Crystallization and X-ray analysis of Borrelia burgdorferi β-barrel assembly machinery A

Borrelia burgdorferi β-barrel assembly machinery A (BbBamA) is an essential translocator located in the outer membrane. The N-terminal periplasmic polypeptide-transport-associated (POTRA) domains of BbBamA are vital for assembly of the BAM complex. Here, the production, crystallization and initial X-ray crystallographic analysis of the three N-terminal POTRA domains of BbBamA (BbBamA-POTRA P1–P3) are reported.

Abstract

Mitochondria, chloroplasts and several species of bacteria have outer membrane proteins (OMPs) that perform many essential biological functions. The β-barrel assembly machinery (BAM) complex is one of the OMPs of Borrelia burgdorferi, the pathogenic spirochete that causes Lyme disease, and its BamA component (BbBamA) includes a C-terminal β-barrel domain and five N-terminal periplasmic polypeptide-transport-associated (POTRA) domains, which together perform a central transport function. In the current work, the production, crystallization and X-ray analysis of the three N-terminal POTRA domains of BbBamA (BbBamA-POTRA P1–P3; residues 30–273) were carried out. The crystals of BbBamA-POTRA P1–P3 belonged to space group P2₁, with unit-cell parameters a = 45.353, b = 111.538, c = 64.376 Å, β = 99.913°. The Matthews coefficient was calculated to be 2.92 ų Da⁻¹, assuming the presence of two molecules per asymmetric unit, and the corresponding solvent content was 57.9%. Owing to the absence of an ideal homology model, numerous attempts to solve the BbBamA-POTRA P1–P3 structure using molecular replacement (MR) failed. In order to solve the structure, further trials using selenomethionine derivatization are currently being carried out.

1. Introduction  

Gram-negative bacteria are packaged by double membranes, and some proteins located in the bacterial outer membrane (OM) are critical for transferring substances between the cells and their surroundings. The β-barrel assembly machinery (BAM) complex is an essential outer-membrane protein (OMP) complex that forms a channel for the transport of specific molecules (Webb et al., 2012 ▸). For example, the BAM complex allows the uptake of ions and nutrients, the expulsion of waste products and antibiotics, and the translocation of effectors such as bacterial enzymes and toxins into the extracellular setting (Silhavy et al., 2010 ▸). Furthermore, the assembly of many OMPs into the bacterial inner surface of the outer membranes requires the BAM complex. In Gram-negative bacteria, nascent OMPs bound to periplasmic chaperones transit across the periplasm to the outer membrane. In this process, a signature sequence termed the ‘β-signal’ that is often found in the C-terminal β-strand of OMPs can bind to the central subunit (β-barrel assembly machinery A; BamA) of the BAM complex to create a hybrid barrel. In this hybrid barrel model, exposed unpaired β-strands of BamA may be used to nucleate the proper folding and membrane insertion of OMPs (Noinaj et al., 2014 ▸).

In fact, BamA is the channel core component of the BAM complex, which belongs to the Omp85 protein family, and has homologues that are found not only in prokaryotes but also in eukaryotic organelles (such as mitochondria and chloroplasts; Day et al., 2014 ▸). Structural analysis has shown that BamA proteins are evolutionarily well conserved, and all contain a C-terminal integral OM β-barrel domain protein and one or more N-terminal periplasmic polypeptide-transport-associated (POTRA) domains (Fig. 1 ▸ a; Gentle et al., 2005 ▸). The β-barrel domain is responsible for forming the transfer channel, and the POTRA domains serve as specific interfaces for protein–protein interactions in the formation of the BAM complex (Ni et al., 2014 ▸). Furthermore, the BamA POTRA domains might mediate initial substrate binding and subsequent folding (Knowles et al., 2008 ▸). BamA proteins from different organisms appear to display different structures and functions. The POTRA domains of BamA from Escherichia coli (EcBamA) interact with BamB, BamC, BamD and BamE (Bakelar et al., 2016 ▸; Kim et al., 2012 ▸). Neisseria meningitidis was the first member of the Betaproteobacteria class in which a BAM complex was identified (Voulhoux et al., 2003 ▸). Compared with the BAM complex of E. coli, that of N. meningitidis differs in that it lacks the BamB protein but has an additional protein called RmpM (Volokhina et al., 2009 ▸). Moreover, unlike the case in N. meningitidis, in the BAM complex of the Betaproteobacteria class organism N. gonorrhoeae only one reported accessory protein (BamD) interacts with BamA (NgBamA; Fussenegger et al., 1996 ▸). Similarly, Borrelia burgdorferi, the pathogenic spirochete that causes Lyme disease, also has a BAM complex (Lenhart & Akins, 2010 ▸). A co-immunoprecipitation assay showed that B. burgdorferi BamA (BbBamA) can form a high-molecular-weight complex by interacting with BB0028 and BB0324. BB0324 is an ortholog of the E. coli and N. meningitidis BamD lipoproteins. In contrast, BB0028 shares no obvious homology with any other known BAM-complex protein (Lenhart et al., 2012 ▸). These observations indicate that BamA proteins, and especially their POTRA domains, have distinct functions during the formation of BAM complexes in different species. Although no research has been reported on the interaction between the B. burgdorferi BAM complex and host cells, the B. burgdorferi BAM complex and its components have both become important targets for the development of novel vaccines and antibiotics (Dunn et al., 2015 ▸).

Figure 1.

Domain architecture and purification of BbBamA. (a) Schematic diagram of the domain organization in the primary sequence of BbBamA. (b) Chromatograms of a sample containing BbBamA-POTRA P1–P3 on a Superdex 75 Increase 10/300 GL column (GE). The blue and red lines indicate the absorbance at 280 and 254 nm, respectively. The inset shows a 15% SDS–PAGE analysis of the peak containing the final sample.

Many BamA and BAM-complex structures, including those of NgBamA, EcBamA and EcBAM (the BAM complex from E. coli), have now been determined using X-ray crystallo­graphy or cryo-electron microscopy (cryo-EM) (Noinaj et al., 2013 ▸; Sikora et al., 2018 ▸; Gu et al., 2016 ▸; Iadanza et al., 2016 ▸). These structural studies have allowed the diversity in the assembly of BAM complexes to be determined, to some degree, by the differences between the BamA POTRA domains. The architecture of the BAM complex from B. burgdorferi is poorly understood, and no structure of BbBamA is available in the Protein Data Bank (PDB). Determining the structure of BbBamA should provide essential insight into its crucial function in the B. burgdorferi BAM complex. Here, we describe the cloning, expression, purification, crystallization and preliminary X-ray diffraction analysis of the N-terminal three POTRA domains of BbBamA and lay the foundation for further structural and functional analyses.

2. Materials and methods  

2.1. Macromolecule production  

Codon-optimized cDNA for the three N-terminal POTRA domains of BbBamA (BbBamA-POTRA P1–P3; residues 30–273) was synthesized by GENEWIZ. The BbBamA-POTRA P1–P3 forward primer included a BamHI restriction site, and the reverse primer included an XhoI restriction site. The PCR products were inserted into the pET-28a vector (Novagen). The recombinant plasmid for BbBamA-POTRA P1–P3 was transformed into E. coli BL21(DE3) cells (TransGen). The cells were cultured at 37°C in 800 ml LB medium supplemented with 50 mg l⁻¹ kanamycin. Once the optical density (OD₆₀₀) had reached 0.6, the temperature of the culture was reduced to 16°C. Protein expression was then induced using 0.5 mM isopropyl β-d-1-thiogalactopyranoside (IPTG) for an additional 18 h. The harvested cells were resuspended in 15 ml lysis buffer (20 mM HEPES, 500 mM NaCl pH 7.0) per gram of pellet and disrupted with a low-temperature ultrahigh-pressure cell disrupter (JNBIO) at 4°C.

The lysate was centrifuged at 12 000 rev min⁻¹ for 30 min at 4°C to remove cell debris. Affinity purification was performed by passing the supernatant through 2.5 ml Ni–NTA affinity resin (GE) three times. The resin was washed four times with 80 ml lysis buffer. Recombinant BbBamA-POTRA P1–P3 was eluted with 40 ml wash buffer (20 mM HEPES, 500 mM NaCl, 500 mM imidazole pH 7.0) and then exchanged into low-salt buffer (20 mM HEPES, 20 mM NaCl pH 7.0) using a 10 kDa molecular-weight cutoff filter unit (Millipore). The sample was subsequently further purified using a HiTrap Heparin HP column (1 ml; GE). The purified BbBamA-POTRA P1–P3 was exchanged into storage buffer (20 mM HEPES, 200 mM NaCl pH 7.0) and concentrated to 30 mg ml⁻¹. The protein homogeneity was analysed by size-exclusion chromatography using a Superdex 75 Increase 10/300 GL column (GE; Fig. 1 ▸ b). The purity of BbBamA-POTRA P1–P3 was assessed by 15% SDS–PAGE. Macromolecule-production details are described in Table 1 ▸.

Table 1. Macromolecule production.

Source organism	B. burgdorferi
Forward primer†	CGGGATCCAAAATCATTAAAGGT
Reverse primer‡	CCCTCGAGTTATTCGCTCAGAAAG
Expression vector	pET-28a
Expression host	E. coli BL21(DE3)
Complete amino-acid sequence of the construct produced§	MGSSHHHHHHSSGLVPRGSHMASMTGGQQMGRGS KIIKGINFEGLKNKKERDFINILKPYIGVSYSNEIFDKLQIDLYSLDYFSGLIKPIFKIDGEDLFITFIVKEKSLVNSVVFSDSSRVFWNSELVEKVNIKTNEPLNLASVNKGIGKLEEMYKDMGYLEVSANFEIKEEGNLVDIIFNIVAGPKYVVKGIDFEGNLSFKSSTLRKSLASRVVSLFSDGKYLKSNVDKDKRQLESFYKNNGYIDVKIINSTVDIKDSLKDSKRLEKEVFLKYFLSE

^†The BamHI site is underlined.

^‡The XhoI site is underlined.

^§The tag residues are shown in italics. The BbBamA-POTRA P1–P3 residues are underlined.

2.2. Crystallization  

Initial crystallization trials for BbBamA-POTRA P1–P3 were performed in a 48-well format using a 1:1 ratio of well solution to protein solution (20 mg ml⁻¹) by screening at 289 K with commercial crystal screening kits including the Wizard I–IV kits from Emerald BioSystems and the Index, Crystal Screen, Crystal Screen 2, SaltRx 1, SaltRx 2, PEG/Ion and PEG/Ion 2 screens from Hampton Research. Small crystals of BbBamA-POTRA P1–P3 first appeared after one week in 1.26 M ammonium sulfate, 0.1 M MES sodium hydroxide pH 6.0 (Fig. 2 ▸ a). Further optimization of the crystallization conditions was performed by adjusting the pH, the temperature and the concentrations of ammonium sulfate and BbBamA-POTRA P1–P3, and the final optimized crystallization condition was 1.2 M ammonium sulfate, 0.1 M MES sodium hydroxide pH 5.5 with 15 mg ml⁻¹ BbBamA-POTRA P1–P3. Crystals were produced by the hanging-drop method at 285 K (Fig. 2 ▸ b). In order to verify that BbBamA-POTRA P1–P3 had indeed crystallized, several optimized crystals were picked up with a loop and subjected to 15% SDS–PAGE (Fig. 2 ▸ c). The major protein band above 25 kDa was excised and sent for mass-spectrometric analysis. Crystallization information is summarized in Table 2 ▸.

Figure 2.

BbBamA-POTRA P1–P3 crystals and X-ray diffraction. (a) Initially obtained BbBamA-POTRA P1–P3 crystals. (b) Optimized BbBamA-POTRA P1–P3 crystals. (c) 15% SDS–PAGE analysis of washed optimized BbBamA-POTRA P1–P3 crystals. (d) Diffraction image from the best BbBamA-POTRA P1–P3 crystal. (e) The peptide coverage of BbBamA-POTRA P1–P3 crystals. The peptide coverage detected by fingerprint mass spectrometry (blue) indicated that it was BbBamA-POTRA P1–P3 that crystallized.

Table 2. Crystallization information.

Method	Sitting drop (screening), hanging drop (production)
Temperature (K)	285
Protein concentration (mg ml−1)	15
Buffer composition of the BbBamA-POTRA P1–P3 solution	20 mM HEPES, 200 mM NaCl pH 7.0
Composition of the reservoir solution	1.2 M ammonium sulfate, 0.1 M MES sodium hydroxide pH 5.5
Volume of reservoir	100 µl (screening), 200 µl (production)
Volume and ratio of drop	2 µl, 1:1 ratio of protein:reservoir solution

2.3. Data collection and processing  

Crystals were flash-cooled in liquid nitrogen after first soaking them in reservoir solution containing 25%(v/v) glycerol as a cryoprotectant. A data set was collected to a resolution of 2.00 Å on the BL17U beamline at the Shanghai Synchrotron Radiation Facility (SSRF), Shanghai, People’s Republic of China (Fig. 2 ▸ d). All collected diffraction data sets were indexed, integrated and scaled with HKL-3000 (Minor et al., 2006 ▸). The Matthews coefficient and solvent content were calculated using phenix.xtriage from the Phenix package (Liebschner et al., 2019 ▸). Data-collection and processing statistics are summarized in Table 3 ▸.

Table 3. Data collection and processing.

Values in parentheses are for the highest resolution shell.

Diffraction source	BL17U, SSRF
Wavelength (Å)	0.97918
Temperature (K)	100
Detector	EIGER 16M
Crystal-to-detector distance (mm)	300.00
Rotation range per image (°)	0.5
Total rotation range (°)	360
Exposure time per image (s)	0.3
Space group	P21
a, b, c (Å)	45.353, 111.538, 64.376
α, β, γ (°)	90, 99.913, 90
Mosaicity (°)	0.3
Resolution range (Å)	50.00–2.00 (2.03–2.00)
Total No. of reflections	160477
No. of unique reflections	41610 (2108)
Completeness (%)	98.1 (99.6)
Multiplicity	3.9 (3.3)
〈I/σ(I)〉	26.2 (2.2)
R merge † (%)	9.4 (32.8)
Overall B factor from Wilson plot (Å2)	35.83

^† R _merge = , where 〈I(hkl)〉 is the mean of the observations I _i(hkl) of reflection hkl.

2.4. Choosing models for molecular replacement  

The BbBamA-POTRA P1–P3 sequence was aligned with the sequences of other proteins in the PDB using BLASTP (https://blast.ncbi.nlm.nih.gov/Blast.cgi?PAGE=Proteins). Only the sequence of the POTRA domains of NgBamA (NgBamA-POTRA P1–P3; PDB entry 4k3b; Noinaj et al., 2013 ▸) was selected from the database (Fig. 3 ▸). At the same time, we constructed a predicted three-dimensional model of BbBamA-POTRA P1–P3 by submitting the primary sequence to the Phyre2 server (http://www.sbg.bio.ic.ac.uk/phyre2/html/page.cgi?id=index). According to a search using the DALI server (Holm & Rosenström, 2010 ▸), the Z-score for comparison of the predicted BbBamA-POTRA P1–P3 model and NgBamA-POTRA P1–P3 is 17.0 and the root-mean-square deviation (r.m.s.d.) for 230 aligned residues is 0.9 Å, which indicate that structural similarities exist between the proteins. The experimentally determined structure of NgBamA-POTRA P1–P3 and the predicted BbBamA-POTRA P1–P3 model were then used as molecular-replacement (MR) search models in MOLREP and Phaser-MR in the CCP4 suite (Winn et al., 2011 ▸) and in Phaser-MR in the Phenix package (Liebschner et al., 2019 ▸), respectively.

Figure 3.

Sequence alignment of BbBamA-POTRA P1–P3 and NgBamA-POTRA P1–P3. Residues in red with blue frames are conserved, and those highlighted in red are identical. Numbers correspond to the BbBamA-POTRA P1–P3 sequence.

3. Results and discussion  

Recombinant BbBamA-POTRA P1–P3 was purified from E. coli BL21(DE3) cells using a two-step method (Ni–NTA and HiTrap Heparin HP affinity). Protein purity and homogeneity were assessed by 15% SDS–PAGE and size-exclusion chromatography, respectively. SDS–PAGE analysis revealed a purity level of greater than 95% for the final purified recombinant protein. A comparison with gel-filtration standards indicated the molecular weight of the recombinant BbBamA-POTRA P1–P3 sample (eluting at ∼11.4 ml) to be between those of ovalbumin (eluting at ∼10.5 ml; ∼43 000 Da) and ribonuclease A (eluting at ∼13.5 ml; ∼13 700 Da). Based on this information and on the theoretical molecular mass of the monomer (31.5 kDa), the purified BbBamA-POTRA P1–P3 was most likely to be a monomer in solution (Fig. 1 ▸ b).

Initially, poor crystals were obtained from the purified sample using conditions consisting of 1.26 M ammonium sulfate, 0.1 M MES sodium hydroxide pH 6.0 with 20 mg ml⁻¹ purified protein at 289 K (Fig. 2 ▸ a). After optimization, high diffraction-quality crystals of BbBamA-POTRA P1–P3 were obtained using 1.2 M ammonium sulfate, 0.1 M MES sodium hydroxide pH 5.5 and 15 mg ml⁻¹ BbBamA-POTRA P1–P3 using the hanging-drop method at 285 K (Fig. 2 ▸ b). A 15% SDS–PAGE analysis indicated that slight degradation of the protein occurred in the crystals, but the major band in the SDS–PAGE gel was broadly consistent (94% coverage) with the recombinant BbBamA-POTRA P1–P3 by fingerprint mass spectrometry (Fig. 2 ▸ e).

The best crystal diffracted to a resolution of 2.00 Å (Fig. 2 ▸ d) on the BL17U beamline at the SSRF and belonged to space group P2₁, with unit-cell parameters a = 45.353, b = 111.538, c = 64.376 Å, β = 99.913°. Analysis of the crystal contents with the Phenix package showed the presence of two molecules per asymmetric unit, corresponding to a Matthews coefficient of 2.92 ų Da⁻¹ and a solvent content of 57.9%. Comparing the primary sequence of BbBamA-POTRA P1–P3 with those of experimentally determined structures in the PDB indicated a low sequence homology (less than 21% sequence identity) between BbBamA-POTRA P1–P3 and NgBamA-POTRA. This result portended the challenge of determining phases using MR. Nevertheless, we built a model based on the NgBamA-POTRA P1–P3 conformation (PDB entry 4k3b) with the BbBamA-POTRA P1–P3 sequence using the Phyre2 server and used this predicted model and the NgBamA-POTRA P1–P3 structure as search models for MR phase determination. However, no useful results were observed. Therefore, we are now in the process of preparing additional crystallization experiments with selenomethionine-containing BbBamA-POTRA P1–P3 protein to provide phases for the X-ray intensities.

Funding Statement

This work was funded by National Natural Science Foundation of China grants 31622007 and 31670237. Taishan Scholar Foundation of Shandong Province grant .