Ln-M4B3T-DOTA – a new sterically overcrowded LCT for protein pseudocontact shift NMR spectroscopy

Abstract

Paramagnetic lanthanoid chelating tags (LCTs) enable the measurement of long-range structural restraints in biomolecular NMR spectroscopy via pseudocontact shifts and residual dipolar couplings. Herein, we report a new LCT based on the M4-cyclen scaffold bearing leucic acid side arms that yields very large anisotropic susceptibility parameters.

An M4-cyclen based lanthanoid tag with leucic acid side arms combines simplified synthesis with large magnetic susceptibility anisotropy parameters for PCS and RDC NMR applications.

The development of novel lanthanoid chelating tags (LCTs) is of significant interest in the field of paramagnetic NMR spectroscopy of proteins.^1–6 The introduction of paramagnetic centres into biomolecules provides access to long-range structural information that is difficult to obtain by conventional NMR experiments alone.^7–10 These tags allow the observation of various paramagnetic effects, thereby enabling the study of the structure, dynamics, and interaction of proteins in solution. Among these, pseudocontact shifts (PCSs) and residual dipolar couplings (RDCs) have proven especially valuable, as they provide complementary structural restraints over distances that extend well beyond those accessible by nuclear Overhauser effects.^11,12

PCSs arise from the anisotropic magnetic susceptibility of the paramagnetic metal center and provide information on both the distance and the spatial orientation of the observed nuclei relative to the metal ion. The magnitude of the observed PCS is directly dependent to the magnetic susceptibility anisotropy (Δχ) of the properties of the conjugated lanthanoid complex.^13,14 Apart from the complexed metal, the Δχ parameters can be influenced by the ligand scaffold.^15–18 Over the past decades a variety of linear and macrocyclic LCTs have been developed. These include acyclic polyaminocarboxylate-based ligands such as EDTA-^19–21 and DTPA-derived tags,^22,23 as well as macrocyclic systems based on DOTA,^24–26 DO3A, PCTA, and related cyclen derivatives. In addition, alternative concepts such as caged lanthanoid NMR probes (CLaNPs),^27–29 lanthanoid-binding peptides (LBTs),³⁰ and genetically encoded lanthanoid-binding proteins^31,32 have further expanded the available toolbox for paramagnetic NMR spectroscopy. Among these, 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid (DOTA) derivatives bearing stereospecifically introduced methyl substituents have been proven to be highly suitable scaffolds. These systems exhibit high binding affinities for lanthanoid cations and form only a single stereoisomer in solution.^33,34 Furthermore, the rigid macrocyclic framework minimizes conformational averaging, while a highly asymmetric ligand field maximizes the magnitude of the observed PCSs. The methyl substituents were introduced in such a way that intramolecular steric repulsion is minimized, and that lactic acid side arms were conformationally locked in one position.^34,35 In combination with a pyridine thiazole linker, efficient and selective conjugation to cysteine residues in proteins can be achieved.³⁶

Joss et al. further demonstrated that replacing the methyl substituents in the cyclen framework with bulkier isopropyl groups significantly enhances the tensor properties possibly by reducing translational and rotational motion of the tag.^37,38 However, one drawback of this system is the comparatively laborious and time-consuming synthetic work. We therefore set out to develop an LCT based on the established M4-cyclen scaffold³⁹ in which the lactic acid side arms were replaced by (S)-2-hydroxyisocaproic acid substituents (cf.Scheme 1).

Scheme 1. Overview of the different structures, from left to right: Ln-M7T-DOTA, Ln-P4T-DOTA and Ln-M4B3T-DOTA.

To evaluate the resulting paramagnetic properties, the Lu-, Dy- and Tm complexes of M4B3T-DOTA ((2R,2′R,2″R)-2,2′,2″-((2S,5S,8S,11S)-2,5,8,11-tetramethyl-10-((2-(methylthio)thiazolo[5,4-b]pyridin-5-yl)methyl)-1,4,7,10-tetraazacyclododecane-1,4,7-triyl)tris(4-methylpentanoic acid)) were synthesized and conjugated to uniformly ¹⁵N labelled ubiquitin as well as selectively ¹⁵N-leucine labelled human carbonic anhydrase II S166C. Comparison with the previously reported P4T-DOTA³⁷ system revealed enhanced susceptibility tensor parameters for the Dy³⁺ complex of the new tag, whereas the Tm³⁺ complex exhibits somewhat smaller tensor magnitudes. These results highlight the influence of the ligand environment on the paramagnetic properties of the LCTs.^17,40

The synthesis and characterization of the Ln-M4B3T-DOTA complexes is described in detail in the SI. Starting from l-leucic acid, the ligand was prepared, complexed with the corresponding lanthanoid ions, and subsequently activated for site specific conjugation to cysteine residues. The resulting Lu, Dy, and Tm complexes are conjugated to uniformly ¹⁵N labelled ubiquitin S57C (Fig. 1 and 2) and selectively ¹⁵N leucine labelled hCAII S166C (Fig. 4 and 5). Tagging reactions were performed as previously described by Joss et al.³⁷ The corresponding tensor parameters were calculated using the paramagpy software package⁴¹ and are summarized in Table 1. The distance between the metal center and the C_β of the cysteine was determined to be 6.7 Å, which is consistent with previously reported values.^36,37 The weak alignment introduced by the LCT enables the measurement of RDCs. The corresponding RDC tensor parameters were determined using the FANTEN software package.⁴²

Fig. 1. Overlay of ¹H-¹⁵N HSQC spectra of Dy- (blue), Tm- (red), and Lu-M4B3T-DOTA-Ubi^S57C.

Fig. 2. Tensor generated by the Dy- (left) and the Tm- (right) complex and their relative orientation to Ubi^S57C (PCS isosurfaces set at 1.5 ppm (outer layer) and 4.0 ppm (inner layer)).

Fig. 4. Overlay of ¹H-¹⁵N HSQC spectra of Dy- (blue), Tm- (red), and Lu-M4B3T-DOTA sel. ¹⁵N leucine labelled hCAII^S166C.

Fig. 5. Tensor generated by the Dy- (left) and Tm- (right) complex and their relative orientation to sel. ¹⁵N leucine labelled hCAII^S166C (PCS isosurfaces set at 0.5 ppm (outer layer) and 2.0 ppm (inner layer)).

Comparison table for the properties of the induced axial and rhombic components of the paramagnetic susceptibility tensors (Δχ_ax and Δχ_rh) for PCS and RDC, metal position in PDB coordinate frame (X_metal, Y_metal, Z_metal), Euler angles (α, β, γ) and quality factor (Q) on ubiquitin S57C (pH 6.0) at 298 K.

LCT	#PCS	Ln3+	Δχax (10−32 m3)	Δχrh (10−32 m3)	X metal (Å)	Y metal (Å)	Z metal (Å)	α (°)	β (°)	γ (°)	Q (%)
M7T-DOTA36	96	Dy	41.9	26.1	15.4	−9.8	5.1	177.3	14.2	47.5	5.0
	120	Tm	35.4	11.3	15.4	−9.8	5.1	136.4	97.4	73.5	6.0
P4T-DOTA37	50	Dy	54.0	27.0	15.5	−9.1	4.5	10.4	167.5	142.7	4.5
	72	Tm	39.3	14.6	15.5	−9.1	4.5	129.8	91.9	69.4	4.6
M4B3T-DOTA	54	Dy	73.2	27.4	14.6	−7.5	9.5	154.8	25.7	70.3	4.2
	91	Tm	32.4	9.5	14.6	−7.5	9.5	136.8	107.9	49.3	4.2

	#RDC	Ln3+	Δχax (10−32 m3)	Δχrh (10−32 m3)	X metal (Å)	Y metal (Å)	Z metal (Å)	α (°)	β (°)	γ (°)	Q (%)
P4T-DOTA37	24	Dy	−45.5	−36.2	15.5	−9.0	4.5	79.2	33.4	156.9	15.1
	36	Tm	−36.3	−12.3	15.5	−9.0	4.5	40.1	90.7	168.8	23.2
M4B3T-DOTA	11	Dy	57.0	−36.3	14.6	−7.5	9.5	147.3	116.1	167.6	9.8
	20	Tm	−26.2	−8.7	14.6	−7.5	9.5	137.6	94.4	17.5	22.7

The determined tensor parameters for the Dy complex attached to Ubi^S57C exceed those reported for the P4T system, particularly with respect to the axial component. For Tm, however, smaller values for Δχ_ax and Δχ_rh were obtained. These results indicate that the effect of the M4B3T-DOTA ligand on the magnetic susceptibility tensor is strongly lanthanoid-dependent rather than a general property of the ligand. This behaviour can be rationalized in terms of the characteristic distribution of the 4f electron density in lanthanoid ions. Depending on the electronic configuration, the 4f electron cloud can adopt either an oblate or a prolate shape, which strongly influences how the ion interacts with the surrounding ligand field. To maximize the magnetic anisotropy, the ligand field should ideally complement the shape of the 4f electron distribution and minimize repulsive interactions. Dy possesses an oblate 4f electron distribution and therefore benefits from a ligand environment in which electron density is concentrated along the axial direction. In contrast, Tm features a prolate 4f electron cloud and is stabilized by ligand electron density located predominantly in the equatorial plane. In the case of the M4B3T-DOTA scaffold, the introduction of the bulkier side arms is expected to decrease the effective electronegativity in the equatorial region while enhancing it along the axial direction. This results in a coordination environment that is particularly favourable for oblate lanthanoid ions such as Dy³⁺. This interpretation is supported by the large paramagnetic shift range observed for the Dy-M4B3T-DOTA complex, where ¹H-NMR resonances span a range exceeding 2000 ppm, thus increasing the values for Dy-M7T-DOTA by more than 500 ppm. At the same time, an environment optimized for oblate lanthanoid inevitably becomes less favourable for prolate ions. Consequently, the larger tensor anisotropy of the Dy complex translates into larger PCSs, whereas the smaller tensor of the Tm gives rise to correspondingly smaller, yet well-resolved, PCSs. A comparison with the ¹H-NMR data of the Dy-P4T scaffold (Fig. 3) further supports this hypothesis. The larger shift range arises possibly from the enhanced steric crowding resulting in a further distortion of the already asymmetric coordination polyhedron. As a consequence, this more favourable ligand field, together with a further reduction in symmetry, and a decreased flexibility provides a possible reasoning for the observed increases in tensor anisotropy and ¹H-NMR shift range. Accordingly, the larger Δχ tensors for the Dy³⁺ complex translate into larger PCSs, thereby extending the range over which informative PCSs can be detected. This may be particularly advantageous for the structural investigation of larger proteins and protein assemblies, where larger PCSs allow to avoid tagging positions close to relevant interaction sites.

Fig. 3. Overlay of ¹H-NMR spectra of Dy-M7T-DOTA (magenta), Dy-P4T-DOTA (blue) and Dy-M4B3T-DOTA (black).

It is conceivable that the side arms of the Ln-P4T-DOTA complexes retain some residual mobility, which appears to be further suppressed in the leucic acid derivatives. As an additional factor contributing to the enhanced tensor properties, the highly apolar surface of the Ln-M4B3T-DOTA complex should be considered. While it promotes favourable interactions with the protein surface, thereby decreasing tag mobility, it may also increase the likelihood of nonspecific interactions with exposed hydrophobic patches. A comparison of the metal positions for Dy and Tm in M4B3T-DOTA-Ubi^S57C shows a small, but significant displacement from the metal positions in DOTA-M7T or P4T-DOTA-Ubi^S57C (5.4 and 5.1 Å rsp.), while the metal- S57C_β distances remain unchanged. This indicates a different interaction with the protein compared to the other two LCTs, which might contribute to the observed stronger PCSs. To evaluate whether the tag perturbs the protein structure, the spectrum of untagged Ubi^S57C was overlayed with that of Lu-M4B3T-DOTA-Ubi^S57C (Fig. S10). No significant differences were observed, indicating that the overall protein fold remains unaffected upon tagging. Furthermore, ¹⁵N T1 and T2 times were recorded for Lu-, Tm- and Dy-M4B3T-DOTA-Ubi^S57C (Fig. S17 and S18) which also gave no indication of major changes of the fold. PRE effects up to 20 Å away from the metal are more pronounced for Dy than for Tm as expected. Although ubiquitin represents a relatively small and well-folded model protein, larger proteins often present more extensive and diverse surface environments. Therefore, additional validation may be advisable when conjugating the tag to larger protein systems where transient hydrophobic interactions could become more pronounced. Nevertheless, no significant structural perturbations were observed for the two proteins upon conjugation with the highly apolar Ln-M4B3T-DOTA complex, supporting its suitability for biomolecular NMR applications. In addition to PCSs, the LCTs also introduce weak alignment of proteins, enabling the measurement of RDCs. These provide orientational restraints that complement PCSs and therefore represent an additional source of structural information. RDCs were measured for the ubiquitin conjugates using IPAP-HSQC experiments. The measured and predicted RDCs are in good agreement, which indicated that the lanthanoid complex is well immobilized relative to the protein frame.

To demonstrate the applicability of the tag in a more complex biomolecular system, it was further conjugated to selectively ¹⁵N-leucine labelled hCAII S166C. For the Dy³⁺ complex, the new tag introduced a larger Δχ tensor than the corresponding P4T-DOTA complex, resulting in pronounced paramagnetic effects (Table 2). In contrast, a smaller tensor magnitude was obtained for the Tm³⁺ complex, although substantial PCSs remain readily observable. Similar lanthanoid-dependent differences were already observed for the ubiquitin system and reflect the strong influence of the ligand field on the anisotropic magnetic susceptibility and spatial distribution of the 4f electron density of the lanthanoid ion.

Table 2. Comparison table for the properties of the induced axial and rhombic components of the paramagnetic susceptibility tensors (Δχ_ax and Δχ_rh) for PCS, metal position in PDB coordinate frame (X_metal, Y_metal, Z_metal), Euler angles (α, β, γ) and quality factor (Q) on hCA II S166C (pH 6.8) at 298 K.

LCT	#PCS	Ln3+	Δχax (10−32 m3)	Δχrh (10−32 m3)	X metal (Å)	Y metal (Å)	Z metal (Å)	α (°)	β (°)	γ (°)	Q (%)
P4T-DOTA37	40	Dy	−46.1	−30.5	−15.9	−3.9	−10.8	24.4	43.7	38.9	11.4
	46	Tm	44.3	4.8	−15.9	−3.9	−10.8	174.7	157.0	29.0	5.8
M4B3T-DOTA	28	Dy	56.2	29.5	−16.1	−2.7	10.8	155.5	69.9	143.0	9.8
	34	Tm	29.9	3.3	−16.1	−2.7	10.8	2.3	23.4	177.3	3.6

In conclusion, we synthesized a new, strongly paramagnetic LCT that exhibits an exceptionally large paramagnetic ¹H-NMR shift range for the Dy complex. The new tag was successfully conjugated to two different protein constructs, ubiquitin S57C, and selectively ¹⁵N-leucine labelled hCAII S166C, demonstrating its applicability to both small model proteins and larger biomolecular constructs. For oblate lanthanoid ions such as Dy³⁺, a pronounced enhancement of the susceptibility tensor properties was observed. Owing to these favourable tensor properties, the complex represents a promising alternative to established systems such as Ln-M7T-DOTA, as it enables the observation of PCSs at even larger distances from the metal center with comparable synthetic work.

Author contributions

AM: synthesis, NMR experiments, interpretation of data and co-writing of the manuscript. DH: interpretation of data and co-writing of the manuscript.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting this article have been included as part of the supplementary information (SI). Supplementary information is available. See DOI: https://doi.org/10.1039/d6cc03248f.

Biological structures were generated using the open-source software PyMol (https://www.pymol.org/).