Crystal structure of the N domain of Lon protease from Mycobacterium avium complex

Abstract

Lon protease is evolutionarily conserved in prokaryotes and eukaryotic organelles. The primary function of Lon is to selectively degrade abnormal and certain regulatory proteins to maintain the homeostasis in vivo. Lon mainly consists of three functional domains and the N‐terminal domain is required for the substrate selection and recognition. However, the precise contribution of the N‐terminal domain remains elusive. Here, we determined the crystal structure of the N‐terminal 192‐residue construct of Lon protease from Mycobacterium avium complex at 2.4 å resolution，and measured NMR‐relaxation parameters of backbones. This structure consists of two subdomains, the β‐strand rich N‐terminal subdomain and the five‐helix bundle of C‐terminal subdomain, connected by a flexible linker，and is similar to the overall structure of the N domain of Escherichia coli Lon even though their sequence identity is only 26%. The obtained NMR‐relaxation parameters reveal two stabilized loops involved in the structural packing of the compact N domain and a turn structure formation. The performed homology comparison suggests that structural and sequence variations in the N domain may be closely related to the substrate selectivity of Lon variants. Our results provide the structure and dynamics characterization of a new Lon N domain, and will help to define the precise contribution of the Lon N‐terminal domain to the substrate recognition.

1. INTRODUCTION

Mycobacterium avium complex (Mac), often cause nontuberculous mycobacteria (NTM) pulmonary disease, which has been increasing worldwide and causes significant burden.1, 2 Therapy to Mac‐infected patients usually takes long time with combined medicines and high recurrence rates because of the high tolerance, and no effective vaccines are currently available in the market.3, 4, 5, 6 Structure determinations of Mac important functional proteins could help to develop new therapeutic strategies against this pathogen.

Lon protease, belonging to the AAA+ (ATPases associated with diverse cellular activities) superfamily, is evolutionarily conserved in prokaryotes and eukaryotic organelles.7 Its primary function is to selectively degrade abnormal proteins and regulatory proteins, and maintain the homeostasis in vivo.8, 9, 10 Lon functions as a homo‐hexamer and each subunit mainly consists of three functional domains, the N‐terminal non‐catalytic domain, the central ATPase (A) domain, and the C‐terminal proteolytic (P) domain, among which the feature of the N‐terminal domain is less defined.11, 12, 13 The N‐terminal domain of bacterial Lon consists of ~300 amino acids and is further divided into two or more subdomains including the N domain (~200 amino acids) and the coiled‐coil (CC) domain (~100 amino acids).13, 14, 15 Mutagenesis experiments have indicated that the N‐terminal domain is required for substrate selection and recognition, and the direct interaction between the N domain of the Escherichia coli Lon (EcLon) and the sul20 degron has been detected, but the precise contribution of the N domain remains unclear.12, 16, 17

The high resolution structure of the full‐length Lon has not been determined, and the determined structures were focused on the Lon ATPase and proteolytic domains, including the α subdomain of the Lon ATPase domain from Brevibacillus thermoruber,18 the Lon ATPase domains from Meiothermus taiwanensis and Bacillus subtilis,14, 19 and the Lon proteolytic domain from four organisms.14, 19, 20, 21, 22, 23 Whereas structural information about the N‐terminal domain is still scarce, only the N domains of B. subtilis and E. coli Lon proteases (BsLon‐N, EcLon‐N) have been structurally determined.14, 15, 24 However, sequence identities of BsLon‐N and EcLon‐N with the N domain of Mac Lon (MacLon‐N) are only 23% and 26%, respectively. In this work, we determined the crystal structure of MacLon‐N (residues 1–192, MacLon‐N192) at 2.4 å resolution using the single‐wavelength anomalous dispersion (SAD) method, as no structure has been deposited in the PDB database that has a sequence identity of more than 30% with MacLon‐N192. We also performed a structural comparison between MacLon‐N192 and its homologs. Meanwhile, we applied solution NMR spectroscopy to characterize its secondary structure and measure backbone ¹H—¹⁵N relaxation parameters. Our report provides initial data on the MacLon structure and will potentially contribute to defining substrate‐binding sites and understanding the N‐terminal domain functions of Lon in depth.

2. RESULTS AND DISCUSSION

2.1. Overall structure and backbone dynamics of MacLon‐N192

The N domain structures of E. coli and B. subtilis Lon proteases have been resolved with PDB IDs of 3LJC and 3M65, respectively.14, 15 According to their sequences together with homology modeling, we made two Mac Lon constructs, MacLon‐N231 (residues 1–231) and MacLon‐N192 (residues 1–192). Crystals of the MacLon‐N231 construct diffracted poorly even after optimization with around 1000 crystallization conditions, and its ¹H—¹⁵N HSQC spectrum shows crowded peaks in the spectral center (Figure 1a), which restricted the structure determination by either X‐ray crystallography or NMR. In comparison, the MacLon‐N192 construct could produce well‐diffracted crystals and well‐resolved ¹H—¹⁵N HSQC spectrum (Figure 1a) and is better for the further structure investigation.

Figure 1.

Structure and backbone dynamics of MacLon‐N192. (a) 2D ¹H—¹⁵N HSQC spectra of MacLon‐N192 (black) and MacLon‐N231 (orange). Assignments for MacLon‐N192 are indicated. (b) Crystal structure of MacLon‐N192 shown in cartoon representation with two different directions of views. Secondary structure elements are labeled on the structure. (c) ¹H—¹⁵N relaxation parameters (bottom) for MacLon‐N192. Secondary structures predicted by TALOS+ (α‐helix: red; β‐strand: black) based on assigned chemical shifts agree well with the crystal structure regardless of unassigned regions (top). The crystal secondary structure elements are labeled at the top for comparison

The crystal structure of MacLon‐N192 was determined in the space group of P3₁21 by the SAD method at 2.4 å resolution. In the crystal, each asymmetric unit contains one monomer, and the size exclusion chromatography also illustrates that MacLon‐N192 exists as a monomer (Figure S1). The X‐ray diffraction data and refinement statistics are shown in Table 1. The structure of MacLon‐N192 consists of two distinct subdomains connected by a seven‐residue linker (V106‐T112) (Figure 1b). The N‐terminal subdomain (NS1) is rich in β‐strands, consisting of six β‐strands (β1: 5–12, β2: 20–26, β3: 46–50, β4: 59–71, β5: 74–95, β6: 98–105) and one α‐helix (α1: 29–40) among which strands β4 and β5 are much longer than the other four strands. Moreover, three β‐sheets are formed, one of which is a mixed parallel/antiparallel sheet formed by strands β1/β3/β4/β5, and the other two are antiparallel sheets formed by strands β2/β5/β4 and β1/β6/β5, respectively. The α1‐helix connecting strands β2 and β3 is surrounded by three sheets in a semicircle. The C‐terminal subdomain (NS2) consists of five α‐helices (α2: 113–133, α3: 136–144, α4: 148–157, α5: 163–171, α6: 175–187) that form a compact five‐helix bundle. Helices α3 and α4 are close to the NS1 subdomain, especially to the loop between strands β1 and β2 (L_β1β2), partially contributing to the hydrophobic interactions between the two subdomains.

Table 1.

Data collection and refinement statistics for MacLon‐N192

Data collection	Native	SeMet
Wavelength (å)	0.9785	0.9785
Space group	P3121	P3121
Unit‐cell parameters	a = b = 58.6,	a = b = 58.7,
	c = 118.3,	c = 118.8,
	α = β = 90°, γ = 120°	α = β = 90°, γ = 120°
Resolution range (å)a	50.00–2.40 (2.44–2.40)	50.00–2.50 (2.54–2.50)
Total observations	178997	146954
Unique reflections	9758	8447
Redundancya	18.3 (12.4)	17.4 (11.4)
Completeness (%)a	91.9 (80.7)	78.6 (54.8)
I/σ (I)a	41.6 (3.0)	30.2 (2.0)
Molecules in asymmetric unit	1	1
R merge (%)a	7.6 (63.0)	9.7 (68.9)
CC1/2 a	1 (0.968)	1 (0.935)
CC*a	1 (0.99)	1 (0.98)
Refinement
Resolution range (å)a	24.82–2.40 (2.48‐2.40)
R work/R free (%)b	22.7/26.8
Wilson B factor (å2)	34.8
Average B factor (å2)
Protein	46.51
Water	42.26
Number of atoms
Protein	1409
Water	28
R.m.s. deviations
Bond lengths (å)	0.009
Bond angles (°)	1.11
Ramachandran plot (%)
Most favored	95.65
Allowed	3.80
Outlier	0.54

^aValues in parentheses are for the highest resolution shell.

^b R _work was calculated with 95% of the unique reflections used for refinement, whereas R _free was calculated with the remaining 5% of the unique reflections.

Characterizing dynamics of proteins is helpful for understanding their structures and functions at a detailed level. In this study, we labeled MacLon‐N192 with ¹³C and ¹⁵N isotopes, assigned more than 83% backbone resonances (BMRB accession code: 12025), and measured backbone ¹H—¹⁵N relaxation parameters of R₁, R₂, and NOE (Figure 1a,c). The secondary structures were predicted by TALOS+25 based on the assigned chemical shifts. MacLon‐N192 in solution exhibits the similar secondary structures as those in crystal except for the unassigned regions such as parts of strand β2 and helix α3 and the C‐terminal portion of strand β5 that was predicted as loop. Sequence variations of the relaxation parameters were observed for residues I71‐T76 in the loop L_β4β5 and the N‐terminal portion of strand β5, residues S93‐A98 in the loop L_β5β6 and the C‐terminal portion of strand β5, and residues A109‐T112 in the loop L_β6α2. These regions are highly flexible with ¹H—¹⁵N NOE values <0.6, among which the most flexible one is loop L_β6α2 with ¹H—¹⁵N NOEs even <0.4. Moreover, these regions show significantly higher R₁ and lower R₂ values than most of structural regions, indicating the nanosecond motion. Interestingly, residues in loop L_β3β4 have ¹H—¹⁵N NOE values ≥0.7, reflecting the relatively restricted high‐frequency motion that results from the conformation stabilized by H‐bond between E53 and R55 (i → i + 2 turn), salt‐bridge between D54 and R55, and hydrophobic interactions between L52 and Y56. Residues P50 and P57 may also contribute to the stabilization of loop L_β3β4 by mediating the formation of turn‐like structures. Proline‐stabilized conformations have recently been found for in vitro designed bicyclic peptides with irregular turns.26 Residues V17 and L18 in loop L_β1β2, stabilized by hydrophobic interactions with the NS1 and NS2 subdomains respectively, show ¹H—¹⁵N NOE values ≥0.8. Unexpectedly, the structural flexibility was observed for parts of strand β5, which was potentially caused by its highly distorted conformation. The structural distortion may shift the strand chemical shifts to random coil as well.

2.2. Structural comparison with homologs

Crystal structures of the N domains of B. subtilis and E. coli proteases (BsLon‐N and EcLon‐N) have been solved, and both consist of two subdomains (NS1, NS2) connected by an extended linker.14, 15 The NS1 subdomains of MacLon‐N192 and BsLon‐N209 adopt similar structures with an RMSD of 2.3 å for 100 residue pairs superposed by DALI (Z‐score: 15.3), whereas the NS2 subdomains are significantly different with a compact five‐helix bundle for MacLon‐N192 but a differently oriented four‐helix bundle for BsLon‐N209 [Figure S2(a)]. Furthermore, MacLon‐N192 is in a monomeric form as identified by size exclusion chromatography (Figure S1), while BsLon‐N209 exists as a homo‐dimer.14 The structure and amino acid sequence differences of the NS2 subdomains might contribute to the different oligomeric states. BsLon‐N209 exists as a domain‐swapped dimer through intermolecular interactions including coiled–coil interaction and hydrophobic interaction mainly resulting from the helical bundle of the NS2 subdomain,14 which are not present in the NS2 domain of MacLon‐N192 without a long α helix and specific hydrophobic pockets [Figure S2(b,c)].

In addition, the DALI search provided a hypothetical protein BPP1347 with unknown functions showing the similar structure to the N domains of both EcLon and MacLon. The structural comparisons among EcLon‐N, BsLon‐N, and BPP1347 have been extensively investigated.14, 15, 24 In this study, we focus on the structural comparison between MacLon‐N192 and EcLon‐N.

The overall structure of MacLon‐N192 is very similar to that of EcLon‐N245 with an RMSD of 2.3 å for 185 residue pairs superposed by DALI (Z‐score: 19.6) (Figure 2a). Because the crystal structure of EcLon‐N245 was solved under the condition of poor electron density for the NS1 subdomain,15 for accuracy, only the structure of the NS2 subdomain of EcLon‐N245 together with the structure of EcLon‐N119 (the isolated NS1 subdomain) was used for the comparison (Figure 2a). The significant diversities are observed in the NS1 subdomain. The region between strands β3 and β4 comprises eight residues in MacLon‐N192 but 16 residues in EcLon‐N, and includes a turn structure in MacLon‐N192 but a four‐residue helix in EcLon‐N. The strand β5 in MacLon‐N192 is four‐residue longer than that in EcLon‐N, and the structures of the C‐terminal regions of strands β4 and β5 are more distorted in MacLon‐N192. In addition, the region representing helix α3 in the NS2 subdomain of MacLon‐N192 is rotated ~ 90° in EcLon‐N mainly caused by the proline at the beginning of the helix. The very long C‐terminal helix observed in EcLon‐N245 is only seen for a quarter in length in MacLon‐N192 due to the truncation, and the compact structure is almost not affected by the truncated part, which can be observed from the crystal structure and is also suggested by the well overlaid ¹H—¹⁵N HSQC spectra between MacLon‐N192 and MacLon‐N231 constructs.

Figure 2.

Comparison of MacLon and its homologs. (a) Superposition of MacLon‐N192 (green, PDB ID: 6IHG) and EcLon‐N245 (orange, PDB ID: 3LJC)/EcLon‐N119 (blue, PDB ID: 2ANE) in different directions of views. Part of the long C‐terminal helix of EcLon‐N245 are deleted in comparison. Arrows indicate the structural deviation of MacLon‐N192 from EcLon‐N. (b) Sequence alignment of MacLon‐N192 with the corresponding fragments of Escherichia coli Lon (EcLon; UniProt ID P0A9M0), Bacillus subtilis Lon (BsLon; UniProt ID P37945), Mycobacterium tuberculosis Lon (MtLon; UniProt ID A0A0T9M6Q7), and human Lon (HuLon; UniProt ID P36776). (c) Superposition of possible binding residues of MacLon‐N192 (green) and EcLon‐N245 (orange)/EcLon‐N119 (blue) to the potential hydrophobic substrates

Structural variations of the N domains even being small may contribute to the specific substrate recognition in different species of Lon proteases. The substrate selection of Lon proteases is organism dependent, and three domains were all observed functioning in the substrate selectivity revealed by the domain swap experiments.17 The N domains of bacteria and human Lon proteases share the sequence identity less than 10% (Figure 2b), whereas the sequence identity of the AP domains is more than 30%,19 suggesting that the N domain may contribute more to the substrate selectivity. Thus, the different conformations of the region between strands β3 and β4 probably indicate the different substrate binding for EcLon and MacLon although no substrates have been identified for MacLon so far. Moreover, a similar substrate cleavage pattern exists in various Lon proteases, and the hydrophobic or nonpolar amino acids such as Leu, Ala, Val, Pro, and Phe are preferred at the preceding position of the cleavage site,17 suggesting that hydrophobic interactions between Lon and its substrates may be important for the substrate selection and cleavage. By analysis of conserved residues in EcLon‐N, a hydrophobic patch mainly consisting of V22, P24, L66, V79, I138, Y141, L145, and Y157 was found and an interaction with the substrate β20 rich in aromatic residues was suggested (Figure 2c).15, 27 Interestingly, a similar hydrophobic patch consisting of V26, P28, L78, L91, V152, L156, I159, and A171 was also observed in MacLon‐N (Figure 2c). Thus, the hydrophobic interactions between the N domain and the substrate may involve in the initial substrate selection and recognition.

3. CONCLUSION

In summary, we determined the crystal structure of MacLon‐N192 at 2.4 å, obtained the relaxation parameters of backbones, and performed the structural comparison between MacLon‐N192 and EcLon‐N. The MacLon‐N192 structure consists of two subdomains connected by a flexible linker and is similar to the overall structure of EcLon‐N. The relaxation parameters reveal two relatively rigid loops L_β1β2 and L_β3β4 in which the L_β1β2 is stabilized by the interactions with two subdomains and the L_β3β4 is stabilized by H‐bond, salt‐bridge, hydrophobic interactions, and specific prolines. The sequence and structural diversities between MacLon‐N192 and EcLon‐N may contribute to the substrate selectivity, and the hydrophobic interactions between Lon protease and its substrates may involve in the initial substrate selection and recognition. More structures of the N domains of various Lon proteases need to be solved to uncover the precise contribution to substrate recognition.

4. MATERIAL AND METHODS

4.1. Protein expression and purification

The gene for MacLon‐N192 (residues 1–192) was synthesized by GenScript and incorporated into the pET28a vector with an N‐terminal His₆ tag followed by a thrombin cleavage site via NdeI/XhoI restriction sites. The protein was expressed in E. coli BL21(DE3) cells. The bacteria were grown in LB medium at 37°C until the OD₆₀₀ reached 0.6–0.8, and then 1 mM IPTG was added to induce protein expression for 5 hr. To express ¹³C/¹⁵N‐ or Selenomethionine (SeMet)‐labeled proteins, the cells were grown in 1 L of LB medium to an OD₆₀₀ of ~0.8 and then centrifuged at 4000 rpm for 4 min at 4°C. For ¹³C/¹⁵N‐labeled MacLon‐N192, the cell pellets were resuspended in 0.5 L of M9 medium containing 2 g of uniformly ¹³C‐labeled glucose and 0.5 g of ¹⁵N‐labeled ammonium chloride, and for SeMet‐labeled MacLon‐N192, the cell pellets were resuspended in 0.5 L of M9 medium containing 50 mg each of SeMet, Val, Leu, and Ile and 100 mg each of Phe, Lys, and Thr. After 30 min of growing at 37°C, 1 mM IPTG was added to induce protein expression for 5 hr. Then the cells were centrifuged at 4°C and 9000 rpm for 4 min and stored at −80°C.

The cell pellets were suspended in the lysis buffer (50 mM Tris–HCl, 300 mM NaCl, pH 8.0), and lysed in an ice‐water bath for 30 min using ultrasonic cell crusher VCX800 (Sonic & materials). The lysate was centrifuged at 4°C and 12,000 rpm for 30 min. The supernatant containing the soluble protein was loaded onto a Ni‐NTA column and the column was washed stepwise with the lysis buffer containing 20 mM and 40 mM imidazole. The protein was finally eluted using the lysis buffer containing 250 mM imidazole. After removal of imidazole by dialysis against the buffer of 20 mM Tris–HCl (pH 8.0) and 150 mM NaCl, the thrombin protease (Sigma) was added to cleave the protein overnight at 4°C. The cleaved MacLon‐N192 was further purified using a Superdex 75 10/300 GL gel‐filtration column (GE Healthcare) and the collected fraction was concentrated using Amicon Ultra‐15 (Millipore). The protein purity was verified by SDS‐PAGE and the protein concentration was measured using NanoVue ultra micro spectrophotometer (GE Healthcare).

4.2. Crystallization

Crystallization screening of MacLon‐N192 was performed at 22°C using crystallization kits (Hampton Research) and the sitting‐drop vapor‐diffusion method. The protein was prepared at a concentration of 13 mg/mL for crystallization. The best crystals were obtained within 8 days under the condition of 100 mM sodium acetate trihydrate (pH 4.8) and 3.3 M sodium formate for both native and SeMet‐labeled MacLon‐N192.

4.3. Data collection and structure determination and refinement

The X‐ray diffraction data were collected using a pilatus3 6M detector on the 19U1 beamline at Shanghai Synchrotron Radiation Source. All crystal data were indexed, integrated, and scaled by HKL2000.28 The crystal structure of MacLon‐N192 was determined using the SAD method. The initial phase was calculated with CCP4 Crank2 program29 by using three selenium atoms identified in one asymmetric unit of MacLon‐N192 and the SAD method. The crystal structures were refitted manually in Coot30 and refined with Phenix.refine.31 The crystal diffraction data and refinement statistics are shown in Table 1. The crystal structure of MacLon‐N192 was deposited to Protein Data Bank (PDB) with an accession code of 6IHG.

4.4. NMR experiments

The ¹⁵N‐ and ¹³C/¹⁵N‐labeled samples were prepared by dissolving 0.5 mM proteins in NMR buffer (20 mM Na₂HPO₄‐NaH₂PO₄, 100 mM NaCl, 0.02% NaN3, 10% D₂O, pH 6.4). All NMR experiments were conducted at 308 K on a Bruker 850 MHz spectrometer equipped with a z‐gradient triple‐resonance cryoprobe. Two‐ and three‐dimensional spectra of ¹H—¹⁵N HSQC, ¹H—¹³C HSQC, ¹H—¹³C‐¹⁵N HNCACB, CBCA(CO)NH, HNCA, HN(CO)CA, HNCO, and HN(CA)CO were recorded for backbone resonance assignments. For ¹⁵N T₁ measurements, the delay times were set to 10, 50, 100 (×2), 200, 400, 600, 800 (×2), 1200, 1600, and 2000 ms, and for ¹⁵N T₂ measurements, the delay times were set to 16.96, 33.92 (×2), 50.88, 67.84, 84.8, 101.76, 118.72, 135.68 (×2), 152.64, and 169.6 ms. The repeated experiments were used to estimate uncertainties of T₁ and T₂ values. In the ¹H—¹⁵N NOE experiments, a delay of 2 s was followed by ¹H saturation of 3 s, and in the control experiments without ¹H saturation, a total delay of 5 s was applied. All NMR spectra were processed using NMRPipe32 and analyzed using NMRFAM‐SPARKY.33 The assigned chemical shifts of MacLon‐N192 were deposited to Biological Magnetic Resonance Data Bank (BMRB) with an accession code of 12025.

CONFLICT OF INTEREST

The authors declare no conflicts of interest.

Supporting information

Supplementary Figure S1 Size exclusion chromatography analysis of MacLon‐N192.

Supplementary Figure S2. Structural comparison of MacLon‐N192 with its homologs.

Chen X, Zhang S, Bi F, et al. Crystal structure of the N domain of Lon protease from Mycobacterium avium complex . Protein Science. 2019;28:1720–1726. 10.1002/pro.3687