of TRPM7 is the only difference between induced cells and uninduced cells (Monteilh-Zoller et
al., 2003), Ca“ and Mg“ must be having an effect on Zn* movement through TRPM7.
One explanation for the difference between 0,0 and 2,2 data is that the additional
divalents competitively interfered with Zn“ binding and transport across and through the
TRPM7 channel. This may have occurred, though Zn“ has been found to have high
permeability relative to Mg2t and even higher permeability relative to Ca“ through TRPM7
(Monteilh-Zoller et al., 2003). The Monteilh-Zoller et al. findings in turn may not be applicable
to our study, since 10 mM Zn“ was used in that experiment, whereas we used 100 uM Zn“. A
more likely explanation is that TRPM7 is regulated by intracellular Mg-ATP (Nadler et al., 2001)
Page 6 of 8
or even free Mg“ (Kozak and Cahalan, 2003). Cells bathing for up to -2.5 hours in 2 mM Mg?
would likely have higher intracellular levels of Mg-ATP or free Mg* than cells bathing in 0 mM
Mg“, and would thus show an increased inhibition of TRPM7 activity.
However, based on the preliminary 2,0 data (shown) and other preliminary 2,0
experiments (data not shown), it seems that Ca“ alone does have a negative effect on Zn’
permeation. We speculate that intracellular Ca“ may promote PLC activity and resultant
TRPM7 inhibition via PIP, hydrolysis (Runnels et al., 2002), or that extracellular Ca’t may
significantly interfere with zinc binding to TRPM7 at 100 uM Zn2. Besides collecting more
meaningful data from better 2,0 experiments, in future 0,2 and 0,0 tests we will also loaded cells
with BAPTA AM, a strong Ca“ chelator. Comparing a non-BAPTA 0,2 experiment with a
BAPTA 0,2 experiment will further elucidate calcium’s effects on Zn’ TRPM7 permeation,
since even in 0 mM Ca“ there is some intracellular Ca2t.
Also, the only missing piece of information necessary to calculate the actual ion
concentration at any given time is the Fyny. This value will be determined by treating cells with
TPEN (Sensi et al., 1997). Finally, since the [Zn’ we used was 100 uM, and since
physiological levels are closer to 10 uM (Perveen et al., 2002), it will also be important to
examine 10 uM Zn“ permeation.
Conclusions
This study shows that cells overexpressing TRPM7 load Zn’ more than cells with normal
expression of TRPM7, supporting the claim that TRPM7 is the dominant pathway for Zn“
movement into cells. This study also shows that Ca“ and Mg“ inhibit Zn“ movement through
TRPM7. Further experiments will be necessary to determine just how effective Ca“ and Mg-
are at inhibition of Zn“ permeation at physiological concentrations. Ultimately, thorough
Page 7 of 8
investigation of zinc permeation through TRPM7 may help us understand how zinc levels in the
brain become imbalanced, especially once stable transfects of normal human TRPM7 and
Guamanian mutant TRPM7 exist. Establishing the relative magnitudes of physiological divalent
effects on metal permeation will help elucidate the mechanism of TRPM7 trace metal transport.
Acknowledgments
I thank Stuart Thompson for guidance and advice. I also thank Christian Reilly for random acts
of help, Jenny Ta and Bonita Song for their assistance, and John Lee for machine shop
assistance.
Page 8 of 8
Tables
Table 1. Stock solution components.
Solution
KCI
Nacl
140.0
2,2
2.8
145.0
0,0
2.8
2.8
145.0
0,2
2,0
145.0
2.8
final mM
Hepes Naoh
10.0
10.0
10.0
Caci2
2.0
0.0
2.0
Mgc12
2.0
2.0
0.0
Figures
Fig. 1. Induced cells in 2 mM Ca2, 2 mM Mg“,O uM Zn2t before addition of 100 uM Zn2 in 2
mM Ca’t and 2 mM Mg?.
Fig. 2. Cells from Fig. 1 approximately 3 minutes after addition of 100 uM Zn2t.
Fig. 3. Cells from Fig. 2 approximately 3 minutes after addition of 10 uM pyrithione in 100 uM
Zn?t.
Fig. 4. Average of 6 induced and 6 uninduced cells' responses to 100 uM Zn* in 2 mM Ca2 and
0 mMMg“. Cells started in 2 mM Ca“,0 mM Mg“, O uM Zn’. Results were normalized to
pyrithione-induced Fyxx of FluoZin-3 AM, as well as to baseline.
Fig. 5. Average of 6 induced and 6 uninduced cells' responses to 100 uM Zn* in 0 mM Ca* and
2 mMMg“. Cells started in 0 mM Ca“, 2 mM Mg“,O uM Zn2. Results were normalized to
pyrithione-induced Fygx of FluoZin-3 AM, as well as to baseline.
Fig. 6. Average of 6 induced and 6 uninduced cells' responses to 100 uM Zn2t in 2 mM Ca2t and
2 mM Mg“. Cells started in 2 mM Ca“, 2 mM Mg“, O uM Zn’. Results were normalized to
pyrithione-induced Fygx of FluoZin-3 AM, as well as to baseline.
Fig. 7. Average of 6 induced and 6 uninduced cells' responses to 100 uM Zn2 in 0 mM Ca* and
0 mMMg“. Cells started in 0 mM Ca“,0 mM Mg“, O uM Zn’. Results were normalized to
pyrithione-induced Fyxx of FluoZin-3 AM, as well as to baseline.
Fig. 1
Fig. 2
Fig. 3
Fig. 4
1.6-

inducec
1.4-

1.2-
v


uninduced
110-
0.8-
0.6 -
0.4-

0 20 40 60 80 100 120 140 160 180 200
seconds
Fig. 5
1.6-
induced



1.4-

1.2-
uninduced
10 a
0.8-
0.6-
0.4-
0 20 40 60 80
100
120 140 160 180 200
seconds
Fig. 6
1.6-
1.4-
induced

1.2-

f

1 1og
uninduced
W
0.8-
0.6-
0.4-

vakaa-
0 20 40 60 80 100
120 140 160 180 200
seconds
Fig. 7
1.6-
aes
1.4-


1.2-

1opsav

uninduced
0.8-
0.6 -
0.4-

0 20 40 60 80
100
120 140 160 180 200
seconds
References
Bossy-Wetzel, E., M.V. Talantova, W.D. Lee, M.N. Schölzke, A. Harrop, E. Mathews, T. Götz,
J. Han, M.H. Ellisman, G.A. Perkins, and S.A. Lipton. 2004. Crosstalk between nitric
oxide and zinc pathways to neuronal cell death involving mitochondrial dysfunction and
p38-activated K' channels. Neuron. 41:351-365.
Clapham, D.E., L.W. Runnels, and C. Strübing. 2001. The TRP ion family. Nature Reviews.
2:387-396.
Gellein, K., R.M. Garruto, T. Syversen, T.E. Sjøbakk, and T.P. Flaten. 2003. Concentrations of
Cd, Co, Cu, Fe, Mn, Rb, V, and Zn in formalin-fixed brain tissue in amyotrophic lateral
sclerosis and parkinsonism-dementia complex of Guam determined by high-resolution ICP-
MS. Biol. Trace Element Res. 96:39-60.
Koh, J.Y., S.W. Suh, B.J. Gwag, Y.Y. He, C.Y. Hsu, and D.W. Choi. 1996. The role of zinc in
selective neuronal death after transient global cerebral ischemia. Science. 272:1013-1016.
Kozak, J.A., and M.D. Cahalan. 2003. MIC channels are inhibited by internal divalent cations
but not ATP. Biophys. J. 84:922-927.
Monteilh-Zoller, M.K., M.C. Hermosura, M.J.S. Nadler, A.M. Scharenberg, R. Penner, and A.
Fleig. 2003. TRPM7 provides an ion channel mechanism for cellular entry of trace metal
ions. J. Gen. Physiol. 121:49-60.
Nadler, M.J., M.C. Hermosura, K. Inabe, A.L. Perraud, Q. Zhu, A.J. Stokes, T. Kurosaki, J.P.
Kinet, R. Penner, A.M. Scharenberg, and A. Fleig. 2001. LTRPC7 is a Mg-ATP-regulated
divalent cation channel required for cell viability. Nature. 411:590-595.
Perveen, S., W. Altaf, N. Vohra, M.L. Bautista, R.G. Harper, and R.A. Wapnir. 2002. Effect of
gestational age on cord blood plasma copper, zinc, magnesium and albumin. Early Hum.
Dev. 69:15-23.
Runnels, L.W., L. Yue, and D.E. Clapham. 2001. TRP-PLIK, a bifunctional protein with kinase
and ion channel activities. Science. 291:1043-1047.
Runnels, L.W., L. Yue, and D.E. Clapham. 2002. The TRPM7 channel is inactivated by PIP,
hydrolysis. Nat. Cell Biol. 4:329-336.
Ryazanova, L.V., M.V. Dorovkov, A. Ansari, and A.G. Ryazonov. 2004. J. Biol. Chem.
279:3708-3716.
Schmitz, C., A.L. Perraud, C.O. Johnson, K. Inabe, M.K. Smith, R. Penner, T. Kurosaki, A.
Fleig, and A.M. Scharenberg. 2003. Regulation of vertebrate cellular Mg“ homeostasis by
TRPM7. Cell. 114:191-200.
Sensi, S.L., L.M.T. Canzoniero, S.P. Yu, H.S. Ying, J.Y. Koh, G.A. Kerchner, and D.W. Choi.
1997. Measurement of intracellular free zinc in living cortical neurons: routes of entry. J.
Neurosci. 17:9554-9564.
Shankar, A.H., and A.S. Prasad. 1998. Zinc and immune function: the biological basis of
altered resistance to infection. Am. J. Clin. Nutr. 68:447S-63S.
Siddiq, M.M. and S.E. Tsirka. 2003. Modulation of zinc toxicity by tissue plasminogen
activator. Mol. Cell. Neurosci. 25:162-171.
Yasui, M., K. Ota, and R.M. Garruto. 1993. Concentrations of zinc and iron in the brains of
Guamanian patients with amyotrophic lateral sclerosis and parkinsonism-dementia.
Neurotoxicology. 14:445-50.
Zatta, P., R. Lucchini, S.J. van Rensburg, and A. Taylor. 2003. The role of metals in
neurodegenerative processes: aluminum, manganese, and zinc. Brain Res. Bul. 62:15-28.