Lanthanide Binding Tags (LBTs)

• LBTs are short lanthanide-chelating peptides that form defined coordination pockets for trivalent lanthanide ions, enabling luminescence, paramagnetism, and quantum spin manipulation.
• Their engineered sequence architecture, including variants like LBTC and daLBT, allows tunable metal-binding properties that facilitate applications in quantum information, molecular spintronics, and rare-earth capture.
• LBT applications span structural studies using EPR and NV-diamond biosensing to the development of self-assembled monolayers and interfacial films, though challenges in coherence and ligand specificity remain.

Searching arXiv for the cited LBT papers and closely related records to ground the article. arxiv_search.query({"search_query":"id:(Rosaleny et al., 2017) OR id:(Gao et al., 20 Jul 2025) OR id:(Crane et al., 2024)", "max_results": 10, "sort_by": "submittedDate", "sort_order": "descending"}) Lanthanide Binding Tags (LBTs) are short lanthanide-chelating peptides, typically about 15–17 amino acids long and also described as short peptides around 20 amino acids, that form locally folded coordination pockets for trivalent lanthanide ions. Originally developed in biotechnology for protein structure and dynamics studies, they have been used as luminescent or paramagnetic tags inserted at controlled positions in proteins; subsequent work treats them as tunable metallopeptide coordination modules for quantum information, molecular spintronics, interfacial rare-earth capture, and NV-diamond magnetic biosensing (Rosaleny et al., 2017, Crane et al., 2024, Gao et al., 20 Jul 2025).

1. Origin, biochemical role, and conceptual scope

LBTs were originally developed in biotechnology from calcium-binding motifs of proteins, with classic optimization by Imperiali and co-workers using peptide-library screening to improve lanthanide selectivity and luminescence (Rosaleny et al., 2017). In that setting, they functioned as luminescent or paramagnetic tags for structural studies, including controlled insertion into proteins and DEER-based distance measurements. A separate line of work describes them as a prominent group of fluorescent probes extensively utilized in biological detection, while also emphasizing that their small size allows integration into proteins without substantially perturbing protein structure or function (Gao et al., 20 Jul 2025).

This biochemical origin remains central to later materials-oriented uses. In the quantum-information and molecular-spintronics context, the peptide is not the qubit basis; rather, it provides a programmable coordination environment for a lanthanide-centered spin degree of freedom (Rosaleny et al., 2017). In rare-earth separation at fluid interfaces, the same motif is repurposed as the recognition element in peptide surfactants designed to capture rare-earth ions at the air–water interface (Crane et al., 2024). In quantum biosensing, Gd-loaded LBTs are used as magnetic labels whose fluctuating magnetic fields shorten the spin-lattice relaxation time of near-surface NV centers in diamond (Gao et al., 20 Jul 2025).

A common simplification is that LBTs are merely fluorescent labels. The cited literature contradicts that view in three distinct ways: LBTs can host coherent lanthanide spin manipulation, they can act as interfacial rare-earth binders whose selectivity depends on detailed coordination-sphere design, and they can serve as genetically encodable paramagnetic reporters in magnetic biosensing (Rosaleny et al., 2017, Crane et al., 2024, Gao et al., 20 Jul 2025).

2. Sequence architecture and coordination chemistry

In the quantum-spin study, the “most standard” LBT is identified as YIDTNNDGWYEGDELLA, with a shortened 15-residue folding core YIDTNNDGWYEGDEL argued to behave as an autonomous folding domain (Rosaleny et al., 2017). A cysteine-terminated surface-binding variant, YIDTNNDGWYEGDELC, is termed LBTC, and a fused asymmetric double tag, YIDTDNDGWYEGDELYIDTNNDGWYEGDELLA, is termed daLBT. In the interfacial-separation work, related variants are formulated in terms of net charge and ligand placement, notably \mathrm{LBT}^{5-}, \mathrm{LBT}^{3-}, and \mathrm{RR\!-\!LBT}^{3-} (Crane et al., 2024).

The structural basis of lanthanide binding is an oxygen-rich coordination environment. The donor set is described as arising from carboxylate oxygens from Asp/Glu side chains, carboxamide oxygens from Asn side chains, and one peptide amide oxygen from the backbone (Rosaleny et al., 2017). The coordination sphere is stated to contain 8 oxygen atoms and to be substantially conserved across related crystal structures when the amino-acid sequence is conserved. In the interfacial study, the binding loop is described as forming a 3-dimensional coordination sphere of several $6$–$8$ ligands, primarily carboxylates and carbonyls, around the rare-earth ion (Crane et al., 2024). The geometry is characterized as oxygen-dominated, irregular, and asymmetric rather than assigned to a strict idealized polyhedron (Rosaleny et al., 2017).

The ability to create distinguishable metal sites is a major design principle. In daLBT, one tag differs from the other by an Asn $\rightarrow$ Asp substitution, replacing a neutral carboxamide donor with an anionic carboxylate donor and thereby creating two inequivalent coordination environments within one peptide scaffold (Rosaleny et al., 2017). In the interfacial-selectivity context, the decisive issue is not only overall charge but whether anionic ligands remain outside the primary coordination sphere. \mathrm{LBT}^{3-} was engineered from \mathrm{LBT}^{5-} by C-terminal amidation and D11N mutation so that it retains identical coordinating ligands inside the loop but contains no anionic ligands outside its coordination sphere; \mathrm{RR\!-\!LBT}^{3-} instead compensates charge by adding two Arg residues while retaining D11 outside the coordination sphere (Crane et al., 2024).

These sequence-level distinctions establish a recurring theme across the literature: LBT function depends simultaneously on canonical inward-facing coordination chemistry and on the exterior chemistry left accessible after complexation. This suggests that sequence design for LBTs is not reducible to affinity alone.

3. Qubits, coherent control, and modular multiqubit construction

The quantum-computing study explicitly treats LBT-bound lanthanides as molecular spin qubits, with the lanthanide-centered spin degree of freedom split by spin–orbit coupling and crystal field and manipulated by pulsed EPR (Rosaleny et al., 2017). The peptide scaffold defines the crystal field, magnetic anisotropy, coupling to vibrations, and local spin bath. The lanthanides tested were $\mathrm{Nd, Gd, Tb, Dy, Ho, Er, Tm, Yb}$, and the rationale for choosing lanthanides included prior promise in quantum technologies, the existence of atomic clock transitions, the fact that single ions can host multiple qubits, and prior spintronic readout of nuclear spin qubits at the single-molecule level.

Under the reported conditions, preliminary CW EPR detected signals only for NdLBT and GdLBT, likely in part because concentrations were in the range $50$–$100~\mu\mathrm{mol}$ (Rosaleny et al., 2017). For NdLBT, X-band pulsed EPR at $4~\mathrm{K}$ in frozen solution showed continuous-wave EPR, echo-detected EPR, Rabi oscillations, and determinations of $T_1$ and $T_2$. Reported performance values were $T_1 \simeq 3~\mu\mathrm{s}$ and $8$0 for NdLBT, and $8$1 and $8$2 for GdLBT. Approximately 20 coherent Rabi oscillations were observed for both NdLBT and GdLBT under Hartmann–Hahn conditions, interpreted as roughly 20 coherent quantum operations before coherence is lost. The authors highlight NdLBT as, to the best of their knowledge, the first reported Nd-based molecular spin qubit (Rosaleny et al., 2017).

Modularity is extended to multiqubit architectures by concatenating locally folding tags. Structural alignments in different protein contexts were used to argue that the 15-aa core folds similarly as an isolated tag, fused to ubiquitin, and inserted into interleukin-$8$3, supporting the concept of autonomous folding (Rosaleny et al., 2017). This underpins the design of daLBT as a molecularly encoded asymmetric two-qubit scaffold and of an asymmetric chain of 9 spin qubits with a spin-spin separation of about $8$4 and an arbitrarily chosen sequence of coordination environments, represented as 010011101, where 0 and 1 denote the carboxylate and carboxamide variants.

Scaling beyond short synthetic peptides was approached through recombinant protein expression in E. coli using a pGEX-2T plasmid and a GST tag (Rosaleny et al., 2017). For daLBT, the recombinant product was reported as consistent with $8$5, interpreted as $8$6 plus $8$7. In the 9-tag design, unique linkers based on permutations of GASAG were introduced to permit flexible local arrangement, avoid interfering with independent tag folding, and enable specific PCR amplification of constructs containing between 1 and 9 tags. The paper does not report quantified inter-qubit couplings for daLBT or the 9-tag chain, so the architecture is motivated primarily by site differentiation and scalable assembly rather than demonstrated gate coupling strength (Rosaleny et al., 2017).

4. Surface-grafted monolayers and molecular spintronics

A further extension of LBT chemistry is the preparation of paramagnetic, chiral, self-assembled monolayers on Au(111) using the cysteine-bearing variant LBTC, YIDTNNDGWYEGDELC, in which the cysteine thiol provides Au binding (Rosaleny et al., 2017). In this construct, the peptide is chiral by composition, paramagnetic once loaded with a lanthanide, and surface-addressable through Au–S anchoring. The monolayer was prepared by immersing cleaned Au substrates overnight in buffered aqueous solution containing HEPES $8$8, pH $8$9, NaCl $\rightarrow$0, TbCl$\rightarrow$1 $\rightarrow$2, TCEP $\rightarrow$3, and LBTC $\rightarrow$4, followed by thorough rinsing to remove physisorbed material.

The characterization suite combined AFM, MALDI-TOF, XPS, Quartz Crystal Microbalance, DFT, and NEGF-DFT (Rosaleny et al., 2017). AFM indicated that roughness remained nearly unchanged, consistent with homogeneous coverage without large aggregates. MALDI-TOF showed a strong fragment peak at $\rightarrow$5, assigned to $\rightarrow$6, and a smaller peak at $\rightarrow$7, assigned to intact LBTC + Na$\rightarrow$8, indicating peptide integrity on the surface. XPS detected Tb $\rightarrow$9 peaks at $\mathrm{Nd, Gd, Tb, Dy, Ho, Er, Tm, Yb}$0 and $\mathrm{Nd, Gd, Tb, Dy, Ho, Er, Tm, Yb}$1, directly confirming terbium on the surface. Quartz Crystal Microbalance gave a frequency shift $\mathrm{Nd, Gd, Tb, Dy, Ho, Er, Tm, Yb}$2, corresponding to $\mathrm{Nd, Gd, Tb, Dy, Ho, Er, Tm, Yb}$3–$\mathrm{Nd, Gd, Tb, Dy, Ho, Er, Tm, Yb}$4 coverage depending on the full-coverage model.

The transport calculations identified the nearest transmission peaks at

$\mathrm{Nd, Gd, Tb, Dy, Ho, Er, Tm, Yb}$5

so conduction was inferred to require either a moderate gate voltage or strong bias (Rosaleny et al., 2017). Peak broadening was described as moderate, indicating low hybridization between biomolecule and metal electrode, and the local DOS suggested that the conduction channels involve the amide-bond backbone $\mathrm{Nd, Gd, Tb, Dy, Ho, Er, Tm, Yb}$6 system. This geometry is notable because one amide oxygen from the backbone directly coordinates the lanthanide, placing a likely transport pathway close to the magnetic center.

The spintronics discussion considered two transport regimes. In an extended SOMO regime, an extended conducting orbital bonded to the lanthanide via amide coordination was associated with exchange coupling in the range $\mathrm{Nd, Gd, Tb, Dy, Ho, Er, Tm, Yb}$7–$\mathrm{Nd, Gd, Tb, Dy, Ho, Er, Tm, Yb}$8 (Rosaleny et al., 2017). In a discrete tunneling-step regime, transport was described as proceeding between backbone amide $\mathrm{Nd, Gd, Tb, Dy, Ho, Er, Tm, Yb}$9-bonds, effectively creating a series of local spin polarizer/analyzer segments. The estimated local magnetic field from the lanthanide on the directly coordinated NCO amide group was of order $50$0 under assumed full lanthanide spin polarization, while the rest of the chain experienced fields one to two orders of magnitude smaller. Within that framework, LBT/LBTC monolayers were proposed as candidates for CISS and possibly magnetoresistance.

5. Interfacial rare-earth capture and rheological selectivity

At the air–water interface, LBT-derived peptide surfactants were studied as capture agents for rare-earth separations by foam fractionation (Crane et al., 2024). The mechanistic question was whether peptides designed to form $50$1 complexes with a single lanthanide remain selective and structurally intact at an interface, or instead recruit additional cations through exposed charged groups. Tb$50$2 was used as the model rare-earth ion because the chosen LBTs are known to bind terbium strongly, with prior dissociation constants in the nanomolar range; for \mathrm{LBT}^{5-}:\mathrm{Tb}^{3+}, $50$3 was cited.

The parent \mathrm{LBT}^{5-} carries net charge $50$4 and is known to bind Tb$50$5, but prior XFNTR measurements showed that at greater than equimolar terbium the adsorbed interfacial ratio is

$50$6

rather than $50$7 (Crane et al., 2024). The interpretation advanced in the paper is that the primary binding loop remains intact while excess Tb$50$8 associates with anionic ligands outside the coordination sphere, specifically D11 and the C-terminal carboxylate, thereby linking neighboring peptide–Tb complexes. By contrast, \mathrm{LBT}^{3-} was engineered by C-terminal amidation and D11N mutation to remove anionic functionality outside the coordination sphere while preserving the coordinating ligands inside the loop; prior interfacial X-ray work showed a Tb:peptide ratio of exactly $50$9. The control \mathrm{RR\!-\!LBT}^{3-} also has net charge $100~\mu\mathrm{mol}$0 but retains D11 outside the coordination sphere, because neutrality is achieved instead by appending two Arg residues at the N-terminus.

Interfacial rheology provided the central readout. For purely viscous interfaces, the Boussinesq number was defined as

$100~\mu\mathrm{mol}$1

and particle tracking used the mean squared displacement

$100~\mu\mathrm{mol}$2

with $100~\mu\mathrm{mol}$3 in the fit $100~\mu\mathrm{mol}$4 corresponding to diffusive, purely viscous motion (Crane et al., 2024). \mathrm{LBT}^{5-} without Tb$100~\mu\mathrm{mol}$5 produced a highly viscous layer with $100~\mu\mathrm{mol}$6; at equimolar Tb$100~\mu\mathrm{mol}$7, $100~\mu\mathrm{mol}$8; and above equimolar Tb$100~\mu\mathrm{mol}$9 the layer underwent a viscous-to-elastic transition. At $4~\mathrm{K}$0 Tb$4~\mathrm{K}$1 with $4~\mathrm{K}$2 peptide, the reported values were $4~\mathrm{K}$3 and $4~\mathrm{K}$4, with $4~\mathrm{K}$5, indicating a thin solid-like film. \mathrm{LBT}^{3-}, by contrast, remained purely viscous over Tb$4~\mathrm{K}$6:peptide ratios from 0 to 2, and \mathrm{RR\!-\!LBT}^{3-} again formed elastic interfacial films in excess Tb$4~\mathrm{K}$7.

Simulation reinforced this interpretation. All-atom interfacial MD indicated that \mathrm{LBT}^{5-} developed behavior consistent with network formation and peptide caging at supra-stoichiometric Tb$4~\mathrm{K}$8, while \mathrm{LBT}^{3-} remained diffusively mobile (Crane et al., 2024). Well-tempered metadynamics mapped free-energy landscapes using Tb–oxygen coordination number and loop C$4~\mathrm{K}$9-RMSD as collective variables, and the resulting low-free-energy conformations differed by $T_1$0. For \mathrm{LBT}^{5-}, more probable states corresponded to Tb$T_1$1 coordinated by 10, 8, and 6 oxygen atoms from the peptide; for \mathrm{LBT}^{3-}, the corresponding values were 9, 7, and 6. Crucially, the authors state that all carboxylate ligands within the binding loop remain associated with Tb$T_1$2 in both peptides, so elasticity in \mathrm{LBT}^{5-} cannot be attributed to loop collapse and instead arises from non-binding-loop ligands (Crane et al., 2024).

The key mechanistic conclusion is therefore not simply that neutrality is desirable. \mathrm{RR\!-\!LBT}^{3-} shows that overall charge neutrality alone does not suppress non-specific rare-earth recruitment if an anionic ligand remains sterically accessible outside the coordination sphere (Crane et al., 2024). In the authors’ formulation, selective interfacial binders should satisfy two criteria: net neutrality of the complex and no anionic ligands outside the binding loop.

6. Gd-loaded LBTs as magnetic labels in NV-diamond biosensing

A distinct recent development is the use of Gd-loaded LBTs as magnetic rather than fluorescent labels, detected via the relaxometry of near-surface NV centers in diamond (Gao et al., 20 Jul 2025). The sensing mechanism is based on the shortening of NV spin-lattice relaxation time $T_1$3 by local magnetic noise from unpaired electrons in Gd$T_1$4. The NV ground-state zero-field splitting is given as

$T_1$5

and the normalized relaxometry signal is

$T_1$6

where $T_1$7 and $T_1$8 are fluorescence signals from complementary pulse sequences (Gao et al., 20 Jul 2025).

In this study the LBT was used in an engineered fusion format rather than only as a free peptide. An LBT–RBD nanobody fusion protein, synthesized by BioBasic, was cloned into pET-28a(+) with N-terminal 6xHis and MBP tags and expressed in E. coli BL21 (DE3) Rosetta (Gao et al., 20 Jul 2025). Purification employed a 5 mL HisTrap column on ÄKTA FPLC followed by HiLoad 16/600 Superdex 200 pg gel filtration. Gd loading was performed by mixing purified LBT fusion protein with $T_1$9 GdCl$T_2$0, vortexing for 1 min, centrifuging to remove unbound Gd$T_2$1, and washing the pellet three times with PBS; the LBT protein precipitated after binding Gd ions.

Direct detection was performed by exposing amine-functionalized diamond to LBT solutions of different concentration. Before LBT treatment, $T_2$2; after the $T_2$3 reference condition, fitted $T_2$4; and at $T_2$5 LBT, fitted $T_2$6 (Gao et al., 20 Jul 2025). The reported detection limit for direct LBT sensing was $T_2$7, with data averaged from five randomly selected locations on the diamond surface. Surface blocking was essential: without BSA, treatment corresponding to $T_2$8 gave $T_2$9, whereas with $T_1 \simeq 3~\mu\mathrm{s}$0 BSA blocking the value was $T_1 \simeq 3~\mu\mathrm{s}$1, close to the no-LBT baseline.

The target-recognition demonstration used a sandwich architecture on the diamond surface: primary anti-RBD antibodies, BSA blocking, SARS-CoV-2 spike protein receptor-binding domain (RBD), and then anti-RBD nanobody–LBT fusion (Gao et al., 20 Jul 2025). In this geometry, the assay signal was a binding-dependent decrease in NV $T_1 \simeq 3~\mu\mathrm{s}$2. The reported calibration was linear, with slope

$T_1 \simeq 3~\mu\mathrm{s}$3

The paper contains an internal inconsistency regarding the detection limit: the abstract and conclusion state a detection threshold of approximately $T_1 \simeq 3~\mu\mathrm{s}$4, whereas the results-section estimate based on a stated $T_1 \simeq 3~\mu\mathrm{s}$5 error of $T_1 \simeq 3~\mu\mathrm{s}$6 and the reported slope implies approximately $T_1 \simeq 3~\mu\mathrm{s}$7 (Gao et al., 20 Jul 2025). A careful summary is therefore that the demonstrated practical threshold is about $T_1 \simeq 3~\mu\mathrm{s}$8–$T_1 \simeq 3~\mu\mathrm{s}$9, with $8$00 following directly from the paper’s own error-and-slope calculation.

This work is methodologically important for LBT science because it repositions the tag as a quantum-detectable paramagnetic module. The main limitation identified by the authors is distance-dependent sensitivity loss in the antibody–RBD–LBT sandwich geometry, since increased distance between Gd and the NV layer weakens the magnetic coupling (Gao et al., 20 Jul 2025).

7. Advantages, limitations, and research directions

Across these applications, several advantages recur. LBTs are genetically encodable, short, modular, and sufficiently robust in local folding to permit predictable replication or modification (Rosaleny et al., 2017). They are amenable to sequence optimization, biochemical functionalization such as cysteine addition for Au binding, organization in space by molecular-biology methods, and recombinant production of long constructs impractical by standard SPPS (Rosaleny et al., 2017). In biosensing, their small size and compatibility with fusion to recognition proteins provide a route to magnetic labels that avoid background fluorescence noise (Gao et al., 20 Jul 2025). In interfacial separations, canonical loop binding can remain intact even under anisotropic air–water-interface conditions, provided exterior ligand placement is properly controlled (Crane et al., 2024).

The limitations are equally clear. In quantum-coherent operation, coherence remains modest: $8$01 for NdLBT and $8$02 for GdLBT, with about 20 Rabi cycles reported (Rosaleny et al., 2017). The peptide environment is vibrationally rich, and a dense nuclear spin bath can shorten $8$03. Exact crystallization of the relevant peptides proved difficult, and controlled useful couplings in larger multiqubit constructs were not yet demonstrated (Rosaleny et al., 2017). In surface transport, resonances lie away from the Fermi level, so gating or substantial bias is required (Rosaleny et al., 2017). At interfaces, net neutrality is necessary but not sufficient if anionic ligands remain exposed outside the coordination sphere (Crane et al., 2024). In NV biosensing, nonspecific adsorption must be suppressed by blocking, and the target-assay sensitivity remains below RT-PCR; the reported RBD detection limit also contains the noted $8$04 versus $8$05 inconsistency (Gao et al., 20 Jul 2025).

Two misconceptions are directly resolved by the cited literature. First, LBTs are not restricted to fluorescence readout; they can support coherent spin control, interfacial rare-earth capture, and magnetic quantum sensing (Rosaleny et al., 2017, Crane et al., 2024, Gao et al., 20 Jul 2025). Second, selectivity is not determined solely by strong primary lanthanide affinity or by overall charge neutrality; the decisive variable can be whether any anionic ligand remains sterically accessible outside the intended coordination sphere (Crane et al., 2024).

The proposed future directions are correspondingly diverse. For quantum applications, the authors of the spin-qubit study explicitly propose combinatorial screening of peptide libraries to identify sequences in which low-energy vibrations minimally perturb the coordination environment, combined with theory of spin-coupled vibrations and effective electrostatic crystal-field calculations to guide sequence optimization (Rosaleny et al., 2017). Other implied directions include longer multiqubit peptides, protein-based architectures, CISS-oriented monolayers, and coupling to single-molecule transistors. For biosensing, shorter recognition architectures and nanodiamonds were proposed to reduce NV–Gd distance and improve sensitivity (Gao et al., 20 Jul 2025). Taken together, these results suggest that LBTs occupy a distinctive position at the interface of coordination chemistry, peptide engineering, quantum spin physics, and device-oriented biomolecular materials science.