- Open Access
Identification of a heterologous cellulase and its N-terminus that can guide recombinant proteins out of Escherichia coli
© Gao et al.; licensee BioMed Central. 2015
- Received: 4 February 2015
- Accepted: 23 March 2015
- Published: 10 April 2015
The Gram-negative bacterium Escherichia coli has been widely used as a cell factory for the production of proteins and specialty chemicals because it is the best characterized host with many available expression and regulation systems. However, recombinant proteins produced in Escherichia coli are generally intracellular and often found in the form of inclusion bodies. Extracellular production of proteins is advantageous compared with intracellular production because extracellular proteins can be purified more easily and can avoid protease attack, which results in higher product quality. In this study, we found a catalytic domain of a cellulase (Cel-CD) and its N-terminus can be employed as carriers for extracellular production of recombinant proteins.
In this report, we identified the catalytic domain of a cellulase (Cel-CD) from Bacillus sp. that can be secreted into the medium from recombinant E. coli BL21 (DE3) in large quantities without its native signal peptide. By subcellular location analysis, we proved that the secretion was a two-step process and the N-terminal sequence of the full length Cel-CD played a crucial function in secretion. Both the Cel-CD and its N-terminal sequence can serve as carriers for efficient extracellular production of select target proteins.
Fusion of heterologous proteins with N20 from Cel-CD can carry the target proteins out of the cells with a concentration from 101 to 691 mg/L in flask cultivation. The extracellular recombinant proteins with a relative high purity. The results suggested that this system has a potential application in plant biomass conversion and industrial production of enzymes and therapeutic proteins.
- Escherichia coli
- Extracellular protein production
- N-terminal sequence
The Gram-negative bacterium Escherichia coli has been widely used as a cell factory for the production of enzymes and therapeutic proteins, because it is the best characterized host with many available expression and regulation tools [1-3]. However, the common laboratory strains of E. coli are poor secretors of proteins under normal culture conditions, because this bacterium has a complex cell envelope with two layers [4-6]. Therefore, heterologous proteins produced in recombinant E. coli are generally intracellular and often in the form of inclusion bodies, from which the biologically active proteins can only be recovered by complicated and costly processes . Extracellular production of heterologous proteins in E. coli will not only provide a simple and convenient production and purification process, but also provide fast and direct screening capabilities for target therapeutic proteins or enzymes that are heterologously expressed in recombinant E. coli .
Significant effort to produce target proteins extracellularly in E. coli has been made, with the research efforts split into two categories. (1) Targeted accumulation of the heterologous protein in the periplasmic space through the inner membrane (IM) using a leader peptide, such as PelB, then the heterologous protein is released to the medium through the outer membrane (OM) using cell envelope mutants or lysis proteins [9-11]; (2) Fusion of the heterologous proteins to fusion partners that can be secreted from the cytosol out of the cells via known or unknown systems.
A number of signal peptides have been used for secretory production of recombinant proteins in both eukaryote and prokaryote [12,13]. The typical signal sequence most located in the amino terminal of proteins that functions as a targeting and recognition signal and contains a cleavage site which can be cleaved by a special signal peptidase after transportation . However, the recombinant proteins fused with signal sequence are transported to the periplasmic space instead of culture medium due to the double membrane structure of E. coli . To release the target proteins to the culture medium, cell envelope mutants or lysis proteins were employed. But it also suffers from the purity of the secreted proteins and cell sensitivity to the environment due to the leakage of the cell envelope.
Regarding the application of the second approach, several proteins, including heterologous proteins, were found to be secreted directly into the medium from recombinant E. coli. However, only a few of proteins were investigated as fusion partners for extracellular recombinant protein production, because the secretion efficiency and universality were not sufficient to scale to a larger expression volume [16-20]. In 2005, Majander et al., modified the flagellar type III secretion apparatus of E. coli. Then, by fusing a heterologous protein between the 173-bp untranslated region upstream of the gene fliC (encoding flagellin) and a transcriptional terminator from fliC, they found that the target proteins can be secreted into the medium via the type III secretion apparatus at levels of 1 to 15 mg/L . Subsequently, researchers identified a 13 kDa endogenous bacterial peptide in E. coli, YebF, which can be secreted into the medium in relatively large amounts by a two-step process via the periplasmic space . Recently, an osmotically inducible protein Y (OsmY) from E. coli BL21 (DE3) was identified using a proteomic method. OsmY can also be secreted out of the cells with target proteins at concentrations ranging between 5 and 64 mg/L . Furthermore, ESETEC®, WACKER’s patented secretion system, a technology for producing proteins and antibody fragments on a patented strain of E. coli K12, was the only one commercial solution to our knowledge .
Above all, it is generally considered that no fusion secretion systems are suitable for all heterologous proteins, because current fusion partner systems have been shown to have limitations [24-26]. In this study, we report that the catalytic domain of a cellulase (Cel-CD) from Bacillus sp.Z-16 can be efficiently secreted from E. coli when it was heterologously overexpressed. The accumulation of Cel-CD in the culture medium reached 514 mg/L. As the cellulase plays an important role in cellulosic biomass transformation, the extracellular expression of Cel-CD in E. coli provides a platform for cellulose production. Both the Cel-CD and its N-terminal sequence have potential as fusion partners in the production of various recombinant proteins.
Cel-CD is secreted from E. coli into the medium without typical signal peptides
A previously screened Bacillus sp. Z-16 was found to produce alkaline cellulase with endo-beta-glycosidase activity. This enzyme is composed of a 30-residue signal peptide at N-terminus, a catalytic domain (GH5, residues 30 to 404), followed by two carbohydrate-binding domains at C-terminal region (CBM_17_28, residues 405 to 821). The catalytic domain of this enzyme (Cel-CD) is 100% identical with a cellulase from Bacillus subtilis in the NCBI (DNA similarity is 99%, accession No.: M84963, 1,494 to 2,618). The full length Cel-CD consists of 375 amino acids with a molecular mass of 41.53 kDa.
Analysis of the N-terminal sequence of purified extracellular Cel-CD showed that it was MEGNTREDNF, which is the same as the predicted N-terminal sequence of the Cel-CD. The result proved that Cel-CD was translocated from E. coli BL21 (DE3) without removing any N-terminal sequence.
The secretion of Cel-CD is a two-step process via the periplasmic space
Cell lysis determination
A recent report showed that the heterologously expressed cutinase from Thermobifida fusca was secreted from cells because of its hydrolytic activity toward phospholipids . To exclude the possibility that Cel-CD secretion was also caused by its hydrolytic activity in the periplasmic space, the active site of Cel-CD was mutated according to the results of partial amino acid sequence alignment with cellulase. Two mutants, E160Q, and E160Q and E274Q, in which hydrolytic activity was completely lost, were still detected in the medium (Figure 3B). The morphology of E. coli BL21 (DE3) strains were analysised by Transmission electron microscopy (TEM). It showed that the cells expressing either Cel-CD or inactivated Cel-CD remained intact, supporting the aforementioned result (Figure 3C).
The N-terminus of Cel-CD plays a key role in secretion
Extracellular production of heterologous proteins in E. coli
Since the N-terminal region plays an important role in secretion, we tried the N-terminal 20 residues (N20) as a signal peptide to guide the target proteins out of E. coli (Figure 5B). The N-terminal 20 residues from Cel-CD were added to the target proteins by direct PCR ligation. Fusion of N20 with GlpQ and MBP led to a significant increase in secretion when compared with the Cel-CD fusion (Figure 5C). The extracellular protein was 691 and 350 mg/L, respectively, while the secretion of N20-CBD and N20-pelC reduced to 101 and 106.7 mg/L, respectively. However, N20-NRN1 was not secreted. Analysis of the expression of N20-NRN1 showed that it was deposited as inclusion bodies (data not shown).
In this study, we have shown that Cel-CD, the catalytic domain of a cellulase from Bacillus sp.Z-16, could be secreted into the culture medium without a typical signal sequence when expressed in E. coli BL21 (DE3). As a heterologous protein, the overexpression of Cel-CD was not found any interference to cell growth and metabolism. This may overcome the cell lysis caused by the overexpression of the fusion partner, thus extend the fermentation process to increase the extracellular production. Meanwhile, the secretion of Cel-CD and its recombinant proteins was a two-step pocess, the oxidizing environment and Dsb system of periplasm benefits the formation of disulfide bonds . Further experiments showed that Cel-CD can also be expressed and secreted from E. coli K strains, such as E. coli BW25113, MG1655, though the level of secretion was much lower. Thus, the application of Cel-CD fusion partner is not strain dependent.
We have proven that the N-terminal sequence of full-length Cel-CD plays a key role in secretion. This indicated that N-terminal 20 amino acid residues of Cel-CD can serve as a carrier for the secretion of heterologous target proteins out of E. coli. Fused with Cel-CD or N20, five selected proteins found to secrete into the medium at high levels. The secretion efficiency depends on the properties of the target proteins. Cel-CD appears to be suitable for the secretion of small and low-solubility proteins because of its high solubility and large size in E. coli, whereas the N20 peptide is more suitable for large proteins. Notably, all secreted fusion proteins exhibited relative high solubility; fusion proteins with low solubility are not efficiently secreted. We suggest that the solubility of the fusion protein is a prerequisite for secretion, while the N-terminal sequence and three dimensional structure of the protein is a crucial factor. Fusion of heterologous proteins with N20 from Cel-CD can carry the target proteins out of the cell. This is the first report to show that a short peptide can serve as a signal peptide and guide heterologous proteins across both the inner and outer membranes of E. coli.
Secretory production of proteins in E. coli has many possible applications. For example, a mutagenized protein library could be directly screened on a plate if the protein or its product is directly visible, which can greatly facilitate protein engineering research. Extracellular accumulation of Cel-CD with other cellulosic enzymes may also be useful for the creation of consolidated bioprocess for simultaneous biomass conversion product formation . Finally, secretory production of proteins will simplify the protein production process and will decrease production costs of enzymes and antibodies.
Recombinant proteins produced in Escherichia coli are generally intracellular and often found in the form of inclusion bodies. In this study, a heterologous expressed catalytic domain of a cellulase (Cel-CD) in E. coli was found to be secreted out of the cells in large quantities with endo-beta-glycosidase activity. We demonstrated that the N-terminal sequence of Cel-CD played a crucial role in this secretion, which suggests that both Cel-CD and its N-terminus can be employed as carriers for extracellular production of recombinant proteins. Recombinant E. coli expressing the fusion protein of Cel-CD and other types of hydrolytic enzymes can be used for creation of consolidated bioprocess.
Bacterial strains and plasmids construction
The strains and plasmids used in this study
E. coli DH5α
F− endA1 hsdR17(r K - m K +) supE44 thi-l λ − recA1 gyrA96 ΔlacU169 (Φ80dlacZM15)
E. coli BL21(DE3)
hsdS, gal(ΔcIts857 ind1, Sam7, nin5, lacUV5-T7, gene1
5.4 kb, f1 ori, T7 promoter, KanR
pET28a, cel-cd gene under T7 promoter
pET28a, 60 bp at 5′ region of cel-cd under T7 promoter
pET28a, cel-cd gene that truncated 15 bp at 5′ region
pET28a, cel-cd gene that truncated 30 bp at 5′ region
pET28a, cel-cd gene that truncated 45 bp at 5′ region
pET28a, cel-cd gene that truncated 60 bp at 5′ region
pET28a, cel-cd gene that truncated 90 bp at 5′ region
pET28a, cel-cd gene fusion with nrn1from Homo sapiens
pET28a, cel-cd gene fusion with male from Escherichia coli
pET28a, cel-cd gene fusion with cbd from Bacillus circulans
pET28a, cel-cd gene fusion with glpQ from Escherichia coli
pET28a, cel-cd gene fusion with pelc from Bacillus subtilis
pET28a, n20 gene fusion with nrn1
pET28a, n20 gene fusion with male
pET28a, n20 gene fusion with cbd
pET28a, n20 gene fusion with glpq
pET28a, n20 gene fusion with pelc
pET28a, the mutated cel-cd gene under T7 promoter
pET28a, the mutated cel-cd gene under T7 promoter
The primers used in this study
The DNA sequence encoding Cel-CD with 5 amino acid residues deletion at the N-terminus was amplified from the plasmids pCel by polymerase chain reaction using PrimeSTAR (TaKaRa) with the Cel-CD5-NcoI-Forward and Cel-CD-XhoI-Reverse as primers, generating plasmids pΔ5cel. The same method was used to construct pΔ10cel, pΔ15cel, pΔ20cel and pΔ30cel.
To fuse the target protein genes to the 3′ end of cel-cd, the DNA fragments were amplified by PCR, and were digested with restriction enzyme BamHI and XhoI on the both ends. The resulting fragments were then ligated to the vector pET-28a/cel by T4 DNA ligase to obtain the objective plasmids. The same method was used to construct N20 fusion proteins.
Cel-CD activity assay
The hydrolytic activity of the secreted Cel-CD was determined by Dinitrosalicylic Acid (DNS) method. 0.1 mL culture medium was added to 2 mL Britton-Robinson buffer containing 1% CMC-Na in a test tube. Reaction mixture was incubated at 37°C for 30 min, then was added with 2 mL DNS reagent and boiled for 5 min before it was diluted to 12.5 mL. The enzyme activity was measured with a spectrophotometer at 540 nm.
The total secretory proteins were quantified by Bradford method using Easy Protein Quantitative kit (Lot #G10322, TransGen Biotech). The proteins strained with Coomassie were detected by the UV spectrophotometer at 595 nm. The protein samples were loaded on 12% or 15% SDS-PAGE and stained with Coomassie dye. The stained gels were scanned by Scanner and the target protein bands were normalized and quantified by Bio-Rad Quantity One Version 4.0.
SDS-PAGE and western blotting
Cells were harvested at indicated time after induction and washed once with 1 mL of 100 mM MOPS (pH 7.0). The secreted protein samples were obtained directly by centrifugation. The cell pellets were then disrupted by ultrasonification and centrifuged to obtain the total soluble protein samples. The samples were then prepared by adding an equal volume of electrophoresis loading buffer. Proteins were separated on 12% SDS-PAGE and either stained with Coomassie dye or transferred to a PVDF membrane (100 V, 1 h) for immunoblotting, blocked with 5% (W/V) BSA in PBS/Tween, and probed using the appropriate antibody. Antibody to Cel-CD (Sigma, 1:2,000) was used for the detection of secreted Cel-CD in the culture medium and cells. Antibody to MalE (Sigma, 1:5,000) and GroEL (Sigma, 1:10,000) were used for the control of potential lysis or leakage from the periplasmic space (MalE) and cytoplasmic (GroEL).
Site-directed mutagenesis of Cel-CD
Site-directed mutagenesis mutants were constructed using Easy Mutagenesis System Kit (TransGene). The primers Glu160Gln-Forward/Reverse were included 5′end overlap region, mutation site and 3′ end extension region. In a certain system, the target DNA vectors containing the cel-cd gene were amplified by PCR and the templates were digested by DMT enzyme. After purified by gel extraction, the target vectors containing the E160Q mutant and E160Q&E274Q mutant was obtained. The Gln in the active sites of Cel-CD (Gln160 and Gln274) were all mutated to Glu. All the resulting constructs were confirmed by DNA sequencing and hydrolytic activity determination.
Transmission electron microscopy (TEM)
E. coli cells containing the target plasmids were collected by centrifuging at 4,000 g for 10 min. Pelleted cells were fixed by 2.5% glutaraldehyde and 1% osmic acid. After gradually dehydrated by ethyl alcohol, the cells were embedded in Epon-812 araldite. The samples were cut into 70 nm thick slice, stained using 2% uranyl acetate and lead citrate, and observed by the transmission electron microscope at 200 kV. The resulting microscopic images were photographed.
Analysis of the N-terminal sequence
The protein samples were purified and separated by 2D-SDS PAGE, and were transferred to polyvinylidene difluoride (PVDF) blotting membrane. The N-terminal sequence of the protein was analyzed by EDMAN degradation method  using the PROCISE491 protein sequencer by the Center for Protein Science, Peking University.
Congo red staining
To determine the hydrolytic activity of Cel-CD, an agarose plate-based assay was used. 1% (wt/vol) Carboxymethylcellulose (CMC-Na) was added in the agarose plate as a substrate (0.8%, wt/vol). Samples containing the Cel-CD were loaded into the hole of plate. The plate was incubated at 30°C for 3 h, then flooded with 0.1% (wt/vol) Congo red solution and stained for 15 min at room temperature. The dye was removed, and 5 mL of water was used to wash the plate. Finally, 5 mL 0.9% NaCl solution was applied for 15 min, and the plates were then dried and photographed.
Periplasmic protein was separated from cytoplasm by the osmotic shock method . The cells were harvested by centrifugation (6,000 g for 5 min) and were suspended in 100 mM Tris–HCl containing 20% sucrose and 1 mM EDTA (pH 8.0), then pelleted by centrifugation (8,000 g for 5 min) followed by re-suspension in ice-cold water for 10 min. After the addition of MgCl2 to a final concentration of 1 mM, the cell suspension was incubated on ice for a further 10 min before being pelleted by centrifugation (8,000 g for 5 min).
This work was financially supported by research grants from the National Basic Research Program of China (2012CB725202) and the National Natural Science Foundation of China (31370085).
The provision of strain Bacillus sp. Z-16 by Prof. Liu, Xiangmei is highly appreciated. Critical comments of Prof. Xun, Luying on this manuscript are also appreciated.
- Jong WS, Sauri A, Luirink J. Extracellular production of recombinant proteins using bacterial autotransporters. Curr Opin Biotechnol. 2010;21:646–52.View ArticleGoogle Scholar
- Ferrer-Miralles N, Domingo-Espin J, Corchero JL, Vazquez E, Villaverde A. Microbial factories for recombinant pharmaceuticals. Microb Cell Fact. 2009;8:17.View ArticleGoogle Scholar
- Langer ES. Capitalizing on novel expression systems. Genet Eng Biotechnol News. 2009;29:12–2.Google Scholar
- Desvaux M, Hebraud M, Talon R, Henderson IR. Secretion and subcellular localizations of bacterial proteins: a semantic awareness issue. Trends Microbiol. 2009;17:139–45.View ArticleGoogle Scholar
- Wai SN, Lindmark B, Soderblom T, Takade A, Westermark M, Oscarsson J, et al. Vesicle-mediated export and assembly of pore-forming oligomers of the enterobacterial ClyA cytotoxin. Cell. 2003;115:25–35.View ArticleGoogle Scholar
- Pugsley AP, Francetic O. Protein secretion in Escherichia coli K-12: dead or alive? Cell Mol Life Sci. 1998;54:347–52.View ArticleGoogle Scholar
- Baneyx F, Mujacic M. Recombinant protein folding and misfolding in Escherichia coli. Nat Biotechnol. 2004;22:1399–408.View ArticleGoogle Scholar
- Choi JH, Lee SY. Secretory and extracellular production of recombinant proteins using Escherichia coli. Appl Microbiol Biotechnol. 2004;64:625–35.View ArticleGoogle Scholar
- Shin HD, Chen RR. Extracellular recombinant protein production from an Escherichia coli lpp deletion mutant. Biotechnol Bioeng. 2008;101:1288–96.View ArticleGoogle Scholar
- Robbens J, Raeymaekers A, Steidler L, Fiers W, Remaut E. Production of soluble and active recombinant murine interleukin-2 in Escherichia coli: high level expression, Kil-induced release, and purification. Protein Expr Purif. 1995;6:481–6.View ArticleGoogle Scholar
- Sevastsyanovich YR, Leyton DL, Wells TJ, Wardius CA, Tveen-Jensen K, Morris FC, et al. A generalised module for the selective extracellular accumulation of recombinant proteins. Microb Cell Fact. 2012;11:69.View ArticleGoogle Scholar
- von Heijne G. Signal sequences. The limits of variation. J Mol Biol. 1985;184:99–105.View ArticleGoogle Scholar
- Austen BM, Ridd DH. The signal peptide and its role in membrane penetration. Biochem Soc Symp. 1981;46:235–58.Google Scholar
- von Heijne G. The signal peptide. J Membr Biol. 1990;115:195–201.View ArticleGoogle Scholar
- Mergulhao FJ, Summers DK, Monteiro GA. Recombinant protein secretion in Escherichia coli. Biotechnol Adv. 2005;23:177–202.View ArticleGoogle Scholar
- Gumpert J, Hoischen C. Use of cell wall-less bacteria (L-forms) for efficient expression and secretion of heterologous gene products. Curr Opin Biotechnol. 1998;9:506–9.View ArticleGoogle Scholar
- Jeong KJ, Lee SY. Excretion of human beta-endorphin into culture medium by using outer membrane protein F as a fusion partner in recombinant Escherichia coli. Appl Environ Microbiol. 2002;68:4979–85.View ArticleGoogle Scholar
- Teng C, Jia H, Yan Q, Zhou P, Jiang Z. High-level expression of extracellular secretion of a beta-xylosidase gene from Paecilomyces thermophila in Escherichia coli. Bioresour Technol. 2010;102:1822–30.View ArticleGoogle Scholar
- Nandakumar MP, Cheung A, Marten MR. Proteomic analysis of extracellular proteins from Escherichia coli W3110. J Proteome Res. 2006;5:1155–61.View ArticleGoogle Scholar
- Qian ZG, Xia XX, Choi JH, Lee SY. Proteome-based identification of fusion partner for high-level extracellular production of recombinant proteins in Escherichia coli. Biotechnol Bioeng. 2008;101:587–601.View ArticleGoogle Scholar
- Majander K, Anton L, Antikainen J, Lang H, Brummer M, Korhonen TK, et al. Extracellular secretion of polypeptides using a modified Escherichia coli flagellar secretion apparatus. Nat Biotechnol. 2005;23:475–81.View ArticleGoogle Scholar
- Zhang G, Brokx S, Weiner JH. Extracellular accumulation of recombinant proteins fused to the carrier protein YebF in Escherichia coli. Nat Biotechnol. 2006;24:100–4.View ArticleGoogle Scholar
- Muecke M, Leonhartsberger S. Increasing Production Process Efficiency. Genet Eng Biotechnol News. 2009;29:56–7.Google Scholar
- Xia XX, Han MJ, Lee SY, Yoo JS. Comparison of the extracellular proteomes of Escherichia coli B and K-12 strains during high cell density cultivation. Proteomics. 2008;8:2089–103.View ArticleGoogle Scholar
- Zhou S, Yomano LP, Saleh AZ, Davis FC, Aldrich HC, Ingram LO. Enhancement of expression and apparent secretion of Erwinia chrysanthemi endoglucanase (encoded by celZ) in Escherichia coli B. Appl Environ Microbiol. 1999;65:2439–45.Google Scholar
- Papi RM, Chaitidou SA, Trikka FA, Kyriakidis DA. Encapsulated Escherichia coli in alginate beads capable of secreting a heterologous pectin lyase. Microb Cell Fact. 2005;4:35.View ArticleGoogle Scholar
- Weber MJ, DeMoss JA. The inhibition by chloramphenicol of nascent protein formation in E. coli. Proc Natl Acad Sci U S A. 1966;55:1224–30.View ArticleGoogle Scholar
- Shuman HA, Panagiotidis CH. Tinkering with transporters: periplasmic binding protein-dependent maltose transport in E. coli. J Bioenerg Biomembr. 1993;25:613–20.Google Scholar
- Horwich AL, Low KB, Fenton WA, Hirshfield IN, Furtak K. Folding in vivo of bacterial cytoplasmic proteins: role of GroEL. Cell. 1993;74:909–17.View ArticleGoogle Scholar
- Su L, Woodard RW, Chen J, Wu J. Extracellular location of Thermobifida fusca cutinase expressed in Escherichia coli BL21(DE3) without mediation of a signal peptide. Appl Environ Microbiol. 2013;79:4192–8.View ArticleGoogle Scholar
- de Marco A. Recent contributions in the field of the recombinant expression of disulfide bonded proteins in bacteria. Microb Cell Fact. 2012;11:129.View ArticleGoogle Scholar
- Bokinsky G, Peralta-Yahya PP, George A, Holmes BM, Steen EJ, Dietrich J, et al. Synthesis of three advanced biofuels from ionic liquid-pretreated switchgrass using engineered Escherichia coli. Proc Natl Acad Sci U S A. 2011;108:19949–54.View ArticleGoogle Scholar
- Edman P. A method for the determination of amino acid sequence in peptides. Arch Biochem. 1949;22:475.Google Scholar
- Nossal NG, Heppel LA. The release of enzymes by osmotic shock from Escherichia coli in exponential phase. J Biol Chem. 1966;241:3055–62.Google Scholar
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.