Characterization of Protein Expression Patterns in Clostridium sporogenes upon Exposure to Rare Earth versus Nonrare Earth Elements

Abstract

To better understand the interactions of rare earth elements (REE) with anaerobic microorganisms, we explored the proteome response of a Clostridium strain to europium (Eu) and its adjacent lanthanide, samarium (Sm), comparing them with the well-studied metals iron, calcium, and zinc. Our findings showed that Eu and Sm induced a number of proteomic changes consistent with Fe, Ca, and Zn, but Eu enhanced expression for ATP synthase and electron transfer flavoproteins, while Sm increased expression of proteins associated with butyric acid metabolism. In contrast, Fe reduced ATP synthase protein expression, and Fe and Ca demonstrated a greater expression change for amino acid metabolism, transcription, and ribosome activity. Differences in ABC transporter protein expressions were observed for the REE. Further characterization of proteins that respond to REE can provide sequence and structural insights to support the design of REE binding molecules or bioinspired systems for REE capture. Given the potential environmental benefits of biological enrichment methods, identifying proteins and metabolic pathways interacting with REE offers an important opportunity to advance the sustainability of supply chains for these metals critical to modern technologies.

Introduction

In recent years, the demand for rare earth elements (REE) has risen dramatically. REE are crucial to producing high-strength permanent magnets, phosphors, and catalysts, making them indispensable in modern technology. However, concerns over the future supply of REE have emerged due to geographically limited production and the potential for geopolitical disruptions to REE supply chains. China currently dominates REE production leading to a strategic vulnerability for many other nations who are consequently striving to develop alternative supply options. A significant bottleneck in REE supply chains lies in the downstream processing of REE. Conventional separation and refining operations are highly resource inefficient and polluting, , highlighting the need for innovative extraction approaches that are sustainable, cost-effective, and environmentally friendly. This has led to increased interest in biological or biologically inspired approaches for REE recovery.

Until the past decade, rare earths were considered biologically inert. In 2014 however, the first report was made of microorganisms that actually utilized lanthanides for a metabolic function. Since that first report by Pol and coworkers of a methane oxidizing bacterium isolated from a volcanic mudpot and its lanthanide-dependent methanol dehydrogenase, several other research groups have described lanthanide-dependent organisms and or associated biomolecules with affinity for REE. ,, Most of the associated metabolic pathways have been linked with the oxidation of methanol or short-chain alcohols. The identification of specific binding proteins and peptide motifs associated with such pathways has spurred vigorous efforts to utilize them for REE extraction and separation. In 2019, Maleke and coworkers reported a different kind of microbial physiology linked with REE. They described a Clostridium strain (sp. 2611) that reduced trivalent europium (Eu³⁺) to its divalent form (Eu²⁺) and observed that the bacterium accumulated Eu intracellularly. Selective Eu bioreduction and accumulation offers the potential to exploit another set of biologically produced or biologically inspired ligands for selective rare earth recovery. To leverage this interaction, it is crucial to identify the proteins that respond to the presence of Eu and determine which proteins directly bind to Eu.

Previous studies of rare earth interactions with microorganisms have identified several Eu-binding proteins, including phosphotriesterase, bacteriohodopsin, streptavidin (STN-NE113), Mex-LanM, Lanpepsy, and a 12 amino acid EF hand fold of calmodulin within prokaryotic proteomes. As the compilation of known REE binding proteins, the lanthanome, increases so too do the opportunities to computationally identify structural moieties involved in REE-protein binding. If obligate anaerobes like Clostridium can directly reduce Eu³⁺ to Eu^2+, there should be specific proteins involved in the bioreduction, and proteins identified as such can be screened as candidates for selective REE binding using computational methods or experimentally. Protein motifs already known to bind metals are a good place to start looking.

In this study, we investigated the differential expression of proteins by Clostridium sporogenes when exposed to europium (Eu), samarium (Sm), iron (Fe), zinc (Zn) or calcium (Ca). Phylogenetically, C. sporogenes (strain NCA 3679) is closely related to the Eu-reducing strain Clostridium sp. 2611 described by Maleke and coworkers (Figure S1). A preliminary proteomics study in our laboratory determined that Fe, Zn, and Ca binding proteins were differentially expressed when C. sporogenes cells were cultured with 6 μM or 2000 μM Eu (data not shown). This finding prompted us to design a follow-up study to specifically compare the cellular response to Eu with the responses to Fe, Zn, Ca and Sm using Tandem Mass Tag (TMT) mass spectrometry (MS) (Figure ). Considering that Fe³⁺ can be reduced to Fe²⁺ by many organisms, including C. sporogenes, we chose Fe³⁺ as a reducible positive control for our study and wanted to compare if the proteome response to Fe³⁺ would be similar to the response to Eu³⁺. We included Sm as a nonreducible negative control because it is more difficult to reduce than its adjacent lanthanide neighbor Eu. The reducing conditions necessary for stable Sm²⁺ are outside of the stability field of water and therefore expected to be incompatible with biological systems. , We postulated that proteins that responded similarly to Eu³⁺ and Sm³⁺ might both be associated with REE binding, but proteins involved with Eu bioreduction would be distinct.

1.

Experimental overview. C. sporogenes cultures incubated with metals were assayed by mass spectrometry for changes in protein expression to identify metal-responsive molecular processes using bioinformatics (photos and images used are taken or created by the authors using bioRender software).

Results and Discussion

Cell Growth Indicators of C. sporogenes Cultured with Eu, Sm, Fe, Zn, or Ca

To assess the impact of Eu, Sm, Fe, Zn, and Ca on cellular activity, we quantified anaerobic cell growth indicators (e.g., cell pellet weight, gas production) after incubating C. sporogenes with each metal for 15 h. We chose this time point, early in the exponential growth phase, to maintain a relevant ratio of metal to cell number. As cell numbers increase, the ratio of metal to cell number decreases, effectively reducing average metal exposure and thereby the impact on average protein expression. The concentration of 500 μM was chosen as our lab had previously demonstrated that 500 μM was high enough for a cellular response but would not be toxic to the microorganism. Concentrations in the 100–1000 μM range are common in REE-rich industrial leachates and biosorption systems, making 500 μM relevant for biorecovery applications. Lastly, we chose to analyze both H₂ and CO₂ as each gas is produced by different pathways of cellular metabolism.

The trends for production of H₂ and CO₂ by the microorganism differed for the added metals (Figure ). At the 9 h mark, Eu, Sm, and Ca treatments exhibited statistically significant increased gas production compared to the control. After 9 h, the gas production trends diverged: the control and Eu treatment showed an increase in the rate of gas production rate compared to Fe, Ca, and Zn while Sm treatment exhibited slower gas production rate for H₂ but consistency for CO₂ production. By the 15 h mark, the Eu treatment showed significantly increased H₂ and CO₂ production compared to the control, Ca, and Zn treatments (Table ). There was no significant difference in cell pellet weight between the Eu treatment and the control (p-value 0.20) (Table ). However, the cell pellet weight for the Eu treatment was significantly higher than for the Ca and Zn treatments, but not for the Sm or Fe treatment. (Table S1).

2.

Percent H₂ and CO₂ in the headspace. The gases H₂ (A) and CO₂ (B) were measured though the first 15 h of C. sporogenes cultured with 500 μM of each metal (Eu, Fe, Zn, Ca, or Sm) or a no metal control.

1. Indicators of Cellular Activity .

Treatment	H2 (%)9 h	H2 (%)15 h	CO2 (%) 9 h	CO2 (%) 15 h	Cell pellet weight (g)
C. Sporogenes control	2.4 ± 0.1	8.9 ± 0.4	2.6 ± 0.1	11.6 ± 0.6	0.44 ± 0.11
C. Sporogenes + 0.5 mM Eu	3.9 ± 0.4	11.2 ± 0.4	4.0 ± 0.5	16.1 ± 0.8	0.34 ± 0.01
C. Sporogenes + 0.5 mM Sm	6.1 ± 0.5	10.2 ± 0.3	7.0 ± 0.8	15.0 ± 0.6	0.30 ± 0.02
C. Sporogenes + 0.5 mM Fe	3.5 ± 1.2	9.7 ± 1.0	4.0 ± 1.3	13.2 ± 1.8	0.30 ± 0.02
C. Sporogenes + 0.5 mM Zn	3.2 ± 0.1	8.4 ± 0.2	3.4 ± 0.0	10.2 ± 0.1	0.28 ± 0.02
C. Sporogenes + 0.5 mM Ca	3.8 ± 0.4	8.9 ± 0.4	4.0 ± 0.8	11.8 ± 0.6	0.28 ± 0.01

^aThe percentage of gas in the headspace at 9 and 15 h and cell pellet weight measured at 15 h of incubation with different metals.

^bIndicates statistical difference from the control (Student’s t test, p < 0.05).

Global Protein Expression of C. Sporogenes Cultured with Eu, Sm, Fe, Zn, or Ca

The TMT assays identified 9,656 peptides corresponding to 2,469 TMT-labeled proteins across all conditions, with a false discovery rate (FDR) less than 0.05 (Table S2). To visualize the protein response for each metal, a principal component analysis (PCA) was performed (Figure ). The PCA grouped the protein responses for Fe, Zn, and Ca into the same quadrant, while the control, Eu, and Sm responses each resided in their own quadrants. Component 1 accounted for 66.6% of the variation, and Component 2 accounted for 14.5%. When the individual proteins in the PCA were visualized, each quadrant had a near equal distribution of proteins, indicating that the protein responses are diverse and not dominated by a few highly expressed proteins. This balance suggests a broad impact from each metal on the proteome.

3.

PCA analysis of the 2,469 TMT labeled proteins identified after exposure to metals Eu, Sm, Fe, Zn, Ca, or the no-metal control C. (3A) PCA for each metal and (3B) displays of the individual proteins.

To visualize differences in protein expressions in response to Eu and the other metals we generated a heat map of the 2,469 proteins quantified in this study. To focus on the differences between the Eu treatment and the other conditions, the protein abundance for each metal or control was subtracted from the protein abundance for the Eu treatment (Figure ). From this heatmap, it was observed that exposure to Eu and Sm resulted in the most similar expression changes as indicated by the column with the most black lines. Hierarchical clustering also indicated that the relative (compared to Eu treatment) patterns of protein expression for Fe and Ca treatment were similar to each other, suggesting a common cellular response to these metals.

4.

Heat map of 2,469 quantified proteins. Protein abundances are displayed as abundance difference relative to Eu-induced protein abundance. Protein abundance for a given metal or the control was subtracted from the protein abundance in response to Eu. A negative value (green) represents a lower Eu-induced protein abundance compared to the other conditions. Conversely, red represents a greater Eu-induced protein abundance compared to the other conditions. Black represents no difference between Eu-induced abundance and protein abundance from the other conditions.

Differential Protein Expression of C. Sporogenes Cultured with Eu, Sm, Fe, Zn, and Ca

To define our differentially expressed proteins (DEPs) we chose a 1.5-fold expression change cutoff and statistical cutoff using the false discovery rate (FDR) of <0.05 for metal induced protein expression changes compared to the protein expression of the control. Table displays the number of DEPs that were up-regulated or down-regulated compared to the control for each metal. These DEP tallies revealed that Ca and Fe generated a more profound cellular response than the REEs or Zn as indicated by the total number of DEPs. To visualize the overlap of DEPs were up- or down-regulated for each metal, we generated a Venn diagram (Figure ). The Venn diagram highlighted that Eu and Sm share more overlap with each other than they do with the other metals, similar to the expression patterns observed in the heatmap. The full list of protein expression changes can be found in Table S1. Table S1 includes all identified proteins and their metal cofactors. Potential REE-binding proteins can be computationally identified from the REE-induced differentially expressed proteins listed in Table S1 This data set provides a basis for identifying sequence and structural conservation compared to known REE-binding proteins characterized by Sajeevan et al. Additionally, the sequences and structures characterized by Sajeevan et al. could be used to probe metal-binding and REE DEPs in this data set and others. Other investigations have identified peptide sequences that display an affinity for REEs and for Eu³⁺ specifically. These diverse methodologies collectively advance the field of protein-directed REE accumulation, offering new avenues for computational and biotechnology approaches.

2. Total Number of Proteins Upregulated or Downregulated in Response to a Given Metal Compared to the No-Metal Control.

Metal	Upregulated	Downregulated
Eu	74	76
Sm	79	77
Zn	21	21
Ca	461	411
Fe	465	417

5.

Venn diagrams depicting overlap of upregulated or downregulated DEPs. The overlapping DEPs with upregulated protein expression (A) and downregulated protein expression (B).

Bioinformatic Analysis of Differentially Expressed Proteins

The network of protein–protein interactions and gene ontology (GO) enrichments was determined by submitting the DEPs to the search tool for the retrieval of interacting genes/proteins (STRING) database. Functional networks for both up-regulated and down-regulated DEPs were displayed together for Eu, Sm, and Zn. However, for Fe and Ca, the STRING analysis was separated by up-regulated and down-regulated DEPs due to the number of DEPs. GO annotations were obtained by submitting only the up-regulated or down-regulated proteins for each metal to the STRING database (Figure S2–S16). Notably, no GO annotation was produced for the DEPs in response to Zn which indicates that Zn did not induce protein expression changes for multiple proteins involved in the same biological processes, unlike the other metals.

STRING analysis was used to organize protein functions into groups and display the number of proteins up- or down-regulated for a given function. The patterns created by STRING distinguished between the specific processes of ATP synthase, electron transfer, amino acid metabolism, and ABC transporters. The differences in protein functions for the DEPs were also characterized using GO. The STRING and GO bioinformatic tools seek to group proteins by their functions but each tool uses different grouping algorithms. Using both of these methods in addition to grouping proteins by fold change (Table S1), we identified key differences in protein expression for the processes as described in the following sections.

Effects on ATP Synthase and Cellular Energy Production

The expression of ATP synthase proteins varies significantly in response to different metals. ATP synthase, a conserved multiprotein enzyme that regenerates ATP from ADP, works alongside the Stickland reactions, which generate substrate-level ATP and other metabolic byproducts. The Stickland reactions involve pairing amino acids, with one acting as the electron donor and the other as the electron acceptor. These redox reactions serve as a primary metabolic driver in Clostridia species, converting amino acids into ATP, CO₂, H₂, secondary metabolites such as butyrate ,− and nicotinamide adenine dinucleotide (NADH), which feeds into ATP synthase. , NADH from carbohydrate metabolism and other metabolites also support ATP synthase activity. Our data demonstrated that Eu significantly increased the expression of seven ATP synthase proteins by at least 50% whereas Fe and Ca increased expression three and decreased six (Table ). Zn had no significant impact on expression of ATP synthase proteins.

3. ATP Synthase Protein Expression Changes in Response to Different Metals .

Gene	C–Eu	C–Sm	C–Zn	C–Fe	C–Ca	Function
atpH	–0.930	–0.825	–0.445	4.019	3.808	ATP synthase subunit delta
atpD	–0.760	–0.614	–0.286	9.953	9.999	V-type ATP synthase subunit D
atpC	–0.723	–0.590	–0.190	2.852	2.898	ATP synthase epsilon chain
atpG	–0.711	–0.452	–0.178	–2.451	–2.622	ATP synthase gamma chain
atpE	–0.696	–0.583	–0.323	5.190	5.291	ATP synthase subunit c
atpB	–0.660	–0.452	–0.209	–1.549	–1.597	ATP synthase subunit a
atpA	–0.611	–0.476	–0.065	6.153	–2.895	ATP synthase subunit alpha
atpF	–0.270	–0.207	0.027	-3.312	–3.384	ATP synthase subunit b
NPD5_3311	–0.177	–0.057	0.115	6.246	6.173	ATP synthase (C/AC39) subunit
atpB2	–0.166	–0.116	0.078	5.529	5.766	V-type ATP synthase beta chain

^aNegative values represent protein expression that was increased relative to the control while positive values were decreased relative to the control. Values are expressed as Log₂ fold change. Bold font indicates protein expression meets threshold of 1.5-fold change (Log₂ = 0.58).

The differential expression levels of ATP synthase subunits in response to the same metal can be attributed to two primary mechanisms: direct binding and indirect binding. Specific subunits of ATP synthase, such as the α subunits of the F1 domain, have the capability to bind metal ions, leading to conformational changes or functional modifications in the ATP synthase complex. Additionally, metals can influence gene expression indirectly by affecting various cellular pathways. For example, metals like iron and calcium play complex roles in regulatory pathways or stress responses within the cell. In contrast, Eu and Sm consistently increased the expression of all ATP synthase subunits, suggesting a more uniform response potentially linked to their interaction with specific regulatory mechanisms that promote ATP synthase expression.

The electron transfer flavoprotein (ETF) plays a crucial role in Clostridium metabolism, acting as a redox center that both accepts and donates electrons across multiple metabolic pathways. A key feature of ETF is electron bifurcation, a mechanism that couples exergonic (energy-releasing) and endergonic (energy-consuming) redox reactions, enabling the simultaneous generation or utilization of low- and high-potential electrons. The EtfAB/Bcd complex, which comprises the electron transfer flavoprotein (EtfAB) and butyryl-CoA dehydrogenase (Bcd), facilitates the transfer of electrons from NADH to various substrates, including ferredoxin and butyrate. − This complex utilizes multiple flavin adenine dinucleotides (FADs) to mediate electron transfer through a series of redox reactions. By doing so, this intricate mechanism ensures efficient energy extraction and conservation within the cell, allowing NADH to receive electrons indirectly from ETF and integrating it into the broader metabolic network.

The balance between Stickland reactions and ETF promotes effective energy extraction from substrates and the generation of secondary metabolites, with hydrogen (H₂) and carbon dioxide (CO₂) serving as byproducts. The production of H₂ and CO₂, which can be measured externally, serves as indicators of cellular activity. In our experiments, both H₂ and CO₂ production were elevated in the presence of Eu and Sm compared to the control (Table ). Additionally, Eu and Sm significantly enhanced ATP synthase protein expression (Table ). Notably, both metals induced similar patterns of increased and decreased expression of ETF proteins (Table ); if a particular protein abundance increased (or decreased) with Eu, it also exhibited similar changes with Sm. These ETF protein expression changes align with Maleke’s findings of increased protein abundance of flavooxidoreductase in Eu-cultured cells. Maleke hypothesized that cytosolic flavooxidoreductases were responsible for the reduction of Eu, rather than membrane-bound flavooxidoreductases. Here, we identify ETF proteins as the primary subclass of flavooxidoreductases influenced by REEs. Interestingly, Sm, which is considered nonbiologically reducible due to its high reduction potential, affected ETF proteins similarly to Eu.

4. Electron Transfer Protein Expression Changes in Response to Different Metals .

Gene	C–Eu	C–Sm	C–Zn	C–Fe	C–Ca	Co factor	Function
P9J83_16070	–1.361	–1.576	0.027	1.219	1.486	FAD	Electron transfer flavoprotein subunit alpha/FixB family protein
P9J83_15250	–1.205	–1.174	–0.659	3.399	3.332		Fe-only hydrogenase system regulator
CGS26_11655	–1.151	–1.452	0.122	7.389	7.627	FAD	Electron transfer flavodomain protein
NPD5_3744	–1.010	–1.059	–0.049	4.981	5.042		Electron transfer flavoprotein small subunit
etfB1	–0.982	–1.021	0.066	16.216	16.535	4Fe-4S, FAD	Electron transfer flavoprotein small subunit
etfA3	–0.363	–0.457	–0.394	–5.248	–5.324	4Fe-4S, FAD	Electron transfer flavoprotein subunit alpha
P9J83_16925	–0.342	–0.385	–0.111	–1.926	–2.031		Electron transfer flavoprotein small subunit
CGS26_05300	–0.332	–0.437	–0.141	–3.870	–3.830	FAD	Electron transfer flavoprotein subunit alpha
NPD5_1046	–0.313	–0.312	–0.122	–0.499	–0.505	FAD	Electron transfer flavodomain protein
etfB3	–0.256	–0.220	–0.195	2.011	1.879		Electron transfer flavoprotein small subunit
BS638_12170	–0.242	–0.219	0.055	–0.036	0.031		Electron transfer flavoprotein subunit alpha
NPD5_2717	0.281	0.293	–0.090	–0.085	–0.062	FAD	Electron transfer flavodomain protein
etfA2	0.671	0.854	–0.046	0.120	0.094		Electron transfer flavoprotein subunit alpha
etfB2	0.817	0.772	0.111	0.117	0.092		Electron transfer flavoprotein small subunit
NPD5_2716	0.881	0.871	0.066	0.189	0.082		Electron transfer flavoprotein small subunit
P9J83_14015	0.890	0.868	0.038	0.127	0.025	FAD	Electron transfer flavoprotein subunit alpha/FixB family protein

^aNegative values represent protein expression that was increased relative to the control while positive values were decreased relative to the control. Values are expressed as Log₂ fold change. Bold font indicates protein expression meets threshold of 1.5-fold change (Log₂ = 0.58).

A notable distinction between Eu and Sm was that the GO enrichment from Sm elicited greater activity of “butyrate metabolic processes and electron transporter activity” (Figure S6–S7). This finding highlighted that while both Eu and Sm induce similar expression patterns for ATP synthase and ETF proteins, the proteins associated with metabolic flux for each metal differ. Our data emphasized how different metals influence broad cellular pathways, presenting unique opportunities for biological REE recovery. Specifically, Eu and Sm increased the activity of ATP synthase and ETF proteins, with Eu exhibiting slightly higher expression for ATP synthase, while Sm demonstrated greater connectivity between butyrate and electron transfer activity (Figure S6–S7). In C. sporogenes, butyrate is primarily formed via Stickland reactions, unlike other Clostridium species where it is NADH-dependent from acetate. Furthermore, Eu and Sm elicit different responses in amino acid metabolism and ABC transporters, suggesting element-specific metabolic pathways for selective REE recovery. Potentiometric studies have measured Sm³⁺ in stable complexes with amino acids and peptides without detection of Sm²⁺ states.

Both Eu and Ca influenced butyrate metabolism proteins, though to a lesser extent than Sm, while Fe decreased butyrate-related protein expression, consistent with observations in C. difficile. Ca and Fe also increased the expression of transporters for cadmium, magnesium, and Ca (Figure S9, S13), with Ca additionally affecting copper and drug export proteins (Figure S13). More research is needed to understand how REEs influence cellular activity through direct metal-protein interactions or indirect mechanisms such as metal displacement, allosteric binding, or toxicity resistance.

In contrast, Fe and Ca had markedly different impacts on ATP synthase and ETF proteins compared to Eu and Sm. Specifically, the differential expression patterns (DEPs) for ETF proteins showed opposing trends for Eu and Sm in contrast to Fe or Ca for most proteins. For some ETF proteins, differential expressions were observed exclusively for either Eu/Sm or Fe/Ca, but not both (Table ). Notably, divergent ATP synthase protein expression patterns were observed between Fe and Ca; for example, Fe decreased atpA expression, while Ca increased it (Table ). We propose that the changes in ATP synthase subunit expression in response to Fe or Ca are attributed to the broad regulatory roles these metals play in Clostridium metabolism. This is supported by observations that Clostridium can continue cell growth even after deletion of atpG or atpA. Lastly, Zn induced differential protein expression for only one ETF protein and had no significant influence on other ETF or ATP synthase proteins. The marginal changes in DEPs and the reduction of CO2 gas production compared to the control at 15 h are consistent with reports indicating that the addition of Zn to C. sporogenes reduced cell viability. By comparing the impact of each metal on protein expression for ATP synthase, ETF, amino acid metabolism, and gas production, we find evidence that Eu and Sm impose a greater demand on electron transfer proteins like ETF and ATP synthase, while Fe and Ca exert a stronger collective influence on Stickland reactions, as discussed in detail below.

Influence on Amino Acid Metabolism

Eu, Sm, Fe, and Ca influenced the expression of proteins involved in metabolism of amino acids. As noted earlier, the Stickland reaction, central to the metabolism of many Clostridium species, involves the pairing of amino acids where one serves as an electron donor and the other as an acceptor. Typical electron acceptors include glycine, asparagine, and proline, while other amino acids such as leucine, phenylalanine, tyrosine, tryptophan, and arginine can function as either donors or acceptors. The remaining amino acids primarily act as electron donors. By modulating which amino acid is the acceptor and which is the donor, the cell can regulate energy production, secondary metabolites, and drive the metabolic state of the cell toward growth. ,

Shifts in amino acid pairs have been observed in several Clostridium species. For instance, in C. sticklandii, different growth phases and oxidative stress can influence amino acid pairs. In C. difficile cell metabolism, toxin production and growth are driven by modulating proline reductase, which functions in proline and arginine metabolism. In C. sporogenes, acetate availability altered which amino acid pairs contributed to cell growth. Additionally, Stickland activity byproducts such as butyrate can undergo redox reactions with amino acids to regenerate NAD+ in C. sticklandii.

Each metal influenced protein expression related to amino acid metabolism. When Eu was added to C. sporogenes, protein expression decreased compared to the control for glycine (acceptor), aspartate (donor), valine (donor), leucine (donor), and isoleucine (donor) metabolic activities (Figure S2–S4). For Sm, Fe, and Ca, GO enrichment identified both increased and decreased protein expression for amino acid metabolism. Sm decreased expression of proteins regulating aspartate (donor), glutamate (donor), glycine (acceptor), valine (donor), leucine (donor), and isoleucine (donor), while glutamate catabolism increased (Figure S5–S7). Fe increased the expression of proteins for biosynthesis of lysine (donor) and tyrosine (donor/acceptor) but decreased expression for metabolism of glycine (acceptor), cysteine (donor), and methionine (donor) (Figure S10–S12). Ca increased aspartate (donor) and lysine (donor) activities while decreasing proline (acceptor), arginine (donor/acceptor), glycine (acceptor), serine (donor), threonine (donor), cysteine (donor), and methionine (donor) (Figure S13–S16). Zn increased the protein expression of two amino acid proteins, one of which functions in phenylalanine fermentation (fldZ) while the other is an alanine t-RNA (alaS) (Figure S8). In addition to amino acid metabolism, Fe and Ca also altered protein expression related to transcription, ribosome activity, tRNAs, and secondary metabolites (Figure S9–S16), with many of these proteins directly binding metals (Table S1). This broad regulatory impact of Fe and Ca on Clostridium beyond amino acid metabolism is well-supported by previous studies. − In summary, each metal distinctly influenced amino acid metabolism.

Transportation of Metals

Fe is a vital micronutrient essential for DNA synthesis, respiration, and metabolism. In Clostridiumspecies, one mode of Fe acquisition is mediated by ABC transporters consisting of a periplasmic binding protein, a membrane-spanning domain, and an ATPase component. ATP-binding cassette (ABC) transporters are a large family of proteins that utilize the energy derived from ATP hydrolysis to transport various substrates across cellular membranes. These transporters are found in all domains of life and play crucial roles in maintaining cellular homeostasis, nutrient uptake, and the efflux of toxic compounds. These transporters exhibit high-affinity binding to Fe, ensuring efficient uptake even in Fe-limited environments. The ability to secure Fe efficiently is important given the competitive nature of microbial communities and the host’s Fe-sequestration strategies during infection or in soil communities. Like Fe, Ca is involved in a wide range of cellular processes such as signal transduction, enzyme coordination, and chemotaxis. , In Clostridium, ABC transporters facilitate the uptake of Ca ions, maintaining intracellular Ca homeostasis. These transporters are critical for processes like sporulation, motility, and metabolic regulation, which are essential for the survival and adaptability of Clostridium species under various environmental conditions. While some metal interactions in C. sporogenes have been explored, potential mechanisms for REE accumulation are poorly understood.

In our study, we identified 104 ABC transporters (Table S2). Among these, there were 36 DEPs in response to Fe, 37 DEPs from Ca, 3 DEPs from Zn, 15 DEPs from Sm, and 12 DEPs from Eu. ABC transport of Eu is a potential mechanism to support the observation made by Maleke 2019 that Eu was accumulated intracellularly. While ABC transporters were the most common transport system identified in our data set, differential expressions of other transport proteins were identified in response to all metals except Zn. As observed in each STRING analysis, non-ABC transport proteins showed increased and decreased expression. While ABC transporters are known to transport metals across the cell membrane, they may not be the only membrane transport system that could accumulate REEs intracellularly. Consistently, ABC transporters and other bacterial membrane transport systems were increased when REE gadolinium­(III) was introduced to a microbial community thereby suggesting a common theme for microorganisms exposed to REEs. However, the increased expression of ABC transporters is not a direct indicator of increased metal accumulation. As an environmental example, some high affinity zinc ABC transporters only become active under low Zn conditions.

Conclusion

This work provides the first detailed proteomic characterization of Clostridium sporogenes in response to exposure to europium (Eu) and samarium (Sm), two REE with distinct redox properties, alongside physiologically relevant metals (Fe, Ca, Zn). Comparative proteomics revealed that Eu and Sm share a core response involving coordinated upregulation of ATP synthase and electron transfer flavoproteins, pointing to increased energetic and redox demands under REE exposure. Subtle but important differences were also evident as Sm elicited a stronger association between butyrate metabolism and electron transport, whereas Eu showed slightly higher ATP synthase induction, suggesting REE-specific metabolic routing even among closely related lanthanides. In contrast, Fe and Ca triggered broader proteomic shifts, including pronounced effects on amino acid metabolism, transcription, ribosome activity, and ABC transporters, underscoring that cellular adaptation to REEs is distinct from responses to more common metals. Notably, Eu and Sm induced changes in ABC transporter expression patterns consistent with potential roles in REE uptake or intracellular sequestration, complementing prior reports of Eu bioreduction and accumulation in related Clostridia. These results, together with the inclusion of metal-binding annotations and cofactor information for all identified proteins, provide a rich candidate set for computational screening and structural analysis aimed at pinpointing novel REE-binding motifs. We acknowledge that Eu and Sm can act as Ca surrogates and may compete for Ca-binding sites. This competition could potentially influence the uptake and treatment responses of Eu and Sm. However, our experimental design included a control group with no added metals except what was inherent in the media, allowing us to compare the effects of Eu and Sm directly against a baseline of 10 mg CaCl₂·2H₂O in the medium. While calcium could theoretically compete with Eu and Sm for binding sites, our data suggests that the observed protein expression changes were primarily driven by the presence of Eu and Sm. Furthermore, the specific differences in protein expression between Eu and Sm compared to the control or to Ca addition suggest that despite Ca competition, Eu and Sm still bound to proteins and affected cellular activity.

It is acknowledged that the current study did not include direct intracellular localization analyses (e.g., TEM-EDS, ICP-MS), and this is recognized as a limitation. Challenges such as the instability of Eu­(II) under sample preparation conditions, strong adsorption of REEs to cell surfaces, and matrix interference in ICP-MS ionization complicate reliable intracellular quantification without extensive protocol optimization. Additionally, the experimental confirmation of Eu-reduction would require redox-state-specific spectroscopy (e.g., XANES, EPR, Mössbauer) conducted under strictly anoxic conditions. These analytical constraints prevent reliable “intracellular-only” data without extensive protocol optimization. Our proteomic results, however, do indicate biologically meaningful interaction between Eu/Sm and C. sporogenes, including altered expression of electron-transfer proteins, membrane transporters, and ATP synthase subunits. These effects suggest cellular sensing or stress response but do not claim intracellular uptake.

Moreover, strain NCA 3679 belongs to Clostridium Cluster I, the same broad clade as the Eu-reducing Clostridium sp. 2611; however, they are phylogenetically distinct isolates, and REE-reducing traits cannot be inferred from shared lineage alone. Although proteomic changes observed here indicate that NCA 3679 senses and responds to Eu and Sm exposure, the present study does not directly evaluate Eu³⁺ → Eu²⁺ reduction. This limitation will be pursued in future studies.

Collectively, our analyses of protein expression changes induced by REE exposure, combined with knowledge of known REE-binding proteins or peptides, could identify novel REE-binding motifs. Understanding these motifs can support the development of innovative bioinspired methods for accumulating, concentrating, and separating individual REEs. Future research should focus on experimentally validating putative REE-binding proteins and exploring their practical applications for more sustainable rare earth production. By linking these molecular signatures to the broader project goal of understanding microbial REE bioreduction mechanisms, this study delivers a foundation for targeted experimental validation, genetic engineering, and synthetic biology strategies. Such approaches could enable the design of biologically inspired systems for selective REE concentration and separation, offering a sustainable, low-waste alternative to conventional extraction technologies. The insights gained here not only advance fundamental knowledge of REE–microbe interactions but also strengthen the path toward environmentally responsible rare earth recovery, critical for securing resilient supply chains for emerging energy and technology sectors.

Methods

Microbial Cultivation and Media Preparation

The following metals were used in this study: Eu (III) chloride hexahydrate, Sm­(III) chloride hexahydrate, Fe­(III) chloride, Zn chloride, Ca chloride dihydrate. Clostridium sporogenes (NCA 3679) was purchased from ATCC. The bacteria were cultivated in DSMZ 104b medium. The medium (per L: 5 g trypticase peptone, 5 g peptone from meat, pepsin–digested, 10 g yeast extract, 10 mg CaCl₂-2H₂O, 20 mg MgSO₄-7H₂O, 40 mg each K₂HPO₄ and KH₂PO₄, 400 mg NaHCO₃, 80 mg NaCl) was boiled for 15 min under a stream of nitrogen, prior to glucose (5 g/L) and cysteine (0.5 g/L) addition. The medium pH was 7.0. All glassware was washed with 0.1 M HCl and extensively rinsed with nanopure water. After transferring the medium (50 mL) into serum bottles (volume 165 mL), the headspace (115 mL) was flushed with nitrogen gas for an additional 5 min. The bottles were then closed with butyl rubber stoppers, equilibrated with N₂ to 1 atm internal pressure, and crimped with aluminum seals. The prepared bottles were autoclaved for 20 min and cooled prior to inoculation.

C. sporogenes Cultivation with REE and Ions

Metals were added to DSMZ 104b medium to a concentration of 500 μM. The REE and metals were first prepared anaerobically as individual stock solutions at 50 mM in nanopure water, then autoclaved (except for the ferric chloride stock, which was filter sterilized). They were then added to the media using sterile anaerobic techniques. Experiments were initiated by the addition of 0.5 mL (∼1% volume inoculation) of C. sporogenes cultures previously grown mid- to late-log. Triplicate bottles for each condition, including biotic control, were prepared, and incubated at 38 °C. Hydrogen and carbon dioxide production was measured using a gastight syringe to remove 7 mL of gas from the headspace for analysis.

Hydrogen, Carbon Dioxide, and Ion Measurements

Headspace was measured by gas chromatography, using a Molsieve 5Å column and argon carrier gas with a thermal conductivity detector (Agilent 490 Micro GC; Agilent Technologies); calibration for hydrogen measurements were performed using standard gas mixtures containing 0.5%, 1% and 15% H₂ with N₂ balance. Calibration for CO₂ measurements were calibrated using 1%, 5%, 10% and 15% CO₂ with N₂ balance. REE concentrations were measured using inductively coupled plasma optical emission spectrometry (ICP-OES; PerkinElmer 8300) or Agilent 7900 ICP-MS in accordance with manufacturer instructions. Commercial standard stock solutions and samples were acidified with ultrapure concentrated nitric acid to a concentration of 1% (v/v).

C. sporogenes Whole Cell Extract Preparation

After 15 h of growth, cultures were poured into preweighed tubes and pelleted (15 min, 5500 rcf, 4 °C) and weighed after supernatant decantation. The cells were then resuspended to 1 g using phosphate buffer saline (PBS) pH 7.4 with protease inhibitors (40 mM 2-Iodoacetamide, 20 mM N-ethylmaleimide, and 1:500 stock Sigma PIC) and 5% sodium dodecyl sulfate (SDS). The cell slurry was transferred one drop at a time into tubes half-full of liquid nitrogen then kept at −80 °C until cryo-milling while submerged in liquid nitrogen using a rotary mill. Cells were mechanically fractured for 3 cycles at 1 min per cycle then stored at −80 °C. Cells were resuspended in ice-cold PBS with protease inhibitors and SDS and slurry was clarified by centrifugation (30 min, 14,435 rcf, 4 °C). Protein concentration was determined by bicinchoninic acid (BCA) protein quantification and samples were kept frozen at −80 °C until shipped overnight to Michigan State University’s Mass Spectrometry and Metabolomics Core for protein digestion and analysis by MS.

Proteolytic Digestion

An aliquot equivalent to 60 μg was taken from each sample and applied to S-trap devices (www.protifi.com) for tryptic digestion according to manufacturers’ instructions. Following digestion, eluates from S-traps were dried by vacuum centrifugation and stored at −20 °C.

Isobaric Peptide Labeling

Peptide samples (∼60 μg each) were suspended in 100 μL of 100 mM TEAB and labeled with TMTpro18 reagents from Thermo Scientific according to manufacturers’ instructions. Aliquots of 2 μL were taken from each labeled sample and used to test labeling and mixing efficiency by MS. The remaining labeled peptides were mixed in equal portions dried by vacuum centrifugation and resuspended in 2% acetonitrile/1% trifluoroacetic acid, desalted using a Waters c18 SepPak, dried by vacuum centrifugation and stored at −20C.

Fractionation

Peptide sample was suspended in 2% acetonitrile/0.1% trifluoroacetic acid to 50 μL and injected onto a Waters Acquity UPLC BEH 2.1 mm × 100 mm c18 column using a Waters Acquity H-class UPLC. Bound peptides were washed using 0.1% trifluoroacetic acid in water for 2 min and then separated over 65 min using a gradient of 2%B to 40% in 53 min, raised to 80%B at 54 min and held at 80%B for the duration of the run (A = 99.9% Water/0.1% Formic Acid; B = 80% Acetonitrile/0.1% Formic Acid/19.9%Water). Fractions were collected at 1.5 min intervals using a Gilson FC403B fraction collector and then concatenated into 12 total fractions postrun. Peptides were dried by vacuum centrifugation and frozen at −20C.

LC/MS/MS Analysis

The peptide fractions were each suspended to 20 μL using 2% acetonitrile/0.1% trifluoroacetic acid. An injection of 10 μL was automatically made using a Thermo EASYnLC 1200 onto a Thermo Acclaim PepMap RSLC 0.1 mm × 20 mm C18 trapping column and washed for ∼5 min with buffer A (Buffer A = 99.9% Water/0.1% Formic Acid). Bound peptides were then eluted onto a Thermo Acclaim PepMap RSLC 0.075 mm × 500 mm C18 resolving column over 95 min with a gradient of 5%B to 8%B in 5 min, 8%B to 35%B in 84 min, ramping to 90%B at 85 min and held at 90%B for the duration of the run at a constant flow rate of 300 nL/min (Buffer B = 80% Acetonitrile/0.1% Formic Acid/19.9% water). Column temperature was maintained constant temperature at 50 °C using and integrated column oven (PRSO-V2, Sonation GmbH, Biberach, Germany). Eluted peptides were sprayed into a ThermoScientific Q-Exactive HF-X mass spectrometer using a FlexSpray spray ion source. Survey scans were taken in the Orbi trap (120,000 resolution, determined at m/z 200) and the top 15 ions in each survey scan are then subjected to automatic higher energy collision induced dissociation (HCD) with fragment spectra acquired at 45,000 resolutions. The resulting MS/MS spectra are converted to peak lists using MaxQuant1, v2.6.7 (and searched against a protein database containing all protein sequences for C. sporogenes available from Uniprot (www.uniprot.org, downloaded on 20240715; 20,323 entries) appended with common laboratory contaminants using the Andromeda2 search algorithm. MaxQuant parameters for all databases were as follows: - allow up to 2 missed tryptic sitesFixed modification of Carbamidomethyl Cysteine, TMTpro to Lysine and peptide N-terminusvariable modification of Oxidation of Methionine, Acetylation of Protein N-terminus, - peptide tolerance of ± 4.5 ppmMS/MS tolerance of ± 20 ppmFDR calculated using randomized database search.

MaxQuant protein group txt file was opened in excel to calculate the fractional intensity values. The fractional intensities were calculated by dividing the protein intensity value for each sample by the total protein intensity across all samples, and this was done for each individual protein. Within Perseus, data was filtered to remove proteins only identified by site, reverse proteins, potential contaminants, and non-numerical values. Numerical abundances were transformed by log2 followed by a 1-Smmple t test FDR 0.05 across the 3 biological replicates. Perseus generated PCA analysis, scatter plots, and heat maps. Proteomic data was uploaded to MassIVE MSV000098474.

Bioinformatic Analysis

Venn diagrams were created using data exported from Perseus into excel and the Venn tool (https://bioinformatics.psb.ugent.be/webtools/Venn/) to generate the numbers for each category. Search tool for the retrieval of interacting genes/proteins (STRING) was used to visualize protein functions associated with differential protein expression and gene ontology. STRING background was the protein database containing all protein sequences for C. sporogenes available from Uniprot (www.uniprot.org; 20,323 entries). This was the same database used for proteomics.

Glossary

Abbreviations

ATP

Adenosine Triphosphate

ABC

ATP-binding cassette

REE

rare earth elements

Eu

Eu

Sm

Sm

Fe

Fe

Zn

Zn

TMT

Tandem Mass Tag

PCA

Principal Component Analysis

DEPs

differentially expressed proteins

GO

gene ontology

ETF

electron transfer flavoprotein

NADH+/, NADPH

Nicotinamide adenine dinucleotide

LC-MS/MS

Liquid Chromatography with tandem Mass Spectrometry

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.5c09729.

• Supplementary Figures S1–S16. Phylogenic tree of Clostridium species; STRING and GO for DEPs induced by Eu, Sm, Fe, Ca, or Zn (PDF)

• Supplementary Table S1. List of proteins quantified in this study (PDF)

• Supplementary Table S2. Protein groups of TMT labeled proteins at 500 μM (PDF)

A.R., J.S., Y.F., D.R., and C.G. conceptualized and designed the project. A. R., J.S., G.M., and M.W. conducted laboratory experiments. A.R., J. S., G.M., M.W. and P.W. contributed data acquisition and analysis and figures and tables. J.S., Y.F, A.R, and D.R prepared the manuscript. All authors read, provided comments and approved the final version of the manuscript.

This work has been supported by Laboratory Directed Research and Development at Idaho National Laboratory through DOE Idaho Operations Office Contract DE-AC07–05ID14517.

The authors declare no competing financial interest.