bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
1 Proton co-transport and voltage dependence enforce unidirectional metal transport in an Nramp
2 transporter
3
4
5 Aaron T. Bozzi1, Lukas B. Bane1,2, Christina M. Zimanyi1,3, and Rachelle Gaudet1,*
6
7 1Department of Molecular and Cellular Biology, Harvard University, 52 Oxford Street, Cambridge,
8 MA 02138, USA
9 2Current address: D. E. Shaw Research, New York, NY 10036, USA
10 3Current address: New York Structural Biology Center, 89 Convent Ave., New York, NY 10027,
11 USA
12 *Corresponding author
1 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
13 Abstract
14 Secondary transporters harness electrochemical energy to move substrate through structurally-
15 enforced co-substrate “coupling”. We untangle the “proton-metal coupling” behavior by a Natural
16 resistance-associated macrophage protein (Nramp) transporter into two distinct phenomena: ΔpH
17 stimulation of metal transport and metal stimulation of proton co-transport. Surprisingly, metal
18 type dictates co-transport stoichiometry, leading to manganese-proton symport but cadmium
19 uniport. Additionally, the membrane potential affects both the kinetics and thermodynamics of
20 metal transport. A conserved salt-bridge network near the metal-binding site imparts voltage
21 dependence and enables proton co-transport, properties that allow this Nramp transporter to
22 maximize metal uptake and prevent deleterious back-transport of acquired metals. We provide a
23 new mechanistic model for Nramp metal-proton symport in which, in addition to substrate
24 gradients determining directionality as in canonical secondary transport, synergy between protein
25 structure and physiological voltage enforces unidirectional substrate movement. Our results
26 illustrate a functional advantage that arises from deviations from the traditional model of symport.
2 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
27 Cells commit significant ATP expenditure to maintain ionic gradients across membranes, a form
28 of energy storage. Selective membrane permeability also generates a stable charge imbalance that
29 produces a membrane voltage, generally negative. Secondary transporters harness these
30 electrochemical gradients by coupling the energetically-favorable movement of abundant ions
31 (Na+, K+, H+, Cl-) to the (often uphill) movement of substrates (Forrest et al., 2011; Gadsby, 2009;
32 LeVine et al., 2016; Shilton, 2015). Tight coupling mechanisms prevent uniport events: “futile
33 cycles” that dissipate the driving ion gradients and, even more deleterious, the back-transport of
34 the primary substrate down its concentration gradient (LeVine et al., 2016).
35 The Nramp (Natural resistance-associated macrophage protein) family of transporters import
36 divalent transition metal ions, essential micronutrients serving as cofactors to myriad metabolic
37 enzymes. Prokaryotic Nramps perform high-affinity Mn2+ scavenging (Cellier, 2012; Ma et al.,
38 2009), while in eukaryotes, including humans, Nramps are essential to iron dietary uptake, iron
39 trafficking to and recycling from erythrocytes, and the metal-withdrawal innate immune defense
40 (Abbaspour et al., 2014; Andrews, 2008; Coffey and Ganz, 2017).
41 Nramps are thought to couple proton transport to metal substrate import (Chen et al., 1999a;
42 Gunshin et al., 1997)—primarily Mn2+ and Fe2+, but also the biologically useful metals Co2+, Ni2+,
43 and Zn2+ (Ehrnstorfer et al., 2014; Illing et al., 2012; Picard et al., 2000); and toxic heavy metals
44 Cd2+, Pb2+, and Hg2+ (Bressler et al., 2004; Vazquez et al., 2015). Given typically low
45 environmental metal ion concentrations, this coupling may make the net transport
46 thermodynamically favorable. However, the substantial negative membrane potential (which
47 favors cation entry) and the plethora of intracellular metal-binding proteins and small molecules
48 (which likely act as a sink) may make metal entry alone energetically-downhill in many cases.
49 Early electrophysiological studies of Nramps showed variable metal-proton transport
3 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
50 stoichiometries depending on pH and membrane potential, with significant proton uniport,
51 suggesting loose coupling (Chen et al., 1999a; Mackenzie et al., 2006; Nelson et al., 2002).
52 Furthermore, multiple Nramp homologs exhibit voltage- and pH-dependence of transport (Chen
53 et al., 1999a; Gunshin et al., 1997; Mackenzie et al., 2007; Xu et al., 2004). Nramps have a
54 common secondary-transporter fold (Ehrnstorfer et al., 2014) and an alternating-access
55 conformational cycle (Bozzi et al., 2016b; Bozzi et al., 2018; Ehrnstorfer et al., 2017). Several
56 conserved residues are important for substrate selectivity (Bozzi et al., 2016a; Bozzi et al., 2018)
57 and proton-metal coupling (Ehrnstorfer et al., 2017; Pujol-Gimenez et al., 2017).
58 We use the Deinococcus radiodurans (Dra)Nramp homolog, for which we determined snapshots
59 of the structure in multiple conformations (Bozzi et al., 2016b; Bozzi et al., 2018), to demonstrate
60 how it harnesses the membrane potential to both accelerate metal import and prevent deleterious
61 metal export. We further demonstrate how the Nramp “proton-metal coupling” phenotype
62 encompasses two distinct phenomena: ΔpH stimulates metal transport, and metal stimulates proton
63 co-transport, with some metal substrates surprisingly undergoing uniport—metal transport without
64 accompanying proton(s)—rather than the expected symport. We show how a conserved salt-bridge
65 network allosterically imparts voltage dependence and forms the pathway for proton co-transport.
66 Our results mechanistically explain how an Nramp provides unidirectional transport for its
67 transition metal substrates through its many deviations from the canonical model for symport.
68
69 Results
70 The membrane potential drives DraNramp metal transport
71 Our investigation of the role of transmembrane voltage in DraNramp’s mechanism stems from
72 trying to reconcile an apparent inconsistency between in vivo and in vitro protein behavior. Purified
4 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
73 DraNramp reconstituted into proteoliposomes (Figure 1A and Figure 1—figure supplement 1A)
74 with large metal substrate gradients transported Cd2+—a toxic metal known to be a good Nramp
75 substrate (Courville et al., 2008; Illing et al., 2012)—but not the biological substrates Mn2+ or Fe2+,
76 or similar metals like Zn2+ or Co2+ (Figure 1B and Figure 1—figure supplement 1D). In contrast
77 DraNramp robustly transported Co2+ when expressed in Escherichia coli (Figure 1—figure
78 supplement 1B).
79 Bacteria, including E. coli, maintain a membrane potential (ΔΨ) between -140 and -220 mV (Bot
80 and Prodan, 2010), which greatly influences ion transport thermodynamics. When we applied
81 voltage across proteoliposome membranes (see Methods), DraNramp transported Mn2+ at -40 mV
82 or below (Figure 1B), and Fe2+, Zn2+, and Co2+ at -80 mV (Figure 1B and Figure 1—figure
83 supplement 1D), with notable acceleration at -120 mV (Figure 1B and Figure 1—figure
84 supplement 1D). Indeed, by measuring at 10 mV increments, we saw metal-specific threshold
85 voltages for metal transport with Mn2+, followed by Zn2+ and Fe2+, then Co2+ requiring
86 progressively larger potential differences to observe metal transport, while Cd2+ was transported
87 across the whole measured range (Figure 1C and Figure 1—figure supplement 1E). These trends
88 are similar to the voltage-dependence of mammalian Nramp2 Fe2+ transport observed via
89 electrophysiology (Gunshin et al., 1997; Mackenzie et al., 2006; Pujol-Gimenez et al., 2017;
90 Sacher et al., 2001).
91 We next measured transport across a range of metal concentrations at a physiologically-relevant
92 ΔΨ = -150 mV (Figure 1D) to determine Michaelis-Menten constants. The apparent substrate
2+ 93 affinity (KM) at -150 mV correlated with the requisite threshold voltage (Figure 1C): Mn and
94 Cd2+ had highest affinity and the lowest magnitude threshold ΔΨ, Co2+ had the lowest affinity and
2+ 2+ 95 highest magnitude threshold ΔΨ, and Zn and Fe were intermediate for both. Our KM values are
5 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
96 consistent with in vivo Co2+ uptake data (Figure 1—figure supplement 1F) and of similar
97 magnitude to previously reported values for in vivo Cd2+ and Mn2+ transport by the E. coli Nramp
98 homolog (Courville et al., 2008; Haemig and Brooker, 2004b). The much higher apparent affinity
2+ 2+ 99 (KM < 10 μM) for Mn and Cd than other divalent metal substrates is consistent with trends first
100 observed in E. coli Nramp (Makui et al., 2000), but contrasts with human Nramp2 (DMT1), which
101 has affinities for Mn2+, Fe2+, Co2+, Zn2+, and Cd2+ all within an order of magnitude in the low μM
102 range (Illing et al., 2012). These results suggest that these bacterial homologs have higher
103 specificity for their intended substrate, Mn2+, with incidental high Cd2+ affinity. Next, to
104 recapitulate the voltage-dependence trend in vivo, we varied external [K+], which likely perturbed
105 the membrane potential of E. coli, as they maintain a significant cytoplasmic [K+] (Shabala et al.,
106 2009). High [K+] indeed drastically decreased DraNramp Co2+ transport (Figure 1E and Figure
107 1—figure supplement 1G; see Figure 5—figure supplement 2 below for comparison of DraNramp
108 point mutants).
109 As one explanation for our results, the membrane potential might perturb DraNramp’s gross
110 conformational landscape, with the outward-facing state relatively destabilized at lower magnitude
111 voltages, as seen with the structurally-related SGLT1 transporter (Loo et al., 1998). We therefore
112 measured the accessibility to N-ethylmaleimide (NEM) modification of single-cysteine reporters
113 that are only exposed to bulk solvent in either the inward-open (A53C) or outward-open (A61C)
114 states (Bozzi et al., 2016b; Bozzi et al., 2018) (Figure 1F). However, although high extracellular
115 [K+] greatly impaired Co2+ transport, it had little effect on the sampling of either the inward-facing
116 or outward-facing states (Figure 1G and Figure 1—figure supplement 1H). We conclude that
117 DraNramp voltage dependence arises from more nuanced mechanistic changes. This is more
118 consistent with the observed substrate-specific threshold voltages, as a shift in conformational
6 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
119 equilibrium that completely inhibits transport of all metal does not account for this substrate
120 specificity.
121 Our consistent in vivo and in vitro findings indicate the reconstituted transporter does indeed
122 function the same as in the native membrane, and that transmembrane voltage—the missing piece
123 in our initial reconstitution attempts—is a critical variable controlling metal transport kinetics.
124
125 Proton co-transport is substrate-specific in DraNramp
126 Despite their similar KM and Vmax at physiological ΔΨ, Mn2+ and Cd2+ behave very differently
127 at voltages near zero. Indeed, when extending our liposome assay to values of ΔΨ > 0, which
128 thermodynamically disfavors cation entry, DraNramp still transported Cd2+—but not Mn2+—
129 down its concentration gradient even at +50 mV (Figure 2—figure supplement 1A-C), hinting at
130 a mechanistic difference between the two metals.
131 In addition to ΔΨ’s importance, the well-documented requirement of many Nramp homologs for
132 low pH to stimulate metal transport (Chen et al., 1999a; Gunshin et al., 1997; Makui et al., 2000)
133 suggests that ΔpH is also a critical variable for DraNramp. We therefore compared the pH
134 dependence of transport for both Cd2+ and Mn2+ (Figure 2A). While lower external pH moderately
135 enhanced Cd2+ transport (Figure 2B), it greatly accelerated Mn2+ transport—even enabling
136 transport at 0 mV (Figure 2C). Therefore Mn2+ transport—but not Cd2+—may require
137 thermodynamically favorable proton entry, either through a negative membrane potential or a pH
138 gradient.
139 To test this hypothesis, we measured H+ transport by DraNramp (Figure 2D) at a range of voltages
140 in the absence or presence of metal substrate (Figure 2E and Figure 2—figure supplement 1D).
141 While DraNramp efficiently transported all tested metals (except Ca2+) at -120 mV and to some
7 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
142 degree at -80 mV, only Mn2+, Fe2+, and Co2+ stimulated H+ influx above basal rates (Figure 2F).
143 Despite robust Cd2+ uptake across all potentials, Cd2+ failed to stimulate significant H+ transport.
144 Indeed at -120 mV, where H+ uniport—uptake in the absence of added metal—became significant,
145 Cd2+ noticeably reduced the H+ influx rate (Figure 2E-F). DraNramp’s two best substrates, which
146 should nearly saturate the metal-binding site at -80 and -120 mV, gave initial rate stoichiometries
147 of approximately 1 H+: 1 Mn2+ and 0 H+: 1 Cd2+ (Figure 2F), confirming that metal-proton co-
148 transport is indeed substrate-specific.
149
150 Voltage and pH both shape DraNramp’s free energy landscape
151 To better understand the effects of transmembrane voltage and pH gradients on DraNramp
152 transport, we determined Michaelis-Menten parameters for Mn2+ and Cd2+ transport kinetics in the
153 absence or presence of a pH gradient at three ΔΨ values (Figure 3A-B) and used our results to
154 sketch reaction coordinate diagrams for Cd2+ and Mn2+ transport (Figure 3C-D). For both metals
155 at both pHs, a more negative ΔΨ increased transport efficiency by improving both KM and Vmax.
156 Strikingly, in the absence of ΔpH, hyperpolarizing ΔΨ from -50 to -150 mV decreased the KMs
157 for Cd2+ and Mn2+ more than 20- and 300-fold respectively. Thus, in addition to its impact on ion
158 uptake thermodynamics, ΔΨ also shapes the free energy landscape that dictates DraNramp
159 transport kinetics (Figure 3C-D) and thus likely influences metal binding and other steps in the
160 transport mechanism.
161 While in all six cases, ΔpH increased Vmax, the effects on KM were mixed (Figure 3A-B). Indeed,
162 ΔpH enhanced apparent affinity for both metals at lower magnitude ΔΨ but reduced apparent
163 affinity at higher magnitude ΔΨ. Generally, ΔΨ and ΔpH are more synergistic for Mn2+ and more
164 antagonistic for Cd2+, consistent with the H+ co-transport requirement for Mn2+ but not Cd2+
8 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
165 (Figure 2F). In addition, our data indicate that an insurmountable kinetic barrier precludes even
166 downhill Mn2+—but not Cd2+—transport in the absence of a ΔΨ and/or ΔpH favorable to H+
167 transport (Figure 3C-D). Taken together with the fact that ΔpH affected the transport kinetics of
168 both metals, this indicates that “proton-metal coupling” in Nramps comprises at least two distinct
169 phenomena: (i) low pH/ΔpH stimulates metal transport, and (ii) H+ co-transport is required for
170 certain metal substrates. Below, using insights gleaned from DraNramp’s structures and
171 mutagenesis experiments we begin to mechanistically explain these surprising results.
172
173 DraNramp metal-binding site abuts an extended salt-bridge network
174 Our structures of DraNramp in multiple conformational states (Bozzi et al., 2016b; Bozzi et al.,
175 2018) provide a framework to understand the mechanistic details of voltage dependence, ΔpH
176 stimulation, and proton co-transport. DraNramp’s 11 α-helical transmembrane (TM) segments
177 adopt the common LeuT-fold (Boudker and Verdon, 2010; Forrest et al., 2011; Shi, 2013;
178 Yamashita et al., 2005) (Figure 4A). Conserved residues D56, N59, and M230 coordinate metal
179 substrate throughout the transport process (Bozzi et al., 2018; Ehrnstorfer et al., 2014) (Figure 4B).
180 Below the binding site, conserved H237 on TM6b lines the metal’s exit route to the cytoplasm
181 (Bozzi et al., 2018) (Figure 4C and Figure 4—figure supplement 1A). Adjacent to D56, a
182 conserved network of charged and protonatable residues lead into the structurally-rigid cluster
183 formed by TMs 3, 4, 8, and 9 to provide a parallel ~20 Å H+-transport pathway towards the
184 cytoplasm (Bozzi et al., 2018) (Figure 4C-D and Figure 4—figure supplement 1A-B). These
185 residues form a sequence of three interacting pairs: H232-E134, D131-R353, and R352-E124. This
186 salt-bridge network is unique to the Nramp clade in the LeuT family of structurally- and
187 evolutionarily-related transporters (Bozzi et al., 2018; Cellier, 2012; Cellier, 2016).
9 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
188 In multiple Nramp homologs D56 and N59 contribute to metal transport (Bozzi et al., 2016a; Bozzi
189 et al., 2018; Courville et al., 2008; Ehrnstorfer et al., 2014; Ehrnstorfer et al., 2017), while M230
190 contributes to a selectivity filter that favors transition metals over alkaline earth metals (Bozzi et
191 al., 2016a). Mutations to E124, D131, E134, H232, and H237 abrogate Mn2+ transport in E. coli
192 Nramp (Haemig and Brooker, 2004a; Haemig et al., 2010), while mutations to H232 and H237
193 (Lam-Yuk-Tseung et al., 2003; Mackenzie et al., 2006) and R353 (Lam-Yuk-Tseung et al., 2006)
194 impair metal transport in human DMT1, with the latter mutation also causing anemia (Iolascon et
195 al., 2006). Recent studies using different Nramp homologs implicated E134 or H232 in proton-
196 metal coupling (Ehrnstorfer et al., 2017; Pujol-Gimenez et al., 2017), and we demonstrated the
197 importance of D56 and D131 in addition to E134 and H232 to DraNramp H+-transport (Bozzi et
198 al., 2018). Based on structural analyses we therefore proposed D56 as the initial proton-binding
199 site and D131 as the subsequent proton acceptor required for transmembrane transport (Bozzi et
200 al., 2018).
201 Hypothesizing that this network of conserved charged and protonatable residues contributes to the
202 voltage-dependence and proton-metal coupling phenomena, we designed a panel of point mutants
203 that remove or neutralize these sidechains via alanine and/or asparagine/glutamine substitution.
204 All DraNramp mutants expressed well in E. coli (Figure 4—figure supplement 1C-D) and were
205 readily purified. We reconstituted each in proteoliposomes to further explore how each conserved
206 residue contributes to DraNramp’s metal and proton transport behavior.
207
208 Conserved salt-bridge network imparts observed voltage dependence of metal transport
209 With our mutant panel, we first measured the effect of ΔΨ on in vitro Cd2+ and Mn2+ transport
210 (Figure 5A and Figure 5—figure supplement 1). Most mutants retained some transport of either or
10 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
211 both substrates, except metal-binding D56 mutants. Compared to WT, the remaining mutants
212 cluster in two groups in terms of their ability to harness ΔΨ.
213 Mutations to N59, M230, H232, and H237 largely preserved WT-like voltage dependence. These
214 TM1 and TM6 residues cluster in the metal-binding site or metal-release pathway (Figure 5B).
215 Mutations to TM3 and TM9 salt-bridge network residues E124, R352, R353, D131, and E134
216 reduced voltage dependence of Mn2+ and Cd2+ transport across the tested ΔΨs (Figure 5). All
217 reduced transport rates at -120 mV, but many equaled and several outperformed WT at lower
218 magnitude ΔΨ. Perturbing this network of residues thus alters ΔΨ’s effects on DraNramp’s metal
219 transport kinetics, such that these mutants, while moderately impaired compared to WT under
220 physiological conditions, no longer face WT’s transport restriction in the absence of ΔΨ.
221 We saw similar trends with this panel in vivo: salt-bridge network mutations reduced the effect of
222 high external [K+] on Co2+ transport in E. coli (Figure 5—figure supplement 2), most strikingly
223 with E134A and R352A. Overall these results further implicate the intact salt-bridge network as
224 contributing to DraNramp’s strong voltage dependence and imposing a restriction on most metal
225 transport when ΔΨ ≥ 0.
226
227
228 Metal-binding site and salt-bridge network contribute to DraNramp’s proton-metal
229 coupling
230 We next sought to disentangle the overlapping effects of ΔpH-stimulated metal transport and
231 metal-stimulated proton co-transport, which we refer to as proton-metal coupling. We measured
232 Mn2+ transport by our mutants in liposomes across a range of external pHs (and thus pH gradients)
233 both in the presence and absence of a ΔΨ < 0 (Figure 6A and Figure 6—figure supplement 1A-B).
11 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
234 Mutant phenotypes again clustered in three groups (Figure 6B). As before, mutants to metal-
235 binding D56 and N59 eliminated Mn2+ transport, while H237Q showed slight activity with
236 combined ΔΨ and ΔpH. As with voltage dependence, all mutants to the E124-R352-R353-D131-
237 E134 network reduced the Mn2+-transport enhancement provided by ΔpH, to a level comparable
238 to Cd2+ in WT. Strikingly, mutants M230A, H232Q, and E134Q—all near metal-binding D56—
239 eliminated ΔpH stimulation, such that unlike the WT, Mn2+ transport was not observed in the
240 absence of ΔΨ, no matter the ΔpH (Figure 6—figure supplement 1A-B).
241 Our mutants grouped differently when measuring how metal substrates Cd2+ and Mn2+ affect
242 proton transport at -120 and -80 mV (Figure 6C-D, Figure 6—figure supplement 1C, and Figure
243 6—figure supplement 2). Mutations to D56, H232, E134, and D131 eliminated both basal and
244 metal-induced H+ transport (E134A retained slight Mn2+ stimulation). In contrast to its inhibition
245 of WT, Cd2+ stimulated H+ uptake for mutants to metal-binding N59 and M230 and the 18 Å-
246 distant R352. Strikingly, for the E124 and R353 mutants, either Mn2+ or Cd2+ sharply reduced H+
247 transport—thus Mn2+ now behaved more like Cd2+ did for WT.
248 Overall, any mutation to the extended network perturbed proton-metal coupling and eliminated the
249 distinction between Cd2+ and Mn2+. This is consistent with previous observations of subtle
250 perturbations to this network affecting proton-metal coupling, as seen with an anemia-causing
251 mutation G185R on TM4 in mouse Nramp2 (Fleming et al., 1998; Su et al., 1998) (G153R in
252 DraNramp) directly above the R353-D131 pair (Xu et al., 2004), and an F-to-I substitution in
253 human Nramp2 at the DraNramp equivalent L164 adjacent to R352-E124 (Nevo and Nelson,
254 2004). Curiously, some mutants (D131A, E134A) retained some ΔpH stimulation of Mn2+
255 transport but little Mn2+ stimulation of H+ transport, while others (M230A, R352A) had the
256 opposite effect, with little ΔpH stimulation but robust Mn2+-stimulation. Interestingly, H232Q
12 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
257 eliminated both effects—consistent with the H-to-A mutant in Eremococcus coleocola Nramp
258 (Ehrnstorfer et al., 2017)—while D56 mutations eliminated all transport of metal and proton.
259 These results indicate that pH stimulation of metal transport and metal-proton co-transport are
260 distinct phenomena, likely imparted by separate groups of conserved residues.
261
262 Perturbing the salt-bridge network impairs Mn2+ transport under physiological conditions
263 As high transport efficiency requires selectivity against potential competing substrates, we tested
264 whether our mutants transport Ca2+, an abundant alkaline earth metal that Nramps must
265 discriminate against (Figures 7A-B and Figure 7—figure supplement 1A-B). Most mutations did
266 not increase Ca2+ transport. Thus, their perturbations of proton-metal coupling and voltage
267 dependence do not arise from indiscriminate metal transport. Metal-binding D56 and N59 mutants
268 reduced Ca2+ transport below WT levels, indicating that Ca2+ transport requires the same metal-
269 binding site as other metals. Two mutations enhanced Ca2+ uptake: M230A, which removes the
270 metal-coordinating methionine that preferentially stabilizes transition metals (Bozzi et al., 2016a),
271 and R353A. The altered selectivity of R353A, 15 Å from the bound metal, is reminiscent of G153R,
272 which adds exogenous positive charge to the salt-bridge network region and perturbs the metal-
273 binding site to improve Ca2+ transport (Bozzi et al., 2016b; Xu et al., 2004).
274 We next compared Cd2+ and Mn2+ transport by select mutants across a range of metal
275 concentrations to understand how these residues impact transport under more physiologically-
276 relevant conditions (Figure 7C-D). Most mutations had little impact on Cd2+ transport even at low
277 μM concentrations, and thus retained transport efficiency (calculated as Vmax/KM) within 2-fold of
278 WT. The exception was M230A, which drastically reduced Cd2+ transport, consistent with the
279 importance of this residue for Cd2+ affinity (Bozzi et al., 2016a; Bozzi et al., 2018). In contrast, all
13 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
280 mutations reduced Mn2+ transport at environmentally-relevant low Mn2+ concentrations, resulting
281 in 5 to 50-fold lower efficiency. Thus, mutagenic perturbations to the transporter that reduce
282 voltage dependence or eliminate the proton co-transport requirement also impair function under
283 physiological conditions. That these mutations alter DraNramp’s relative substrate preferences
284 further supports an allosteric functional link between the salt-bridge network and the metal-binding
285 site.
286
287 Voltage dependence and proton-metal coupling prevent deleterious Mn2+ back-transport
288 Lastly, we probed whether the salt-bridge network prevents back-transport of metal ions. As
289 membrane proteins typically orient randomly in proteoliposomes, we reconstituted DraNramp
290 constructs containing the single-cysteine A61C that lines the external vestibule and applied
291 membrane-impermeable 2-(Trimethylammonium)ethyl methanethiosulfonate (MTSET). Covalent
292 modification of A61C with MTSET nearly eliminates metal transport in vivo (Bozzi et al., 2016b)
293 and in vitro (Bozzi et al., 2018), likely by preventing outward-to-inward conformational change
294 essential for metal transport (Bozzi et al., 2018) (Figure 7—figure supplement 1C). In
295 proteoliposomes, MTSET treatment should thus incapacitate outside-out transporters, leaving
296 inside-out proteins unaffected—and thus theoretically capable of metal transport (Figures 7E and
297 S8D). Strikingly, while MTSET essentially eliminated Mn2+ transport for the WT-like protein,
298 salt-bridge network mutants D131N, E134A, and R352A retained significant activity at both
299 positive and negative ΔΨ (Figure 7F and Figure 7—figure supplement 1E). Thus inside-out mutant
300 transporters shuttled Mn2+ down its concentration gradient against ΔΨ —corresponding to back-
301 transport in a cellular context (Figure 7—figure supplement 1F)—while this process was only
302 minimally observed for the WT-like protein. MTSET-treated WT-like protein facilitated greater
14 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
303 Cd2+ transport, which unlike Mn2+ does not stimulate H+ uptake (Figure 2), further underscoring
304 the importance of proton co-transport to enforce unidirectional metal transport. In conclusion,
305 disruption of voltage dependence and proton-metal coupling through mutagenesis or by metal
306 identity imparts a greater risk of metal efflux under physiological conditions.
307
308 Discussion
309 We summarize the results of our mutagenesis experiments in Table S1, which inform the following
310 model (Figure 8A). The D56-E134-D131-R353-R352-E124 network confers strong voltage
311 dependence to the transporter, such that efficient metal transport only occurs at significantly
312 negative ΔΨ (Figures 1 and 5). D56 first protonates to optimally orient the metal-binding
313 residues—donating a hydrogen bond to the metal-interacting N59 carbonyl oxygen rather than
314 receiving a hydrogen bond from the amide nitrogen (Figure 8B)—a process facilitated by E134,
315 M230, and H232. D56 protonation also neutralizes the protein core. Incoming Mn2+ displaces the
316 H+ from D56 and interacts with N59, the A53 carbonyl, and M230, which bonds semi-covalently
317 to selectively stabilize the transition metal substrate. H232 and E134 may stabilize the H+ in a
318 transition state before it reaches D131. This H+ transfer thus redistributes the net added positive
319 charge, such that both the metal-binding site and the salt-bridge network gain a +1 formal charge.
320 Next, the proton is released to the cytoplasm through the salt-bridge network. Lastly, bulk
321 conformational change brings the transporter to the inward-facing state, in which the external
322 vestibule closes and the cytosolic vestibule opens to allow eventual Mn2+ release (Bozzi et al.,
323 2018). The salt-bridge network, a distinguishing feature of the Nramp clade of the LeuT family,
324 thus amplifies voltage dependence and provides for temporal and spatial separation of the H+ and
325 Mn2+ movement.
15 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
326 Substrate-specific differences in stoichiometry of the driving ion, as we demonstrated for
327 DraNramp (Figure 2), have been reported for unrelated transporters (Chen et al., 1999b; Dohan et
328 al., 2007), and different metal ions can trigger distinct allosteric responses upon binding,
329 especially in metal-sensor proteins (Helmann, 2014; Waldron and Robinson, 2009). For DraNramp,
330 the distinction between proton-metal symport (Mn2+, Fe2+, and Co2+) and metal uniport (Zn2+ and
331 Cd2+) follows a familiar inorganic chemistry partition. The former cations have 5, 6, or 7 valence
332 d-electrons and typically prefer a coordination arrangement with more electron-donating ligands
333 than do the latter metals which have a filled d-shell (Barber-Zucker et al., 2017). Therefore, while
334 D56 coordinates Mn2+ in a bidentate manner requiring deprotonation (Figure 8D) as we modeled
335 for DraNramp’s outward-facing state (Bozzi et al., 2018), D56 may instead interact with Cd2+ only
336 through its carbonyl oxygen and thus retain its proton (Figure 8C). The greater importance of
337 M230 for Cd2+ than Mn2+ transport (Figure 7C-D and Bozzi et al., 2016a; Bozzi et al., 2018) and
338 the necessity of a proposed transient metal-binding residue Q378 for high-affinity Mn2+ but not
339 Cd2+ transport (Bozzi et al., 2018), further supports differential binding properties for the two
340 metals. Our Mn2+/H+ symport results are consistent with in vitro findings with Mn2+ for E.
341 coleocola Nramp (Ehrnstorfer et al., 2017). In contrast, the Cd2+ uniport we report contradicts in
342 vivo results showing Cd2+ stimulates intracellular acidification in cells overexpressing E. coli
343 Nramp (Courville et al., 2008)—which has the same conserved salt-bridge network as DraNramp
344 (Figure 4). However, this study did not address the effects of Mn2+ or other metal substrates on
345 intracellular pH. Furthermore, the inherent complexity of such in vivo experiments compared to
346 an artificial proteoliposome means that the observed intracellular pH changes were perhaps not
347 due to flux through E. coli Nramp.
16 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
348 Thermodynamically, proton-coupled transport promotes M2+ uptake under favorable ΔpH and ΔΨ
2+ 2+ 349 conditions and thus enables a more substantial cytoplasmic vs. external Δ[Mn ] than Δ[Cd ]. The
350 distinction in co-transport stoichiometry in DraNramp may therefore have evolved to prevent the
351 excessive accumulation of toxic Cd2+ while preserving the ability to amass large stores (Daly et
352 al., 2004) of an essential resource like Mn2+.
353 Intriguingly, perturbing either the metal-binding site or the adjoining salt-bridge network removes
354 the H+ co-transport distinction between Cd2+ and Mn2+ (Figure 6C). Cd2+ stimulated H+ co-
355 transport by the N59A and M230A variants, suggesting that D56 may compensate for the loss of
356 a native metal ligand by interacting in a bidentate manner with Cd2+ and thus shedding a H+. In
357 addition, mutations at residues ≥ 15 Å from bound M2+ also upset this distinction—the R352A
358 variant symports H+ with both metals, while E124A and R353A shift to uniport of both metals —
359 further underscoring the long-range influence of the salt-bridge network.
360 Overall, our experiments with different metals and point mutations indicate two distinct roles for
361 protons in DraNramp’s mechanism: (i) ΔpH stimulation of metal transport, and (ii) metal
362 stimulation of proton co-transport, as we highlight in Figures 2 and 6. Lower pH likely optimizes
363 the metal-binding site via D56 protonation, and mutations to residues in the immediate vicinity of
364 D56 (M230, E134, H232) completely eliminate pH dependence of transport (Figure 6A-B),
2+ 2+ 365 perhaps by shifting D56’s pKa. Yet low pH still accelerates Cd transport in WT, and Mn
366 transport in proton-pathway mutants like D131A, although less than in WT. Complementarily,
367 proton co-transport is still observed for mutants such as M230A that eliminate the effect of ΔpH
368 but do not compromise the conserved proton-transport pathway (Bozzi et al., 2018). Therefore, a
369 ΔpH effect on the uptake of primary substrate should not be conflated with H+ co-transport and
370 vice versa, whether in other Nramp homologs or unrelated secondary transporters.
17 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
371 The canonical model for secondary transport posits a cycle in which each transition is fully
372 reversible, precise stoichiometry is maintained, and the prevailing ion gradients determine the
373 directionality of net transport (Figure 8E). Effective symport thus requires tight coupling between
374 the co-substrates, which is implemented in transporters by the driving ions either structurally
375 enabling binding of the primary substrate or mechanistically selectively stabilizing a certain
376 conformational state (Perez and Ziegler, 2013; Rudnick, 2013). Indeed, for the LeuT amino acid
377 transporter—a structural homolog of DraNramp—substrate binding depends on two bound Na+
378 ions (Erlendsson et al., 2017): one Na+ stabilizes the outward-open conformation (Claxton et al.,
379 2010; Tavoulari et al., 2016) while a second Na+ binds the transported amino acid’s carboxylate
380 (Yamashita et al., 2005).
381 The DraNramp proton-metal symport mechanism (Figure 8F) deviates from basic principles of
382 canonical symport. First, substrate coupling is asymmetric: Mn2+ transport requires H+ co-transport,
383 to the extent that an Mn2+ gradient cannot drive Mn2+/H+ transport unless H+ entry alone is
384 thermodynamically favorable (from ΔΨ and/or ΔpH, as seen in Figures 1 and 2). In contrast, while
385 Mn2+ stimulates H+ transport, H+ uniport still occurs readily—and without requiring bulk
386 conformational change (Bozzi et al., 2018)—resulting in the variable co-transport stoichiometries
387 previously reported for yeast and human Nramp homologs (Chen et al., 1999a; Mackenzie et al.,
388 2007; Mackenzie et al., 2006). These “futile cycles” are energetically wasteful, depleting ΔΨ and
389 ΔpH without assisting in the uptake of the primary substrate. It is intriguing that evolution has
390 retained such a thermodynamic cost as a general feature of the Nramp family, especially as simple
391 tweaks near the salt-bridge network were shown to reduce H+ uniport without impairing metal
392 uptake (Nevo and Nelson, 2004). Second, in our model, the two substrates do not coexist in the
393 binding site throughout the transport process: Mn2+ binding to D56 evicts the pre-bound H+ into
18 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
394 its distinct exit pathway. Indeed, our results suggest that H+ transport and M2+ binding may actually
395 become competitive processes under some conditions, underscoring the co-substrate’s imperfect
396 synergy (Figure 3).
397 In addition, our results demonstrate an essential kinetic role for the physiological membrane
398 potential in DraNramp metal transport. While voltage necessarily contributes to the net
399 thermodynamics of moving charged substrates across the membrane, we observe a strong voltage
400 dependence of the entire free energy landscape (Figure 3), such that a sufficiently negative ΔΨ is
401 prerequisite for metal binding and/or transport to occur. Although voltage dependence may be a
402 diffuse property for a transport protein, the conserved salt-bridge network lining the ion transport
403 routes in Nramps likely evolved to amplify the inherent voltage dependence of electrogenic
404 transport and establish delicate mechanistic restrictions—including metal-specific proton co-
405 transport—which our mutagenesis data support (Figures 5-6).
406 The key outcome of these deviations from canonical secondary transport—loose proton-metal
407 coupling and strong voltage dependence—is to enforce the unidirectionality of DraNramp metal
408 transport (Figure 8F). Indeed, no Mn2+ transport occurred in the absence of a negative ΔΨ (Figure
409 1; cations leaving the cell would experience a net positive ΔΨ). Furthermore inside-out
410 transporters—even with a favorable ΔΨ—fail to efficiently import Mn2+ into liposomes down a
411 large concentration gradient (Figure 7F). The inward-to-outward metal-bound transition is thus
412 essentially forbidden in the WT protein under physiological conditions. Strikingly, point mutations
413 to the salt-bridge network that perturb proton-metal coupling and voltage dependence can lift these
414 restrictions, such that these protein variants behave more like directionally-unbiased uniporters.
415 DraNramp’s unique structural adaptation—a conserved salt-bridge network that allosterically
416 amplifies the voltage dependence of metal cation transport and provides a multistep, one-way exit
19 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
417 pathway to the cytoplasm for the loosely-coupled proton co-substrate—therefore allows it to avoid
418 the liabilities of reversibility inherent in the canonical model for symport. Similar properly-tuned
419 voltage dependence and asymmetric substrate coupling may be a more general strategy for
420 transporters to prevent deleterious back-transport without jeopardizing transport in the desired
421 direction.
422
423 Methods
424 In vivo Co2+ transport
2+ 425 Co transport in E. coli was performed as described with 200 μM Co(NO3)2 (Bozzi et al., 2018),
426 except that the assay buffer was 50 mM HEPES pH 7.0, 60 mM NaCl, 10 mM KCl, 0.5 mM MgCl2,
427 0.216% glucose. For the varied [K+] measurements, the assay buffer was 50 mM HEPES pH 7.0,
428 60 mM NaCl, 0.5 mM MgCl2, X mM KCl, and (82.5-X) mM choline chloride, with X indicated
429 in the corresponding figure legends. For each biological replicate reported in the figure legends, a
430 separate culture of transformed E. coli was grown and induced to express the exogenous Nramp
431 construct.
432
433 Cysteine accessibility measurements
434 Cysteine accessibility measurements in E. coli were performed as described (Bozzi et al., 2018),
435 except that the assay buffer used was the same as for the Co2+ transport experiments. For each
436 biological replicate reported in the figure legends, a separate culture of transformed E. coli was
437 grown and induced to express the exogenous Nramp construct.
438
439 Cloning, expression and purification of DraNramp constructs for proteoliposome assays
20 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
440 DraNramp constructs were cloned, expressed, and purified as described (Bozzi et al., 2018), with
441 the following changes: protein was purified from cell pellets in a single day, and washed/eluted
442 from nickel beads in buffers with 0.03% DDM. Protein was concentrated to 2.5 mL and buffer-
443 exchanged into 100 mM NaCl, 10 mM HEPES pH 7.5, 0.1% n-Decyl-β-D-Maltopyranoside (DM)
444 on a PD10 desalting column. Protein concentrations were normalized to 1.2 mg/ml and aliquots
445 were flash frozen in liquid nitrogen. Single-cysteine constructs A53C and A61C were purified in
446 the presence of 1 mM DTT.
447
448 Proteoliposome preparation
449 Adjusting the lipid composition (Ehrnstorfer et al., 2017) of a previous protocol (Bozzi et al.,
450 2016a; Tsai et al., 2014), 75% w/w 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoethanolamine
451 (POPE) was mixed with 25% w/w 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoglycerol (POPG) in
452 chloroform (Avanti Polar Lipids), and then dried under nitrogen in a warm water bath, re-dissolved
453 in pentane, and dried again. Lipids were resuspended at 20 mg/ml in 5 mM DM in
454 KCl+NaCl/MOPS buffer (typically ~90 mM KCl, 30 mM NaCl, 0.5 mM or 10 mM MOPS pH 7).
455 Protein was added at a 1:400 w/w ratio to lipid, and the mixture dialyzed at 4°C to remove the
456 detergent in 10 kDa molecular weight cutoff dialysis cassettes against KCl+NaCl/MOPS buffer
457 with 0.2 mM EDTA for 1 day, then with 0.1 mM EDTA for 1-3 days, then overnight at room
458 temperature (RT) against KCl+NaCl/MOPS buffer. For A53C and A61C, 1 mM, and 0.5 mM DTT
459 was included in the first two dialysis steps. Fluorescent dye (either 1:49 v/v 5 mM Fura-2
460 pentapotassium salt or 1:66 v/v 10 mM 2',7'-bis(carboxyethyl)-5(6)-Carboxyfluorescein (BCECF)
461 in dimethyl sulfoxide) was incorporated into proteoliposomes permeabilized by three freeze-thaw
462 cycles in dry ice-ethanol and RT water baths (and sometimes stored at -80°C after the third freeze).
21 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
463 Proteoliposomes were extruded through a 400 nM filter to create uniform-sized vesicles, buffer-
464 exchanged 1-2 times on a PD10 desalting column pre-equilibrated with NaCl/ or KCl/10 mM
465 MOPS pH 7 buffer. Peak proteoliposome-containing fractions were pooled to remove
466 unincorporated dye.
467
468 Proteoliposome transport assays and data analysis
469 Proteoliposomes loaded with either 100 μM Fura-2 or 150 μM BCECF were diluted into buffer
470 containing appropriate [KCl] to establish the desired membrane potential(Fitzgerald et al., 2017;
471 Uzdavinys et al., 2017) and aliquoted into 96 well black clear-bottom plates. Following baseline
472 fluorescence measurement, 5X metal (750 μM final concentration unless otherwise noted) and
473 valinomycin (100 nM final concentration) were added. A fresh stock of 100 mM FeSO4 in 100
474 mM L-ascorbic acid was made for each assay; stocks of 100 mM CdCl2, MnCl2, ZnCl2, and
475 Co(NO3)2 were freshly diluted into appropriate NaCl or KCl buffer. For assays with a pH gradient,
476 the metals were diluted into 100 mM MES at pH 5.5, 5.8, 6.0, or 6.5 with appropriate NaCl/KCl.
477 The reported external pH upon metal addition was determined from proportional mixings of larger
478 volumes of the same buffers. To pre-modify A61C constructs for transport assays, liposomes were
479 diluted into 120 mM NaCl, 10 mM MOPS pH 7 containing 3 mM MTSET and incubated at least
480 30 min at RT before beginning transport assays.
481 Metal transport was monitored by measuring Fura-2 fluorescence at λex = 340 and 380 nm, at λem
482 = 510 nm. Proton transport was monitored by measuring BCECF fluorescence at λex = 450 and
483 490 nm, at λem = 535 nm. To calculate concentrations of imported metal, the Fura-2 340/380 ratio
484 and experimentally determined KD values (Grynkiewicz et al., 1985; Hinkle et al., 1992) were used
2+ 2+ 2+ 485 for Ca and Cd as described previously (Bozzi et al., 2016a); the Ca KD value was used as an
22 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
486 approximation for Zn2+. For Mn2+, Fe2+, and Co2+, the fraction of Fura-2 340 and/or 380
487 fluorescence quenched, normalized to maximum observed quenching, was used to estimate
488 imported metal. For proton uptake, the BCECF 450/490 ratio was used to calculate internal pH,
489 which along with the known total internal buffer (0.5 mM) and dye (150 μM) concentration was
490 used to calculate net proton import via the Henderson-Hasselbalch equation. Initial rates were
491 calculated in Excel and Michaelis-Menten parameters were fit using MATLAB. For each technical
492 replicate reported in figure legends, a separate aliquot of dye-loaded proteoliposomes was diluted
493 into the appropriate outside buffer, including cysteine modifiers if applicable, then fluorescence
494 time course data were collected before and after the addition of valinomycin, metal substrate,
495 and/or ionomycin.
496
497 Data availability
498 The raw biochemical data that support the findings of this study are available from the
499 corresponding author upon reasonable request.
500
501 Author contributions
502 R.G. oversaw and designed the research with A.T.B., L.B.B., and C.M.Z.; L.B.B. and A.T.B.
503 developed the proteoliposome transport assay and performed preliminary experiments; C.M.Z.
504 provided essential structural insights into the DraNramp transport mechanism that guided
505 mutagenesis experiments; A.T.B. performed all experiments and analyzed the resulting data;
506 A.T.B. and R.G. wrote the manuscript, with input from all authors.
507
508 Competing Interests
23 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
509 The authors declare no competing interests.
510
511 Acknowledgements
512 We would like to thank Sarah Farron for assistance with transport kinetics analysis, Casey Zhang
513 for generating the Nramp sequence alignment, Chris Miller and Joe Mindell for critical feedback,
514 and Niels Bradshaw, Jack Nicoludis, and the rest of the Gaudet lab for helpful discussions. This
515 work was funded in part by NIH grant 1R01GM120996 (to R.G.) and a Jane Coffin Childs
516 Postdoctoral Fellowship (to C.M.Z.).
517
518 References
519 Abbaspour, N., Hurrell, R., and Kelishadi, R. (2014). Review on iron and its importance for 520 human health. Journal of Research in Medical Sciences 19, 164‐174.
521 Andrews, N.C. (2008). Forging a field: the golden age of iron biology. Blood 112, 219‐230.
522 Barber‐Zucker, S., Shaanan, B., and Zarivach, R. (2017). Transition metal binding selectivity 523 in proteins and its correlation with the phylogenomic classification of the cation diffusion 524 facilitator protein family. Scientific Reports 7.
525 Bot, C.T., and Prodan, C. (2010). Quantifying the membrane potential during E. coli growth 526 stages. Biophysical Chemistry 146, 133‐137.
527 Boudker, O., and Verdon, G. (2010). Structural perspectives on secondary active 528 transporters. Trends in Pharmacological Sciences 31, 418‐426.
529 Bozzi, A.T., Bane, L.B., Weihofen, W.A., McCabe, A.L., Singharoy, A., Chipot, C.J., Schulten, K., 530 and Gaudet, R. (2016a). Conserved methionine dictates substrate preference in Nramp‐ 531 family divalent metal transporters. Proceedings of the National Academy of Sciences of the 532 United States of America 113, 10310‐10315.
533 Bozzi, A.T., Bane, L.B., Weihofen, W.A., Singharoy, A., Guillen, E.R., Ploegh, H.L., Schulten, K., 534 and Gaudet, R. (2016b). Crystal Structure and Conformational Change Mechanism of a 535 Bacterial Nramp‐Family Divalent Metal Transporter. Structure 24, 2102‐2114.
536 Bozzi, A.T., Zimanyi, C.M., Nicoludis, J.M., Lee, B.K., Zhang, C.M., and Gaudet, R. (2018). 537 Structures in multiple conformations reveal distinct transition metal and proton pathways 538 in an Nramp transporter.
24 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
539 Bressler, J.P., Olivi, L., Cheong, J.H., Kim, Y., and Bannon, D. (2004). Divalent metal 540 transporter 1 in lead and cadmium transport. Redox‐Active Metals in Neurological 541 Disorders 1012, 142‐152.
542 Cellier, M.F.M. (2012). Nramp: from Sequence to Structure and Mechanism of Divalent 543 Metal Import. Metal Transporters 69, 249‐293.
544 Cellier, M.F.M. (2016). Evolutionary analysis of Slcll mechanism of proton‐coupled metal‐ 545 ion transmembrane import. Aims Biophysics 3, 286‐318.
546 Chen, X.Z., Peng, J.B., Cohen, A., Nelson, H., Nelson, N., and Hediger, M.A. (1999a). Yeast 547 SMF1 mediates H+‐coupled iron uptake with concomitant uncoupled cation currents. 548 Journal of Biological Chemistry 274, 35089‐35094.
549 Chen, X.Z., Zhu, T., Smith, D.E., and Hediger, M.A. (1999b). Stoichiometry and kinetics of the 550 high‐affinity H+‐coupled peptide transporter PepT2. Journal of Biological Chemistry 274, 551 2773‐2779.
552 Claxton, D.P., Quick, M., Shi, L., de Carvalho, F.D., Weinstein, H., Javitch, J.A., and McHaourab, 553 H.S. (2010). Ion/substrate‐dependent conformational dynamics of a bacterial homolog of 554 neurotransmitter: sodium symporters. Nature Structural & Molecular Biology 17, 822‐ 555 U868.
556 Coffey, R., and Ganz, T. (2017). Iron homeostasis: An anthropocentric perspective. Journal 557 of Biological Chemistry 292, 12727‐12734.
558 Courville, P., Urbankova, E., Rensing, C., Chaloupka, R., Quick, M., and Cellier, M.F. (2008). 559 Solute carrier 11 cation symport requires distinct residues in transmembrane helices 1 and 560 6. The Journal of biological chemistry 283, 9651‐9658.
561 Daly, M.J., Gaidamakova, E.K., Matrosova, V.Y., Vasilenko, A., Zhai, M., Venkateswaran, A., 562 Hess, M., Omelchenko, M.V., Kostandarithes, H.M., Makarova, K.S., et al. (2004). 563 Accumulation of Mn(II) in, Deinococcus radiodurans facilitates gamma‐radiation 564 resistance. Science 306, 1025‐1028.
565 Dohan, O., Portulano, C., Basquin, C., Reyna‐Neyra, A., Amzel, L.M., and Carrasco, N. (2007). 566 The Na+/I‐ symporter (NIS) mediates electroneutral active transport of the environmental 567 pollutant perchlorate. Proceedings of the National Academy of Sciences of the United States 568 of America 104, 20250‐20255.
569 Ehrnstorfer, I.A., Geertsma, E.R., Pardon, E., Steyaert, J., and Dutzler, R. (2014). Crystal 570 structure of a SLC11 (NRAMP) transporter reveals the basis for transition‐metal ion 571 transport. Nature Structural & Molecular Biology 21, 990‐996.
572 Ehrnstorfer, I.A., Manatschal, C., Arnold, F.M., Laederach, J., and Dutzler, R. (2017). 573 Structural and mechanistic basis of proton‐coupled metal ion transport in the 574 SLC11/NRAMP family. Nature Communications 8.
25 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
575 Erlendsson, S., Gotfryd, K., Larsen, F.H., Mortensen, J.S., Geiger, M.A., van Rossum, B.J., 576 Oschkinat, H., Gether, U., Teilum, K., and Loland, C.J. (2017). Direct assessment of substrate 577 binding to the Neurotransmitter: Sodium Symporter LeuT by solid state NMR. Elife 6.
578 Fitzgerald, G.A., Mulligan, C., and Mindell, J.A. (2017). A general method for determining 579 secondary active transporter substrate stoichiometry. Elife 6.
580 Fleming, M.D., Romano, M.A., Su, M.A., Garrick, L.M., Garrick, M.D., and Andrews, N.C. 581 (1998). Nramp2 is mutated in the anemic Belgrade (b) rat: Evidence of a role for Nramp2 in 582 endosomal iron transport. Proceedings of the National Academy of Sciences of the United 583 States of America 95, 1148‐1153.
584 Forrest, L.R., Kramer, R., and Ziegler, C. (2011). The structural basis of secondary active 585 transport mechanisms. Biochimica Et Biophysica Acta‐Bioenergetics 1807, 167‐188.
586 Gadsby, D.C. (2009). Ion channels versus ion pumps: the principal difference, in principle. 587 Nature Reviews Molecular Cell Biology 10, 344‐352.
588 Grynkiewicz, G., Poenie, M., and Tsien, R.Y. (1985). A New Generation of Ca2+ Indicators 589 with Greatly Improved Fluorescence Properties. Journal of Biological Chemistry 260, 3440‐ 590 3450.
591 Gunshin, H., Mackenzie, B., Berger, U.V., Gunshin, Y., Romero, M.F., Boron, W.F., Nussberger, 592 S., Gollan, J.L., and Hediger, M.A. (1997). Cloning and characterization of a mammalian 593 proton‐coupled metal‐ion transporter. Nature 388, 482‐488.
594 Haemig, H.A., and Brooker, R.J. (2004a). Importance of conserved acidic residues in mntH, 595 the Nramp homolog of Escherichia coli. The Journal of membrane biology 201, 97‐107.
596 Haemig, H.A., Moen, P.J., and Brooker, R.J. (2010). Evidence that highly conserved residues 597 of transmembrane segment 6 of Escherichia coli MntH are important for transport activity. 598 Biochemistry 49, 4662‐4671.
599 Haemig, H.A.H., and Brooker, R.J. (2004b). Importance of conserved acidic residues in 600 MntH, the Nramp homolog of Escherichia coli. Journal of Membrane Biology 201, 97‐107.
601 Helmann, J.D. (2014). Specificity of Metal Sensing: Iron and Manganese Homeostasis in 602 Bacillus subtilis. Journal of Biological Chemistry 289, 28112‐28120.
603 Hinkle, P.M., Shanshala, E.D., and Nelson, E.J. (1992). Measurement of Intracellular 604 Cadmium with Fluorescent Dyes ‐ Further Evidence for the Role of Calcium Channels in 605 Cadmium Uptake. Journal of Biological Chemistry 267, 25553‐25559.
606 Illing, A.C., Shawki, A., Cunningham, C.L., and Mackenzie, B. (2012). Substrate Profile and 607 Metal‐ion Selectivity of Human Divalent Metal‐ion Transporter‐1. Journal of Biological 608 Chemistry 287, 30485‐30496.
26 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
609 Iolascon, A., d'Apolito, M., Servedio, V., Cimmino, F., Piga, A., and Camaschella, C. (2006). 610 Microcytic anemia and hepatic iron overload in a child with compound heterozygous 611 mutations in DMT1 (SCL11A2). Blood 107, 349‐354.
612 Lam‐Yuk‐Tseung, S., Camaschella, C., Iolascon, A., and Gros, P. (2006). A novel R416C 613 mutation in human DMT1 (SLC11A2) displays pleiotropic effects on function and causes 614 microcytic anemia and hepatic iron overload. Blood Cells Molecules and Diseases 36, 347‐ 615 354.
616 Lam‐Yuk‐Tseung, S., Govoni, G., Forbes, J., and Gros, P. (2003). Iron transport by 617 Nramp2/DMT1: pH regulation of transport by 2 histidines in transmembrane domain 6. 618 Blood 101, 3699‐3707.
619 LeVine, M.V., Cuendet, M.A., Khelashvili, G., and Weinstein, H. (2016). Allosteric Mechanisms 620 of Molecular Machines at the Membrane: Transport by Sodium‐Coupled Symporters. 621 Chemical Reviews 116, 6552‐6587.
622 Loo, D.D.F., Hirayama, B.A., Gallardo, E.M., Lam, J.T., Turk, E., and Wright, E.M. (1998). 623 Conformational changes couple Na+ and glucose transport. Proceedings of the National 624 Academy of Sciences of the United States of America 95, 7789‐7794.
625 Ma, Z., Jacobsen, F.E., and Giedroc, D.P. (2009). Coordination Chemistry of Bacterial Metal 626 Transport and Sensing. Chemical Reviews 109, 4644‐4681.
627 Mackenzie, B., Takanaga, H., Hubert, N., Rolfs, A., and Hediger, M.A. (2007). Functional 628 properties of multiple isoforms of human divalent metal‐ion transporter 1 (DMT1). 629 Biochemical Journal 403, 59‐69.
630 Mackenzie, B., Ujwal, M.L., Chang, M.H., Romero, M.F., and Hediger, M.A. (2006). Divalent 631 metal‐ion transporter DMT1 mediates both H+‐coupled Fe2+ transport and uncoupled 632 fluxes. Pflugers Archiv‐European Journal of Physiology 451, 544‐558.
633 Makui, H., Roig, E., Cole, S.T., Helmann, J.D., Gros, P., and Cellier, M.F.M. (2000). 634 Identification of the Escherichia coli K‐12 Nramp orthologue (MntH) as a selective divalent 635 metal ion transporter. Molecular microbiology 35, 1065‐1078.
636 Nelson, N., Sacher, A., and Nelson, H. (2002). The significance of molecular slips in transport 637 systems. Nature Reviews Molecular Cell Biology 3, 876‐881.
638 Nevo, Y., and Nelson, N. (2004). The mutation F227I increases the coupling of metal ion 639 transport in DCT1. Journal of Biological Chemistry 279, 53056‐53061.
640 Perez, C., and Ziegler, C. (2013). Mechanistic aspects of sodium‐binding sites in LeuT‐like 641 fold symporters. Biol Chem 394, 641‐648.
27 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
642 Picard, V., Govoni, G., Jabado, N., and Gros, P. (2000). Nramp 2 (DCT1/DMT1) expressed at 643 the plasma membrane transports iron and other divalent cations into a calcein‐accessible 644 cytoplasmic pool. Journal of Biological Chemistry 275, 35738‐35745.
645 Pujol‐Gimenez, J., Hediger, M.A., and Gyimesi, G. (2017). A novel proton transfer mechanism 646 in the SLC11 family of divalent metal ion transporters. Scientific Reports 7.
647 Rudnick, G. (2013). How do transporters couple solute movements? Molecular Membrane 648 Biology 30, 355‐359.
649 Sacher, A., Cohen, A., and Nelson, N. (2001). Properties of the mammalian and yeast metal‐ 650 ion transporters DCT1 and Smf1p expressed in Xenopus laevis oocytes. Journal of 651 Experimental Biology 204, 1053‐1061.
652 Shabala, L., Bowman, J., Brown, J., Ross, T., McMeekin, T., and Shabala, S. (2009). Ion 653 transport and osmotic adjustment in Escherichia coli in response to ionic and non‐ionic 654 osmotica. Environmental Microbiology 11, 137‐148.
655 Shi, Y. (2013). Common Folds and Transport Mechanisms of Secondary Active 656 Transporters. Annual Review of Biophysics, Vol 42 42, 51‐72.
657 Shilton, B.H. (2015). Active transporters as enzymes: an energetic framework applied to 658 major facilitator superfamily and ABC importer systems. Biochemical Journal 467, 193‐199.
659 Su, M.A., Trenor, C.C., Fleming, J.C., Fleming, M.D., and Andrews, N.C. (1998). The G185R 660 mutation disrupts functions of the iron transporter Nramp2. Blood 92, 2157‐2163.
661 Tavoulari, S., Margheritis, E., Nagarajan, A., DeWitt, D.C., Zhang, Y.‐W., Rosado, E., Ravera, S., 662 Rhoades, E., Forrest, L.R., and Rudnick, G. (2016). Two Na+ Sites Control Conformational 663 Change in a Neurotransmitter Transporter Homolog. Journal of Biological Chemistry 291, 664 1456‐1471.
665 Tsai, M.F., Jiang, D.W., Zhao, L.L., Clapham, D., and Miller, C. (2014). Functional 666 reconstitution of the mitochondrial Ca2+/H+ antiporter Letm1. Journal of General 667 Physiology 143, 67‐73.
668 Uzdavinys, P., Coincon, M., Nji, E., Ndi, M., Winkelmann, I., von Ballmoos, C., and Drew, D. 669 (2017). Dissecting the proton transport pathway in electrogenic Na+/H+ antiporters. 670 Proceedings of the National Academy of Sciences of the United States of America 114, 671 E1101‐E1110.
672 Vazquez, M., Velez, D., Devesa, V., and Puig, S. (2015). Participation of divalent cation 673 transporter DMT1 in the uptake of inorganic mercury. Toxicology 331, 119‐124.
674 Waldron, K.J., and Robinson, N.J. (2009). How do bacterial cells ensure that metalloproteins 675 get the correct metal? Nat Rev Microbiol 7, 25‐35.
28 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
676 Xu, H.X., Jin, J., DeFelice, L.J., Andrews, N.C., and Clapham, D.E. (2004). A spontaneous, 677 recurrent mutation in divalent metal transporter‐1 exposes a calcium entry pathway. PLoS 678 biology 2, 378‐386.
679 Yamashita, A., Singh, S.K., Kawate, T., Jin, Y., and Gouaux, E. (2005). Crystal structure of a 680 bacterial homologue of Na+/Cl‐‐dependent neurotransmitter transporters. Nature 437, 681 215‐223.
29 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
682 683 Figure 1. Metal transport shows strong voltage dependence. 684 (A) Schematic of proteoliposome assay for metal transport at variable membrane potentials (ΔΨ) 685 established by using K+ gradients and valinomycin. (B) Representative time traces (n = 8) of Cd2+, 686 Mn2+, and Co2+ transport at different ΔΨs; only Cd2+ is transported at ΔΨ = 0 mV. (C) Average 687 initial transport rate ± S.E.M. (n = 4) versus membrane potential. Each metal except Cd2+ has a 688 characteristic threshold negative voltage for transport to occur: -10 mV for Mn2+, -50 mV for Fe2+ 689 and Zn2+, and -80 mV for Co2+. (D) Average initial transport rate ± S.E.M. (n = 2-3) versus metal 690 concentration. Comparison of panels C and D (table inset) show that higher-magnitude threshold 691 voltages correlate with lower apparent affinity (KM), indicating that voltage dependence may be 692 enforced through the metal-binding site. Errors in KM and Vmax encompass the uncertainty of fit 693 (shown as curved lines) to data. (E) Time course of Co2+ uptake shows that high extracellular [KCl] 694 depresses Co2+ uptake in E. coli. (F) Complementary single-cysteine reporters to assess 695 conformational cycling of DraNramp. (G) Fraction of modified cysteine versus NEM 696 concentration. Inward (A53C) and outward (A61C) reporters both label even at high [KCl]out that 697 depresses Co2+ uptake. Conformational-locking thus cannot explain reduced Co2+ transport by 698 DraNramp under these conditions. Data in E and G are averages ± S.E.M. (n = 4). See also Figure 699 1—figure supplement 1.
30 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
700 701 Figure 1—figure supplement 1. DraNramp is a voltage-dependent transition metal 702 transporter. 703 (A) Coomassie-stained SDS-gel of His-tagged DraNramp purification via Ni2+-affinity showing 704 load, flowthrough (FT), washes (W1, W2, W3) and elutions (E1, E2). (B) Colorimetric detection 705 of Co2+ taken up by cells over time shows that DraNramp transports Co2+ when expressed in E. 706 coli. Data are averages ± S.E.M. (n = 4). (C) Control liposomes lacking DraNramp are used to 707 determine protein-dependent transport activity. (D) Representative time traces (n ≥ 4) of metal 708 uptake by DraNramp-containing liposomes (top) show that Fe2+ and Zn2+ also exhibit strong 709 voltage dependence, while Ca2+ remains a poor substrate across all voltages. Control liposomes 710 (bottom) have slow but non-negligible, voltage-independent Fe2+ and Zn2+ leak. (E) Representative 711 time traces (n = 4) of metal uptake by DraNramp-containing liposomes at different membrane 712 voltages. Transport of all five tested substrates is accelerated at high magnitude negative
31 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
2+ 713 membrane potentials. (F) Dose-response curve shows that Co transport in E. coli has KM = 700 714 ± 100 μM similar to that obtained at ΔΨ = -150 mV in proteoliposomes, confirming that DraNramp 715 behavior in vitro is physiologically relevant. Data are averages ± S.E.M. (n = 4). Error in KM is the 716 uncertainty of fit to data. (G) Representative image of Co2+ uptake assay results showing that 717 increased extracellular [K+] inhibits DraNramp-dependent Co2+ uptake in E. coli, as detected by 718 precipitating transported Co2+ as the dark solid CoS. (H) Representative western blots showing 719 that A61C is labeled by N-ethylmaleimide (NEM) similarly in high or low external [K+].
32 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
720 721 Figure 2. pH stimulation is distinct from proton co-transport. 722 (A) Schematic for assessing ΔpH effect on metal transport in proteoliposomes. (B-C) 723 Representative time traces (n ≥ 4) of Cd2+ (B) or Mn2+ (C) uptake into proteoliposomes. ΔpH 724 accelerated Cd2+ and Mn2+ transport (at ΔΨ = -60 mV), and enabled Mn2+ uptake when ΔΨ = 0. 725 (D) Schematic for detecting H+ influx into proteoliposomes. (E) Representative time traces (n = 8) 726 of H+ uptake into proteoliposomes. Negative ΔΨ drove DraNramp H+ import. Mn2+ and Co2+ 727 stimulated H+ entry, while Cd2+ did not, and instead reduced H+ influx at ΔΨ = -120 mV. Except 728 for the reporter dye, conditions were identical to those in Figure 1B. (F) Initial metal (black) and 729 H+ (color) uptake rates show that Mn2+, Fe2+, and Co2+ transport stimulated H+ influx above its 730 basal no-metal rate, while Cd2+ and Zn2+ transport did not. Stoichiometry ratios (numbers above 731 bars) were calculated for Cd2+ and Mn2+ and were approximately 1:1 for Mn2+/H+ symport. Data 732 are averages ± S.E.M. (n ≥ 4). See also Figure 2—figure supplement 1.
33 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
733 734 Figure 2—figure supplement 1. Proton co-transport is metal-specific. 735 (A) Schematic for proteoliposome assays with moderate negative and positive ΔΨ values. 736 Liposomes for ΔΨ ≤ 0 contained only KCl inside and were diluted to a lower external [K+], with 737 increased Na+ used to balance osmolarity, before adding valinomycin to establish the membrane 738 potential. Liposomes for ΔΨ ≥ 0 contained mainly NaCl inside, and were diluted into higher 739 outside [K+]. (B-C) Average initial transport rate ± S.E.M. (n = 4) versus membrane potential. 740 Cd2+ transport (B) driven by a [Cd2+] gradient occurs even at ΔΨ ≥ 0, although the rate decreases 741 at more positive ΔΨ. In contrast, Mn2+ transport (C) does not occur unless ΔΨ < 0. Importantly, 742 for both Cd2+ and Mn2+ at 0 mV, the observed transport behavior is the same regardless of whether 743 the dominant cation species on both sides of the membrane is K+ or Na+. Thus, these ions on their 744 own do not enable or inhibit transport, but rather the membrane potential causes the observed 745 variations in transport rates. (D) Representative time traces (n ≥ 4) of H+ uptake. With 746 proteoliposomes (top), Fe2+ stimulated H+ entry, while Zn2+ (substrate) and Ca2+ (non-substrate) 747 did not. ΔΨ-driven H+ import was DraNramp-dependent, as there was no significant H+ leak or 748 metal-stimulated H+ entry into control liposomes (bottom). Ascorbic acid used to prevent Fe2+ 749 oxidation slightly lowered the external pH, which led to slight proteoliposome acidification. 750 Except for the reporter dye, conditions were identical to those for the metal transport in Figure 1— 751 figure supplement 1D.
34 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
752 753 Figure 3. pH and voltage govern DraNramp transport kinetics. 754 (A-B) Dose-response curves for Cd2+ (A) and Mn2+ (B) transport at different external pH and ΔΨ 755 values. Data are averages ± S.E.M. (n = 3). Errors in KM and Vmax reported in the inset table 756 encompass the uncertainty of the fit (shown as curved lines) to the data. (C-D) Proposed free 757 energy landscapes of Cd2+ (C) or Mn2+ (D) transport based on transport data in (A-B) and Figures 758 1-2.
35 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
759 760 Figure 4. A network of highly conserved charged and protonatable residues adjoins the 761 metal-binding site. 762 (A) Crystal structure (PDB ID: 6BU5) of DraNramp in an outward-facing conformation bound to 763 Mn2+ (magenta sphere). TMs 1, 5, 6, 10 are gold, TMs 2, 7, 11 gray, and TMs 3, 4, 8, 9 blue. (B) 764 D56, N59, M230, and the A53 carbonyl coordinate Mn2+ in the outward-facing state, along with 765 two waters (not shown). (C) View from the plane of the membrane of a network consisting of 766 E134, H232, D131, R353, R352, and E124 that extends ~20 Å from D56 to the cytoplasm. TMs 8 767 and 4, in front of TMs 3 and 9 respectively, were removed for clarity. (D) Sequence logos from an 768 alignment of 6878 Nramp sequences. All 11 mutated residues (numbered above) are highly 769 conserved. Canonical helix-breaking motifs at the metal-binding site are DPGN in TM1 and MPH 770 in TM6. TM6’s H237 is a glycine in many fungal homologs. The second TM9 arginine (R353) 771 varies in location (*) due to an extra helical turn in many homologs; this insertion contains a third 772 arginine in some homologs. See also Figure 4—figure supplement 1.
36 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
773 774 Figure 4—figure supplement 1. Conserved charged and protonatable residues form network 775 leading from metal-binding site to the cytoplasm. 776 (A) View down the metal entry pathway to the conserved network. TMs 10 and 11, which abut 777 TMs 6a and 3, have been removed for clarity. TM6b’s H237 lines the metal exit pathway to the 778 cytoplasm, while the rest of the network lines the proton exit route (Bozzi et al., 2018). (B) 779 Schematic of the conserved residues in the salt-bridge network. (C) Time courses of in vivo Co2+ 780 uptake (averages ± S.E.M.; n = 4). Mutants at most positions retained some Co2+ transport in E. 781 coli, with D56, N59, and H237 mutants most impaired. (D) All constructs express similarly when 782 detected via western blot for the N-terminal His-tag. 783
37 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
784 785 Figure 5. Conserved salt-bridge network confers observed voltage dependence. 786 (A) Average initial metal uptake rates ± S.E.M. (n ≥ 4) for DraNramp mutants at different ΔΨ 787 (colored bars), with the fold increase in transport indicated (cyan, 0 to -120 mV for Cd2+; pink, - 788 40 to -120 mV for Mn2+). Mutations to E124, D131, E134, R352, and R353 (red) greatly reduced 789 the ΔΨ dependence of Cd2+ and Mn2+ transport, such that they exceeded WT at low magnitude ΔΨ 790 but lagged WT at physiological larger magnitude ΔΨ. Mutations to N59, M230, H232, and H237 791 (black) retained ΔΨ dependence, and D56 mutants (grey) eliminated all metal transport. (B) 792 Network schematic illustrates clustering of residues most important for ΔΨ dependence (red). See 793 also Figure 5—figure supplements 1 and 2.
38 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
794 795 Figure 5—figure supplement 1. Sample traces illustrate mutant perturbations to voltage 796 dependence. 797 (A-B) Representative time traces (n ≥ 4) of Cd2+ (A) and Mn2+ (B) uptake into proteoliposomes 798 measured at four ΔΨ values for each mutant. (C) Representative time traces (n > 4) of Cd2+ (left) 799 and Mn2+ (right) uptake showed no metal was imported into control liposomes.
39 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
800 801 Figure 5—figure supplement 2. Mutations to conserved residues perturb voltage dependence 802 of in vivo Co2+ transport. 803 (A) Relative Co2+ transport rates at various extracellular [K+] applied to perturb the endogenous E. 804 coli membrane potential; higher extracellular [K+] presumably leads to a less-negative ΔΨ. 805 Mutations to E124, D131, E134, and R352 reduced the effect of extracellular [K+] more than did 806 mutations to M230, H232, and R353. D56N and N59A mutants were not tested. (B) Time course 807 data used to generate the initial rates shown in (A). All data are averages ± S.E.M. (n ≥ 3).
40 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
808 809 Figure 6. Entire conserved network enables ΔpH stimulation and enforces proton co- 810 transport. 811 (A) Average initial Mn2+ uptake rates ± S.E.M. (n ≥ 4) in proteoliposomes at various external pHs 812 at ΔΨ = -60 mV. (B) Schematic shows proximity to binding site of essential residues for ΔpH 813 stimulation (red), while entire extended network makes some contribution to ΔpH effect (orange). 814 (C) Average initial metal-stimulated H+ transport rates ± S.E.M. (n ≥ 4). (D) Schematic shows that 815 residues in immediate vicinity of D56 are essential for metal-stimulated H+ transport, while the 816 entire network affects metal-proton coupling behavior. Interestingly, D131 and E134 mutants 817 retain some ΔpH stimulation while eliminating metal-stimulated H+ co-transport, while M230A 818 shows the opposite pattern. See also Figure 6—figure supplements 1 and 2.
41 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
819 820 Figure 6—figure supplement 1. Effects of mutations on ΔpH and ΔΨ stimulation of Mn2+ 821 transport. 822 (A) Average initial Mn2+ transport rates ± S.E.M. (n ≥ 4) at various external pHs at ΔΨ = 0. E124, 823 D131, E134, R352, and R353 mutants retained significant Mn2+ transport, but little ΔpH 824 dependence. Even with a ΔpH, H232Q and M230A did not transport Mn2+. (B) Representative 825 time traces (n ≥ 4) of Mn2+ influx at four ΔpH values at ΔΨ = 0 and -60 mV for each mutant. (C) 826 Average initial metal-stimulated H+ transport rates ± S.E.M. (n ≥ 4) at -80 mV show the same 827 trends as at -120 mV in Figure 6C.
42 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
828 829 Figure 6—figure supplement 2. Sample traces show effects of mutations on metal-stimulated 830 H+ co-transport. 831 (A-B) Representative time traces (n ≥ 4) of metal-stimulated H+ transport in proteoliposomes at 832 ΔΨ = -120 mV (A) or -80 mV (B). H+ entry was measured in the absence of metal and in the 833 presence of Mn2+ or Cd2+ for each mutant. (C) No significant H+ entry was observed into control 834 no protein liposomes under any of the tested conditions.
43 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
835 836 Figure 7. Perturbations to salt-bridge network impair high affinity Mn2+ uptake and enable 837 deleterious Mn2+ back transport. 838 (A) Average initial Ca2+ transport rates ± S.E.M. (n ≥ 4). (B) Schematic of mutation locations. 839 M230A, or R353A, 15 Å away, both increased Ca2+ transport. D56 and N59 mutations eliminated 840 Ca2+ uptake, and other mutants were similar to WT. (C-D). Dose-response curves of Cd2+ (C) or 841 Mn2+ (D) transport for a subset of mutants that reduce or eliminate ΔΨ dependence, ΔpH 842 stimulation, or proton co-transport. Data are averages ± S.E.M. (n ≥ 2). The resulting transport 843 kinetic values (middle) show significant overlap for Cd2+ transport but wider separation for the 844 physiological substrate Mn2+. M230A is the only mutation that impaired Cd2+ transport more 2+ 845 severely than Mn transport. Errors in KM and Vmax encompass the uncertainty of the fit to the 846 data. (E) Schematic for isolating metal-transport activity from the ~50% inside-out transporters. 847 Only outside-out transporters are incapacitated by A61C modification with MTSET. (F) Average 848 initial transport rates ± S.E.M. (n = 4) of MTSET-treated or untreated proteoliposomes. MTSET 849 essentially eliminated WT-like’s Mn2+ transport. In contrast, D131N, E134A, and R352A retained 850 significant Mn2+ transport, likely reflecting increased activity from inside-out transporters. This 851 Mn2+ back transport still occurs under physiological-like conditions (ΔΨ > 0 for inside-to-outside 852 transport). In addition, the WT-like protein retained greater residual transport of Cd2+, which did 853 not require H+ symport (Figure 2). See also Figure 7—figure supplement 1.
44 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
854 855 Figure 7—figure supplement 1. Test of DraNramp back transport. 856 (A) Representative time traces (n ≥ 4) of Ca2+ uptake into proteoliposomes. M230A and R353A 857 increased Ca2+ transport, D56A and N59A eliminated Ca2+ uptake, and other mutants behaved 858 comparably to WT. (B) Representative time traces (n = 4) of Ca2+-stimulated H+ uptake into 859 proteoliposomes by M230A. (C) Labeling A61C with positively-charged MTSET likely inhibits 860 DraNramp metal transport by locking the protein in an outward-open conformation. (D) 861 Membrane-impermeable MTSET only modifies A61C of DraNramp molecules oriented outside- 862 out after reconstitution into liposomes. Inside-out DraNramp, which we assume is ~50% of the 863 inserted protein, is protected from modification and thus capable of conformational cycling. (E) 864 Representative metal uptake traces (n = 4) of A61C constructs at ΔΨ = -50 or +50 mV in the 865 presence or absence of MTSET. Interestingly, WT-like A61C/C382S does transport some Mn2+ at 866 ΔΨ ≥ 0, unlike the true WT. Mutations to D131, E134, and R352 all increased Mn2+-transport at 867 +50 mV even after MTSET labeling. Cd2+ transport by WT-like was less dependent on voltage, 868 and showed more residual transport after MTSET labeling. Addition of ionomycin shuttles 869 divalent cations to show the maximum signal. (F) A61C liposome assay mimics cellular context 870 for DraNramp back-transport. Metal influx into MTSET-treated proteoliposomes occurs down a 871 concentration gradient but against a ΔΨ > 0, just as metal efflux would in vivo.
45 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
872 873 Figure 8. Complete mechanistic model for DraNramp transport processes. 874 (A-B) Proposed mechanism of voltage-dependent, metal-proton symport. (Top left) D56 875 protonation optimizes the metal-binding site via hydrogen-bonding of D56 to the N59 carbonyl 876 oxygen, providing a better metal-binding ligand than the amide nitrogen (as shown in B). (Top 877 center) Metal substrate binds, displacing the D56 proton, which passes to D131, with H232 and
46 bioRxiv preprint doi: https://doi.org/10.1101/402412; this version posted August 29, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license.
878 E134 stabilizing a high-energy transition state. (Top right) The proton exits to the cytoplasm via 879 the TM3/TM9 salt-bridge network. In tandem, a bulk conformational change closes the external 880 vestibule and opens the cytoplasmic vestibule. (Bottom right) The metal is released to the 881 cytoplasm. (C-D) Cd2+ uniport perhaps occurs due to a monodentate interaction with D56, enabling 882 proton retention. Bidentate binding of D56 by Mn2+ requires deprotonation, passing the proton to 883 D131. (E-F) Models for canonical secondary transport and DraNramp Mn2+/H+ transport. (E) In 884 canonical secondary transporters, each step of the mechanism is reversible and tight coupling of 885 the co-substrates ensures that the prevailing electrochemical gradients determine the net 886 directionality of transport. (F) In contrast, co-substrate coupling is loose for DraNramp due to the 887 requisite spatial separation of the like-charged Mn2+ and H+ during transport. In addition, the salt- 888 bridge network that imparts voltage dependence and provides the proton-transport pathway 889 enforces unidirectional Mn2+ transport, i.e. the Mn2+-bound inward-to-outward transition is 890 kinetically blocked. See also Figure 8—figure supplement 1.
47 certified bypeerreview)istheauthor/funder,whohasgrantedbioRxivalicensetodisplaypreprintinperpetuity.Itmadeavailableunder bioRxiv preprint
Figure 8—figure supplement 1. Summary of the results presented in Figures 4-6. distance transport H+ H+ / Mn2+ Mutant Residue location ΔΨ dep. ΔpH dep. proposed role(s) to M2+ (Å) Cd2+ Mn2+ Ca2+ uniport co-transport doi: D56A metal-binding site 2.9 X X X NA NA X X metal binding, initial proton binding, pH https://doi.org/10.1101/402412 D56N X X X NA NA X X activation, voltage dependence N59A metal-binding site 3.4 <<< X X = NA = = metal binding metal binding, substrate selectivity, pH M230A metal-binding site 3 << = >> - X = = activation proton transfer, pH activation, conformational H232Q H+-transport pathway 6.2 = << = = X X X change E134A H+-transport pathway 5.3 < < = << <<< X <<< proton transfer, pH activation, voltage E134Q << <<< = < X X X dependence
D131A 11 < < = << << X X ; H+-transport pathway secondary proton binding, voltage sensing this versionpostedAugust29,2018. a D131N < < = <<< << X X CC-BY 4.0Internationallicense R353A H+-transport pathway 14.7 < < > << << > inverted voltage dependence, proton exit R352A H+-transport pathway 17.9 < < = <<< <<< > -- voltage dependence, proton exit E124A 17.7 < < = <<< <<< >> inverted H+-transport pathway voltage dependence, proton exit E124Q < < = <<< <<< >> inverted metal-release H237Q 13.4 <<< X X = = X X conformational change, metal exit pathway X, eliminated; <, less than WT; =, comparable to WT; >; more than WT Distances to M2+ were determined for outward-open Mn2+-bound DraNramp (6BU5)
891 . The copyrightholderforthispreprint(whichwasnot