Lasso Peptide Syanodin I: Loop-Pulling Two-Step Thermal Unthreading Mechanism

Abstract

Lasso peptides are a class of ribosomally synthesized and post-translationally modified peptides (RiPPs) where the C-terminal tail is threaded and sterically trapped within an N-terminal macrolactam ring. In the present work, the thermal unthreading process of the lasso peptide syanodin I was characterized using liquid chromatography, trapped ion mobility spectrometry, electron-capture dissociation, and tandem mass spectrometry (LC-TIMS-q-ECD-ToF MS/MS) and steered molecular dynamics (SMD). The analytical workflow allows for the separation (LC and TIMS) and identification (ECD MS/MS) of syanodin I kinetic intermediates as a function of the solution temperature. The syanodin WT with the plug residue Gln13 can be identified from the branched-cyclic topology based on unique retention time (RT), IMS bands, and hydrogen migration events near the molecular knot (c_i•/z_jʹ fragments). After heat treatment for 1 h at 95 °C, the unthreaded syanodin I (branched-cyclic) and an intermediate lasso structure, with Leu15 as plug residue, are identified based on the RT, IMS bands and signature ECD MS/MS fragments. Mutagenesis experiments, substituting to bulkier ([L15W] and [L15W/A16W]) and smaller ([L15A]) residues, confirmed the intermediate plug residue at Leu15. Changes in the initial plug residue Gln13 to a smaller residue ([Q13A]) resulted in a lasso structure with Leu15 as the plug residue. SMD simulations supported a loop-pulling mechanism, but also the possibility of a tail-pulling mechanism, in good agreement with the experimental observations. This is the first report of a two-step, lasso thermal unfolding mechanism driven by the two steric constraints based on loop-pulling of the lasso tail.

Graphical Abstract

INTRODUCTION

Knots in peptides and proteins are one of the strategies nature employs to significantly improve ligand binding and structural stability.¹ Lasso peptides are ribosomally synthesized and post-translationally modified peptides (RiPPs) that share a knot-like structure, in which the C-terminal tail is threaded and sterically trapped within an N-terminal macrolactam ring.^2–4 The macrolactam ring is generated through a condensation reaction between the N-terminal α-amino group with the carboxylic acid side chain of an Asp or Glu residue at position 7–9 of the core peptide.^4–7 The knot-like structure is primarily maintained by steric interactions from bulky residues in the tail, called plug residues, and sometimes assisted by disulfide bonds (Figure S1).^8–11 Lasso peptides are divided into four classes depending on the presence of two (class I), one (class III and IV), or the absence (class II) of disulfide linkages (Figure S1).^{11, 12} The knot topology confers lasso peptides with an outstanding range of biological activities, including antimicrobial, antiviral, inhibitory, receptor antagonistic, and anticancer properties.^{2–4, 12–17} In addition, the rigid lasso knot-like structure often offers high stability against proteolytic degradation and prolonged incubation at elevated temperatures.^{2, 4, 18–21} However, some class II lasso peptides are heat sensitive, i.e., their C-terminal tail unthreads out of the macrolactam ring when heated yielding a branched-cyclic analog.^{18, 22, 23} The small size (~20 amino acids) combined with the simple knot-like structure of lasso peptides make them good model systems for studying the unthreading of entangled biomolecules, both with experiments and through molecular dynamics (MD) simulations. MD simulations were recently reported to distinguish lasso-fold from unthreaded structures and assess handedness for the case of the lasso peptide microcin J25; these studies demonstrated that the pre-lasso folding landscape can be facilitated by sequence deviations, chemical modifications, and reaction conditions.^24–26 The C-terminal tails of lasso peptides can unthread following (i) a tail-pulling mechanism, in which the ring (green) first slides over the upper plug residue (red), and then the rest of the loop residues (blue) are drawn through the macrolactam ring (Figure 1a) and (ii) a loop-pulling mechanism, in which the ring (green) slips over the lower plug residue (red), and then the remaining tail residues (orange) are pulled out of the macrolactam ring (Figure 1a).²² Previous studies have reported the thermal unthreading of the lasso peptide astexin-3 and benenodin-1 by a tail-pulling mechanism, using a combination of liquid chromatography – mass spectrometry (LC-MS) experiments and MD simulations.^{21, 23, 27}

Figure 1.

Schematics showing (a) the mechanisms of lasso peptide unthreading and (b) sequences of syanodin I and its branched-cyclic analog. The macrolactam rings are colored in green, the loop residues in blue, the plugs in red, and the C-terminal tail in orange.

Syanodin I is a 17 residue class II lasso peptide produced from Sphingobium yanoikuyae, for which no biological activity has been reported yet.¹⁸ Its macrolactam ring encompasses eight residues closed by an isopeptide bond between the α-amino group of Gly1 and the side chain carboxyl group of Asp8 (highlighted in green in Figure 1b). The loop region is composed of four residues located above the ring (highlighted in blue in Figure 1b) and a C-terminal tail of four residues located below the ring (highlighted in orange in Figure 1b). Tandem mass spectrometry (MS/MS) experiments using electron transfer dissociation (ETD), have permitted the assignment of the Gln13 residue (highlighted in red in Figure 1b) as being the lower plug residue maintaining the lasso topology.²⁸ Ion mobility spectrometry – mass spectrometry (IMS-MS) also proved to be an efficient alternative strategy for the discrimination between the lasso and branched-cyclic topologies of syanodin I.^29–31 The implementation of an electromagnetostatic (EMS) cell,^{32, 33} capable of performing ECD without the need for long reaction times or ultrahigh vacuum, into a commercially available quadrupole time-of-flight mass spectrometer (q-ToF MS)^34–36 has opened new avenues for the structural elucidation of biomolecules due to the fast speed of the ECD events that are thereby compatible with the LC and IMS timescale. Electron-based dissociation is currently the only MS-based technique capable of assigning the plug residues stabilizing the lasso structure.^{28, 37–39}

Syanodin I has been shown to unthread after prolonged incubation at elevated temperatures, which makes it interesting to decipher the factor governing its thermal unthreading mechanism.¹⁸ In the present work, we studied the steric effects and loop dynamics that drive the thermal unthreading mechanism of the heat sensitive syanodin I. Thermal intermediates were separated and characterized based on liquid chromatography (LC), trapped ion mobility spectrometry (TIMS), and electron capture dissociation (ECD) tandem mass spectrometry. In the following discussion, we take advantage of the capability to separate conformers based on their retention time, ion mobility, and signature ECD MS/MS patterns. Mutagenesis analysis was used to decipher the effect of steric amino acids as well as the energy barriers associated with the unthreading process. Complementary steered molecular dynamics simulations (SMD) were used to evaluate the unthreading mechanisms.

EXPERIMENTAL SECTION

Materials and Reagents.

Details on syanodin I production have been reported previously.¹⁸ Briefly, syanodin I was produced heterologously in Escherichia coli BL21 (DE3) by IPTG-induced expression of a pET41a production plasmid carrying the syanodin I biosynthetic gene cluster in M9 medium at 20 °C for 3 days. Syanodin I was extracted from the cell pellet with methanol and analyzed as crude extracts. The [Q13A], [Q13W], [L15A], [L15W], and [L15W/A16W] variants of syanodin I were obtained by site-directed mutagenesis of the precursor peptide encoding gene on this plasmid, followed by their heterologous production using the thus obtained mutated plasmids under the same expression conditions as employed for WT syanodin I. The branched-cyclic analog of WT syanodin I was obtained by solid-phase synthesis from Genepep (St Jean de Védas, France). Solutions were prepared at a final concentration of 10 μM in 10 mM ammonium acetate (NH4Ac). To investigate the thermal unthreading mechanism of syanodin I, solutions of 10 μM of syanodin I WT or variants were incubated at 25 °C or 95 °C for either 1 h or 16 h. The same procedure was applied to the branched-cyclic analog that was used as a control. Tuning Mix standard (G1969–85000), obtained from Agilent Technologies (Santa Clara, CA), was used for external ion mobility and mass calibration of the LC-TIMS-q-ToF MS instrument.

LC-TIMS-q-ECD-ToF MS Instrumentation.

Reverse-phase liquid chromatography was carried out on a Dionex UltiMate 3000 LC system equipped with a XBridge peptide BEH C18 column (4.6 mm × 250 mm × 5 μm), which was kept at 50 °C. The flow rate was kept at 0.4 mL/min, and the acquisition time was 60 min. LC separations were achieved with H₂O 0.1% formic acid (A) and ACN 0.1% formic acid (B) using the following elution gradient: i) 0–2.7 min with 10% B; ii) 2.7–8 min to 20% B; iii) 8–12 min to 28% B; iv) 12–30 min to 35% B; v) 30–43 min to 40% B; vi) 43–45 min to 95% B; vii) 45–50 min with 95% B; viii) 50–55 min to 10% B.

The TIMS and ECD capabilities were integrated into a Bruker Maxis Impact II ToF MS (Bruker Daltonics Inc., Billerica, MA) instrument equipped with an ESI source operated in positive mode (Figure S2). A capillary voltage of 4500 V with a nitrogen nebulizer, dry gas, and temperature set at 3 bars, 10.0 L/min, and 225 °C, respectively, were used as the source parameters. A detailed layout of this experimental apparatus is described elsewhere.³⁶ The fundamentals of TIMS operation and calibration have been described previously.^40–43 TIMS experiments were carried out using nitrogen (N₂) as buffer gas at ambient temperature (T) with a gas velocity defined by the funnel entrance (P₁ = 2.1 mbar) and exit (P₂ = 0.48 mbar) pressure differences. An rf of 250 Vpp at 880 kHz was applied to all TIMS electrodes. A deflector voltage of 150 V and a voltage ramp of −150 to −15 V were used for all experiments.

A custom-built 19 mm long EMS cell (e-MSion Inc., Agilent Technologies, Corvallis, OR) was attached to a custom-built collision cell and mounted between the quadrupole exit and the pulsing plates of the ToF MS instrument (Figure S2). The filament was operated at a current of 2.5 A. The collision cell was operated by using high-purity argon (oxygen free) to enhance the cooling of the ions. Additional details on the ECD operation are described elsewhere.^{36, 44} ECD MS/MS spectra were collected on quadrupole mass-selected [M+2H]²⁺ precursor ions (isolation window of 10 Da), where each of the ECD events was synchronized with the LC and TIMS time segments. The MS/MS fragment ion annotations were performed using a custom excel table with all theoretical combinations of fragments based on the sequences of syanodin I and syanodin I variants. The ECD spectra were annotated with a mass error <10 ppm.

Molecular Dynamics Simulations.

Simulation systems for the investigated WT syanodin I and variants were set up using CHARMM-GUI.^45–47 A custom force-field/topology patch was defined with CHARMM General Forcefield (CGenFF)⁴⁸ to connect the N-terminus of Gly1 to the γ-carbon (Cγ) of Asp8, for generating the macrolactam ring, entrapping the C-terminal part of the lasso peptides. The patch is included in the Supporting Information. Each system was solvated in a cubic box with 39 Å dimensions with TIP3P water model and 0.15 mM KCl.⁴⁹ Simulations were performed at 300 K and 1 atm. pressure, using the Langevin thermostat implemented in CHARMM⁴⁵ with damping coefficient of 1 ps⁻¹ to control the temperature and the Nose-Hoover Langevin piston method and piston decay time of 25 fs to maintain the pressure.⁵⁰ Covalent bonds, involving hydrogen atoms, were constrained using SHAKEH,⁵¹ and long-range electrostatic interactions were calculated using the Particle Mesh Ewald (PME) method.⁵² For non-bonded interactions, a cutoff distance of 12 Å with a switching function applied at 10 Å was used.

Molecular Dynamics (MD) simulations were performed using GPU-accelerated NAMD 3.0⁵³ with the CHARMM36m force field.⁵⁴ The simulations were initiated with a 10,000-steps of minimization and 1 ns of equilibration. After a 10-ns conventional MD simulation (cMD), ensuring the system stability, a steered molecular dynamics (SMD) simulation was performed to pull the lasso tail out of the ring. SMD was chosen because simulations at elevated temperatures (up to 1,000 K) were unable to unthread the peptide variants for the simulated timescales of up to 500 ns. For SMD, the Cα atoms of the lasso ring residues were restrained and the tail was pulled perpendicular to the ring’s plane for loop-pulling. The Cα atom of Gly12, located near the center of the ring plane, was selected as the SMD atom for pulling. For each variant, three independent SMD replicas were performed. Additionally, the tail of the WT system was pulled in the opposite direction with Cα of Gly17 selected as the SMD atom. The SMD force profiles were extracted, and the corresponding work profiles were calculated using numerical integration with the Numpy trapz function.⁵⁵ The pulling velocity was set to 0.0000025 Å/step, and a spring constant of k = 5 kcal/mol.Ų was applied during the SMD simulations. The cumulative work during the pulling was calculated using: $W={\int }_0^{x}Fdx=v{\int }_0^{t}Fdt$,⁵⁶ where t = 0–40 ns is the duration of SMD simulation. From the SMD trajectories, representative structures were extracted to capture key conformational states: (i) threaded, (ii) during the process of unthreading, and (iii) unthreaded. These structures were used to perform additional cMD simulations ranging from 50 to 400 ns and identified stable and compact conformations. Candidate structures were also generated using LassoPred⁵⁷ for all syanodin I variants.

RESULTS AND DISCUSSION

Thermally Induced Unthreading and Structural Characterization of the Heat Sensitive Lasso Peptide Syanodin I.

Syanodin I and its branched-cyclic analog were investigated using a novel LC-TIMS-q-ECD-ToF MS platform. The LC and TIMS strategies have proven to be efficient in discriminating between the lasso and its corresponding branched-cyclic topology, but are limited in assigning the plug residues responsible for maintaining the C-terminal tail inside the macrolactam ring.^{22, 30} However, we previously demonstrated the utility of ECD for identifying the plug residues in lasso peptides, yet, this approach was lacking pre-separation techniques before the fragmentation events, which is crucial when analyzing mixtures.^{37, 39} The implementation of an ECD cell into an LC-TIMS-q-ToF MS platform therefore opens new avenues for the structural characterization of lasso peptides, as separation between lasso and branched-cyclic topologies can be achieved while the position of the plug residues can also be assigned in a single online experiment.

The MS analysis of both syanodin I and its branched-cyclic peptides analog resulted in the observation of a doubly charge state species ([M+2H]²⁺) at m/z 705.4 together with a lower intensity singly charge state species ([M+H]⁺) at m/z 1409.7 (Figures S3a–b). The LC, TIMS, and ECD profiles for the [M+2H]²⁺ ions of syanodin I (dark blue) and the branched-cyclic (red) peptide are presented in Figure 2. The LC dimension was very effective by providing a baseline separation between the lasso (RT₁ ~ 22.3 min, Figure 2a) and the branched-cyclic (RT₂ ~ 18.9 min, Figure 2b) peptide. The high resolution TIMS analysis resulted in the identification of multiple IMS bands for both the lasso (dark blue) and the branched-cyclic (red) topology, providing additional insights about the structures adopted by the two topologies. In addition, the two peptides exhibited distinct conformational states, for which the main IMS band of the branched-cyclic analog (CCS ~ 397 Ų, Figure 2b) adopted slightly more extended structures than syanodin I (CCS ~ 392 Ų, Figure 2a). The lasso and branched-cyclic topologies of syanodin I were further confirmed through LC-TIMS-q-ECD MS/MS experiments (Figures 2a–b). ECD spectra of the [M+2H]²⁺ ions displayed classic c′_i fragment series, consisting of c′₈ to c′₁₆ product ions (except c′₉ due to the presence of a Pro residue). Note that the absence of classic z_j^• can be explained by the lack of basic residues or amino acids being able to carry a charge in the C-terminal tail region (Ala9-Gly17), as ECD is a charge-driven fragmentation technique. In addition, the lasso structure exhibited an increase in hydrogen migration events near the molecular knot (formation of c₈^• to c₁₃^• product ions) when compared to the branched-cyclic analog, indicating that these hydrogen migration events occurred less frequently in the absence of an entangled structure. ECD MS/MS analysis pointed toward the Gln13 residue as being the lower plug residue maintaining the lasso structure of syanodin I, because the c₁₃^• fragment was the last product ion showing significant hydrogen migration events (Figure 2), in agreement with previously reported ETD MS/MS experiments.²⁸

Figure 2.

LC, TIMS, and ECD profiles of the [M+2H]²⁺ ions (m/z 705.4) of (a) syanodin I (dark blue), (b) branched-cyclic analog (red), (c) syanodin I after 1 h of incubation at 95 °C (light blue), and (d) syanodin I after 16 h of incubation at 95 °C (red). Typical hydrogen migration events are highlighted and labeled on the peptide cartoons (right of each panel). The macrolactam rings, the loop residues, the plugs, and the C-terminal tails are highlighted in green, blue, red and orange, respectively.

Syanodin I was subjected to heat treatment and directly compared to the branched-cyclic analog as reference. After 1 h of incubation at 95 °C, the extracted ion chromatogram (m/z 705.4) exhibited two additional baseline separated peaks, of which one aligned with the retention time of the branched-cyclic analog (RT₂ ~ 18.9 min), thus confirming that syanodin I is a heat sensitive lasso peptide (Figure 2c). The novel LC peak, eluting between the retention times of the original lasso and branched-cyclic topology (RT₃ ~ 20.0 min), displayed a change in the conformational state (TIMS profile), for which more compact structures (CCS ~ 384 Ų) were observed and separated as compared to the original lasso structure of syanodin I (light blue in Figure 2c). LC-TIMS-q-ECD MS/MS of the [M+2H]²⁺ ions of the new species displayed similar classic c′_i fragment series, consisting of c′₈ to c′₁₆ product ions. However, a change in the hydrogen migration events was observed when compared to the original lasso structure of syanodin I and the branched-cyclic analog. In fact, the formation of c₈^• to c₁₅^• product ions (light blue in Figure 2c), having a maximal c^•/c′ at Leu15, of the newly discovered species pointed toward the generation of another lasso structure formed as intermediate of the unthreading process that is using Leu15 as the lower plug residue. This suggests that the thermal unthreading of syanodin I involves a loop-pulling mechanism and represents the first example, where the intermediate plug is formed upon loop-pulling. After 16 h of incubation at 95 °C, the extracted ion chromatogram (m/z 705.4) showed that both lasso structures of syanodin are nearly fully converted into the unthreaded branched-cyclic analog (RT₂ ~ 18.9 min), which was corroborated through ECD MS/MS analysis (red in Figure 2d).

Factors Governing the Thermal Unthreading of Syanodin I.

Most lasso peptides contain bulky residues (Phe, Tyr and/or Trp) acting as plug residues on either side of the part of the tail threading the macrolactam ring, which greatly contributes to the thermal stability of the knot-like structure. In the case of syanodin I, only the Gln13 residue acts as a steric lock, whereas the residues in the part of the tail immediately above ring (Ala11, Gly12) are too small to provide further stabilization. The presence of a relatively small residue stabilizing the lasso structure is probably the main reason why syanodin I is sensitive to thermal unthreading. Hence, mutagenesis experiments were carried out to probe the factors governing the thermal unthreading of syanodin I. In the mutagenesis studies, amino acid substitutions were made exclusively in the C-terminal tail region (Gln13-Gly17) of syanodin I, due to the lasso peptide unthreading via a loop-pulling mechanism. This yielded the syanodin I variants [Q13A], [L15A], [Q13W], [L15W], and [L15W/A16W].

These syanodin I variants were examined using the same LC-TIMS-q-ECD-ToF MS/MS workflow described above to confirm the effect of each amino acid exchange on the respective lasso structure and its thermal stability. The syanodin I variants were confirmed through MS analysis, where the main doubly charge states species ([M+2H]²⁺) were observed at m/z 676.7, m/z 684.4, m/z 734.5, m/z 742.8, and m/z 799.0 for the [Q13A], [L15A], [Q13W], [L15W], and [L15W/A16W] variants, respectively (Figures S3c–g). The LC and TIMS profiles for the [M+2H]²⁺ ions of syanodin I [Q13A] (magenta), [L15A] (black), [Q13W] (yellow), [L15W] (green), and [L15W/A16W] (orange) variants are illustrated in Figures 3 and S4. After 1 h of incubation at 95 °C, the extracted ion chromatogram of [Q13A] (m/z 676.7) exhibited only two baseline separated peaks, for which one was aligned with the retention time of the respective lasso structure of the syanodin I [Q13A] variant (RT ~ 22.0 min, Figure 3a). The novel LC peak (RT ~ 19.6 min) displayed a change in the conformational state (TIMS profile), for which more extended structures (CCS ~ 383 Ų) were observed and separated as compared to the lasso structures of [Q13A] (CCS ~ 377 Ų). Hence, this peak was assigned to the unthreaded analog of [Q13A]. In addition, ECD MS/MS analysis of [Q13A] pointed toward the presence of a lasso structure with the Leu15 residue as the lower plug residue, suggesting that substituting the Gln by an Ala residue at position 13 was too small to maintain the original lasso structure of syanodin I (Figure S5). Similar results were obtained for [L15A] (m/z 684.4), in which the lasso (RT ~ 19.5 min and CCS ~ 381 Ų) and branched-cyclic (RT ~ 16.7 min and CCS ~ 388 Ų) structures were separated in LC and TIMS (Figure 3b). The [L15A] variant also did not yield another lasso structure during the unthreading process, suggesting that with the L15A substitution no residue in the tail beyond Gln13 is able to maintain the intermediate lasso structure, which is in agreement with results for the [Q13A] variant. Surprisingly, a similar pattern was also observed for [Q13W] (m/z 734.5). In fact, the [Q13W] variant directly converted to the branched-cyclic analog upon thermal unthreading (Figure 3c). The absence of an intermediate lasso structure for [Q13W], despite having a much bulkier lower plug residue than WT syanodin I, could be explained by the high energy barrier for overcoming Trp13 to begin with. The energy for accomplishing Trp13 slipping through the ring is likely already higher than the energy barrier needed for Leu15 slipping through the ring. Thus, during the unthreading of [Trp13] via a loop-pulling mechanism, both Trp13 and Leu15 must slip through the ring immediately once a suitable ring conformation has been adopted that allows Trp13 to pass.

Figure 3.

LC and TIMS profiles of the [M+2H]²⁺ ions of syanodin I (a) [Q13A] (magenta), (b) [L15A] (black), (c) [Q13W] (yellow), and (d) [L15W] (green) variants after 1 h/16 h of incubation at 95 °C. The macrolactam rings, the loop residues, the plugs, and the C-terminal tails are highlighted in green, blue, red, and orange, respectively.

A different result was observed for the syanodin I [L15W] and [L15W/A16W] variants after 1 h of incubation at 95 °C. The extracted ion chromatograms of [L15W] (m/z 742.8) and [L15W/A16W] (m/z 799.0) exhibited three baseline separated peaks, for which one peak was aligned for each variant with the retention time of the respective original lasso structure of [L15W] (RT ~ 26.0 min with CCS ~ 396 Ų, Figure 3d) and [L15W/A16W] (RT ~ 32.3 min with CCS ~ 417 Ų, Figure S4). In addition, the LC peaks at RT ~ 20.5 min (CCS ~ 404 Ų) and RT ~ 24.2 min (CCS ~ 409 Ų) were assigned to the unthreaded branched-cyclic structures resulting from complete [L15W] and [L15W/A16W] unthreading, respectively. The third LC peak at RT ~ 21.0 min and RT ~ 27.1 min displayed a change in the TIMS profile, for which more compact structures (CCS ~ 389 Ų and CCS ~ 406 Ų) were observed as compared to the lasso and branched-cyclic structures of [L15W] and [L15W/A16W], respectively. ECD MS/MS analysis for the [M+2H]²⁺ ions of the newly discovered species of [L15W] and [L15W/A16W] pointed toward the presence of lasso structures using the Trp15 residue as the lower plug residue (Figure S6). In addition, some amounts of the intermediate lasso structures with Trp15 as lower plug residue were still observed after 16 h of incubation at 95 °C, while the intermediate lasso structures using Leu15 as lower plug residue were always fully converted to the unthreaded branched-cyclic analog under these conditions (Figures 2d, 3d and S4). These findings suggest that increasing the size of the second plug residue at position 15 further stabilized the intermediate lasso structures against thermal unthreading. The mutational analysis therefore highlighted the importance of both Gln13 and Leu15 for stabilizing the original and newly discovered intermediate lasso structure of syanodin I upon thermal unthreading. A summary of the thermal unthreading pathways, discovered from the LC-TIMS-q-ECD-ToF MS/MS experiments, for syanodin I WT and variants are illustrated in Figure S7.

Thermal Unthreading Mechanism of Syanodin I.

The thermal unthreading mechanism of WT syanodin I was assessed using steered molecular dynamics (SMD) simulations, starting from nuclear magnetic resonance (NMR) structures. It is important to note that the thermal unthreading observed for lasso peptides is completely irreversible. SMD simulations of the WT syanodin I revealed that the C-terminal tail of the lasso peptide resists unthreading mainly due to the interactions involving Gln13 and Leu15. This was supported by the presence of two distinct force peaks in the SMD trajectory, corresponding to the sequential disentanglement of these residues (Figure 4a). The force peaks aligned with two energy barriers, consistent with the observations of an intermediate lasso state. In addition, representative structures were extracted from the SMD trajectories to capture key conformational states: (i) threaded, (ii) during the process of unthreading, and (iii) unthreaded. The theoretical ^TIMSCCS_N2 for each structure were calculated using the trajectory method algorithm implemented in the IMoS software (Table S1).^{58, 59} Candidate structures for each key conformational state were then proposed for those matching (< 4%) the experimental ^TIMSCCS_N2 (Figures 4 and S9 and Table S1).

Figure 4.

Force and work profiles as a function of time during the SMD simulations for syanodin I (a) WT, (b) [Q13A], (c) [Q13W], and (d) [L15W]. The sharp drop in the force (and a kink in the work) corresponds to the exit of a specific residue from the macrolactam ring during unthreading and the residues involved are indicated. The small drops in the force observed during the first unthreading event correspond to the passage of the backbone carboxyl and amine groups before sufficient force builds up to overcome the barrier for the sidechain passage.

SMD trajectories of [Q13A] showed that the disengagement of Ala13 occurred with minimal resistance (low energy barrier), while a significant energy barrier was observed for the disengagement of Leu15 (Figure 4b). This was consistent with a lasso structure with Leu15 as the lower plug residue, and confirmed that Ala13 was too small to maintain the original lasso structure of syanodin I. The [L15A] variant exhibited a single, small force peak at ~5 ns, corresponding to the exit of Gln13 (Figure S9a). The absence of a significant energy barrier for Ala15 highlighted that its reduced steric hindrance provided minimal resistance to disengagement, in agreement with the absence of a lasso intermediate structure. SMD trajectories of [Q13W] displayed a significant single force peak at ~20 ns, indicating simultaneous disengagement of both Trp13 and Leu15 residues (Figure 4b). The high energy barrier needed to pull out Trp13 explained the lack of an intermediate lasso structure, where it already overcome the energy barrier needed for Leu15 slipping through the ring. The [L15W] and [L15W/A16W] variants exhibited two force peaks, corresponding to the sequential disengagement of Gln13 and Trp15, in agreement with the presence of a lasso intermediate structure (Figures 4d and S9b). The force peak associated with Gln13 closely resembled to that of the WT in both variants (~1500 pN and ~250 kcal/mol). However, the force peak for Trp15 in [L15W] (~1500 pN) showed a reduced resistance compared to the one in [L15W/A16W] (~3000 pN), due to the simultaneous disengagement of Trp15 and Trp16. The SMD simulations were consistent with the experimental results, demonstrating that the thermal unthreading of syanodin I involves a loop-pulling mechanism.

The tail-pulling mechanism was explored by pulling the syanodin I tail, using the Cα of the Gly17 residue. The tail-pulling simulation for syanodin I WT resulted in smaller force peaks, corresponding to the passage of the Gly12, Ala11, and Pro10 residues (loop region) through the ring (Figure S10). Note that the tail-pulling SMD simulations were only performed on the syanodin I WT, as the investigated mutation sites (positions 13, 15, and 16) are located below the ring and would not affect the tail-pulling unthreading of the syanodin I variants.

The theoretical results suggest that a competition between the tail-pulling and loop-pulling thermal unthreading mechanisms is possible. A predominance for the tail-pulling unthreading due to the absence of bulky residues is observed (e.g., absence of any intermediate for [Q13W] Figure S11). However, the low abundance of the lasso intermediate species suggests that a small population could undergo a loop-pulling thermal unthreading.

Candidate structures generated using LassoPred⁵⁷ for the syanodin I variants are summarized in Figure S8. The predicted lasso structure with Gln13 as the upper plug and Leu15 as the lower plug residue do not agree with the experimental results. In the case of [L15A], the predicted structure has Gln13 as the upper plug, which should not be stable due to the absence of bulky residues below the ring to stabilize the lasso structure by maintaining the C-terminal tail sterically trapped inside the ring (Figure S8d).

CONCLUSIONS

A two-step thermal unthreading mechanism of the lasso peptide syanodin I was described. Mutational analysis experiments and LC-TIMS-q-ECD-ToF MS/MS analysis in combination with SMD simulations suggested that the syanodin I thermal unfolding mechanism is driven by two steric constraints (plugs) via a loop-pulling of the lasso peptide tail. An intermediate lasso peptide structure was observed after 1 h at 95 °C, utilizing the Leu15 residue as the plug residue. The mutational analysis experiments highlighted the importance of both Gln13 and Leu15 for stabilizing the original and newly discovered intermediate lasso structure of syanodin I upon thermal unthreading. In addition, SMD studies also pointed toward the presence of potential predominant thermal unthreading via a tail-pulling mechanism, due to the absence of bulky residues that could prevent for it.

This study provides the first example of an intermediate lasso peptide topology being formed during a thermal unthreading process. This feature was different from the previously reported thermal induced conformational switching of benenodin-1, for which an equilibrium was observed between two distinct lasso conformers.^{21, 27} The study of the thermal unthreading behavior has become one of the primary characterization steps for newly discovered lasso peptides. The workflow described herein can be easily extended to other class of RiPPs and provides an effective way to characterize thermal unfolding mechanisms of RiPPs.

Supplementary Material

Supporting Information contain additional Figures that illustrate the general classification criteria for lasso peptides, schematic of the LC-TIMS-q-ECD-ToF MS instrument, MS profiles of syanodin I, branched-cyclic, [Q13A], [L15A], [Q13W], [L15W], and [L15W/A16W] variants, LC and TIMS profiles of syanodin I [L15W/A16W] variant, ECD MS/MS of syanodin I [Q13A] variant, LC and ECD profiles of syanodin I [L15W] and [L15W/A16W] variants after 1h incubation at 95°C, Thermal unthreading mechanisms for syanodin I WT, [Q13A], [L15A], [Q13W], [L15W] and [L15W/A16W], LassoPred structure predictions for syanodin I WT, [Q13A], [Q13W], [L15A], [L15W], and [L15W/A16W} variants, Force and work profiles as a function of time during the SMD simulations for syanodin I [L15A], and [L15W/A16W], Force and work profiles as a function of time during the SMD simulations for syanodin I WT, using the Cα atom of Gly17 as the pulling atom in the tail-pulling configuration, Thermal unthreading pathways of syanodin I, showing the loop-pulling and tail-pulling configurations and the selected SMD atom, pulling direction, and mutation sites highlighted, Experimental and theoretical collision cross sections (CCS, Ų) for the proposed candidate structures of all investigated syanodin I variants, and the force field patch for connecting the N-terminus of Gly1 to the Cγ of Asp8. This material is available free of charge via the Internet at http://pubs.acs.org.