Figure 1.
Identification of M-box RNAs located upstream of bacterial Nramp-related genes.
(A) The magnesium riboswitch consists of two portions – a divalent-sensing aptamer and downstream sequences which couple the conformational status of the aptamer with formation of an intrinsic transcription termination site. An increase in intracellular magnesium triggers a compacted conformation of the aptamer domain [42] and sequestration of an oligonucleotide tract that would otherwise disrupt terminator formation (“T”). Therefore, increased magnesium promotes transcription termination, repressing downstream gene expression. (B) We searched for instances of putative magnesium riboswitches located upstream of Nramp-related genes and identified many such occurrences. Two were identified for Clostridium acetobutylicum Nramp-related sequences, Ca_c0685 and Ca_c3329, and are shown schematically herein. (C) To determine whether the Ca_c0685 and Ca_c3329 RNA elements were likely to function as magnesium riboswitches, the respective aptamer domains were incubated with varying magnesium and analyzed by analytical ultracentrifugation. Prior studies of the magnesium riboswitch revealed a striking compaction of the aptamer domain in response to magnesium [41]–[42], [49]. The sedimentation velocity measurements of the Ca_c0685 and Ca_c3329 riboswitches revealed an identical compaction with magnesium, suggesting they are likely to function similar to the previously characterized riboswitch.
Figure 2.
Control of gene expression by a C. acetobutylicum M-box RNA.
The Ca_c0685 and Ca_c3329 riboswitches were fused downstream of a constitutive promoter (PrpsD) and upstream of the yellow fluorescent reporter gene (yfp) to determine whether they could control heterologous gene expression in a divalent cation-dependent manner. Control strains either lacking the yfp reporter construct or containing a constitutive PrpsD-yfp fusion were included in this study. Cells were cultured to mid-logarithmic growth phase in 2xYT rich medium supplemented with 50 µM MgCl2 (-), then incubated with 2 mM chelating agent, EDTA, or incubated in the presence of excess magnesium (2 mM) for one hour. Total RNA was assessed by staining of rRNA bands. Abundance of the yfp gene and of a zinc-responsive control transcript were monitored by S1 mapping. Radiolabeled DNA probes (Table S2) were used for S1 mapping of the yfp transcript.
Figure 3.
Metal specificity of C. acetobutylicum riboswitch-yfp reporter fusions.
(A) The strains expressing either the Ca_c0685 riboswitch-yfp or Ca_c3329 riboswitch-yfp reporter fusions were cultured in glucose minimal medium supplemented with 50 µM magnesium until reaching an OD600 of ∼0.5–0.7, at which point 2 mM EDTA was added and cells were incubated for 1 hour. These cells were harvested by centrifugation and the pellet was washed three times and resuspended with an equal volume of chelated glucose minimum medium (chelated with Chelex-100). Either EDTA (2 mM final concentration) or varying magnesium concentrations were added and the cells were incubated for another 1 hour before harvesting. Total RNA was assessed by staining of rRNA bands. Abundance of the yfp gene was monitored by S1 mapping. (B) To assess the specificity of the C. acetobutylicum riboswitch-yfp reporter fusion, we cultured cells expressing either the Ca_c0685 riboswitch-yfp or Ca_c3329 riboswitch-yfp reporter fusions in glucose minimal medium supplemented with 50 µM magnesium and appropriate antibiotics until reaching an OD600 of ∼0.5–0.7, at which point 2 mM EDTA was added for 1 hour. These cells were harvested by centrifugation and the pellet was washed three times and resuspended with an equal volume of chelated glucose minimum medium (chelated with Chelex-100). Either EDTA (2 mM final concentration) or 100 µM various metals were added and the cells were incubated for another 1 hour before harvesting. Radiolabeled DNA probes (Table S2) were used for S1 mapping of the yfp transcript, and for several control transcripts that are known to respond to other metals (e.g., mgtE (magnesium), dhbA (iron), mntH (manganese), mntA (manganese), copZ (copper), yciA (zinc)). These data demonstrate that the Ca_c0685 specifically controls gene expression in response to magnesium, although addition of zinc also resulted in a moderate reduction in gene expression. Shown is a representative gel with quantification derived from experimental triplicates.
Figure 4.
Heterologous expression of Ca_c0685 and Ca_c3329.
The ca_c0685 and ca_c3329 genes were subcloned under IPTG-inducible control and integrated single-copy into the B. subtilis amyE gene. These expression cassettes were also integrated into various strains containing deletions of different divalent cation transporters. In all instances, expression of ca_c0685 and ca_c3329 was monitored by S1 mapping (Figure S1, S2, S3). To investigate the effect of gene expression in these various strains, cells were cultured alongside control strains. Shown herein are bar graphs plots of stationary phase growth for these respective strains. (A) Growth after entry into stationary phase is shown for B. subtilis control strains, including a wild-type and a manganese transport-deficient strain, and transport-deficient strains complemented with IPTG-inducible control of B. subtilis MntH, Ca_c0685 or Ca_c3329. These strains were cultured in minimal medium with no added manganese in the presence of 0.5 mM IPTG. (B) Heterologous expression of B. subtilis MntH and MntABCD do not rescue a magnesium-transport deficient phenotype. Growth measurements immediately after entry into stationary phase are shown for B. subtilis control strains, including wild-type, a magnesium transport-deficient strain, and transport-deficient strains complemented with either IPTG-inducible MntABCD or MntH. Full growth curves are shown in Figure S2. These strains were cultured in rich medium in the presence of 0.5 mM IPTG. (C) Heterologous expression of Ca_c0685 and Ca_c3329 in a magnesium transport-deficient strain. Growth measurements immediately after entry into stationary phase (full growth curves are included in Figure S3) are shown for B. subtilis strains, including wild-type, a magnesium transport-deficient control strain, and transport-deficient strains complemented with inducible Ca_c0685, Ca_c3329, or the magnesium transporter MgtE. The strains were cultured in rich medium in the presence of 0.5 mM IPTG and 2.5 mM magnesium. Expression of Ca_c0685 and Ca_c3329 both fully rescued growth in this medium. (D) In addition to the liquid culture growth experiments, 3 µl of each of these strains (∼1×104/µl) was spotted onto solid medium containing a gradient of magnesium that ranged from 0 to 5 mM magnesium, respectively. These plates were incubated for 10 hours at 37°C before they were photographed.
Figure 5.
Magnesium repression of Ca_c0685 and Ca_c3329 in C. acetobutylicum.
C. acetobutylicum cells were cultured in growth medium to an OD600 of ∼0.8. They were then treated with 2 mM EDTA for one hour, followed by centrifugation and resuspension in magnesium-free growth medium. The cells were then aliquoted into growth medium containing a range of magnesium concentrations and incubated for 2 hours under standard conditions. Abundance of the Ca_c3329 and Ca_c0685 transcripts was then measured by S1 mapping, as shown by the representative experiment herein.
Figure 6.
Bayesian phylogenetic tree of 45 Nramp family transporters, Nramp outgroup proteins, riboswitch-associated outgroup members, and branched-chain amino acid transporters.
The sequences were aligned using MAFFT v7 using the L-INS-I algorithm [75]. This tree is the consensus of four replicate trees constructed using MrBayes 3.2 [76], [77]. Each replicate tree was constructed using four total chains for Metropolis coupling, and 1,000,000 generations with 25% relative burn-in. Priors used were the defaults for amino-acid models with an equal mixture of amino-acid substitution models. Transporter genes preceded by putative magnesium riboswitches are denoted by stem-loops. Branch support values are indicated in red by each internal node. Branch lengths represent expected number of substitutions per position. Similar trees were obtained with different approaches including maximum likelihood (MetaPIGA), minimum evolution, and maximum parsimony (MEGA6), not shown [78], [79].