Biochimica et Biophysica Acta 1854 (2015) 10–19

Contents lists available at ScienceDirect

Biochimica et Biophysica Acta journal homepage: www.elsevier.com/locate/bbapap

Clostridium thermocellum thermostable lichenase with circular permutations and modifications in the N-terminal region retains its activity and thermostability А.А. Tyurin a,b, N.S. Sadovskaya a, Kh.R. Nikiforova a, О.N. Mustafaev a, R.A. Komakhin c, V.S. Fadeev a, I.V. Goldenkova-Pavlova a,b,⁎ a b c

Institute of Plant Physiology, Russian Academy of Sciences, ul. Botanicheskaya 35, Moscow 127276, Russia Department of Genetics and Biotechnology, Russian State Agrarian University—Moscow Timiryazev Agricultural Academy, ul. Timiryazevskaya 49, Moscow 127550, Russia All-Russia Research Institute of Agricultural Biotechnology, Russian Academy of Agricultural Sciences, ul. Timiryazevskaya 42, Moscow 127550, Russia

a r t i c l e

i n f o

Article history: Received 26 June 2014 Received in revised form 25 September 2014 Accepted 15 October 2014 Available online 23 October 2014 Keywords: Lichenase Clostridium thermocellum Circular permutation Carrier protein Thermostability Activity

a b s t r a c t The Clostridium thermocellum lichenase (endo-β-1,3;1,4-glucan-D-glycosyl hydrolase) displays a high thermostability and specific activity and has a compact protein molecule, which makes it attractive, in particular, for protein engineering. We have utilized in silico analysis to construct circularly permuted (CP) variants and estimated the retained activity and thermostability. New open termini in the region of residues 53 or 99 in two lichenase CP variants (CN-53 and CN-99) had no effect on their activity and thermal tolerance versus another variant CP variant, CN-140 (cut in the region of residue 140), which displayed a dramatic decrease in the activity and thermostability. Construction and further activity and thermostability testing of the modified lichenase variants (M variants) and CP variants with peptides integrated via insertion fusion have demonstrated that the Nterminal regions in the lichenase catalytic domain (53 and 99 amino acid residues) that permit circular permutations with retention of activity and thermostability of the enzyme as well as the region between the C and N termini of the native lichenase in thermostable and active lichenase variants (CN-53 and CN-99) may be used for integrating small peptides without the loss of activity and thermostability. These findings not only suggest that CP predictions can be used in search for internal integration sites within protein molecule, but also form the background for further enzymatic engineering of the C. thermocellum thermostable lichenase aiming to create new fusion proteins. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Protein engineering is targeted to the creation of the molecules with necessary specified properties. Thermostable proteins are of special interest for protein engineering. The efforts here are directed to both the search for new thermostable proteins and modification of the already known ones. There are various motivations in the engineering of thermostable proteins, including study of the functional role of protein structural elements, determination of the folding, decrease in the proteolytic susceptibility, improvement of catalytic activity and/or thermostability, and search for internal sites allowing for integration of other proteins [1–5]. The following approaches are typically used in protein engineering: deletions (excision of domains and/or individual amino acids); random and directed mutagenesis; grouping of two or more enzyme catalytic modules; and circular permutation (CP) method [4–9]. Note that the ⁎ Corresponding author at: Institute of Plant Physiology, Russian Academy of Sciences, ul. Botanicheskaya 35, Moscow,127276, Russia. Tel.: +7 499 231 8315. E-mail address: [email protected] (I.V. Goldenkova-Pavlova).

http://dx.doi.org/10.1016/j.bbapap.2014.10.012 1570-9639/© 2014 Elsevier B.V. All rights reserved.

CP method is a powerful molecular tool for altering protein sequences and, as a consequence, their structure [9,10]. In particular, “cutting” a protein polypeptide chain within a functionally significant element may change the conformation of this protein, thereby leading to loss of its function; however, this “cutting” beyond the functional element or on its boundary frequently causes only insignificant changes in the protein function [11,12]. In other words, the properties of a “cut” protein will vary depending on the location of this breakpoint. Based on this strategy, protein molecules are “cut” to construct and characterize circularly permutated (CP) proteins, detecting the protein functional elements important for their properties via a comparative analysis [9]. An end-to-end fusion is frequently used when engineering bifunctional proteins to connect two enzymes [8,13–16] using different approaches, for example, by optimization of peptide linkers [14] or through all-atom molecular dynamics simulations [16]. However, the chimeric enzymes constructed with the help of this method are frequently more susceptible to proteolytic degradation and structural instability [17]. In such cases, the strategy for constructing chimeric enzymes using the “insertion” fusion when one gene is inserted within the other gene may be more beneficial [12,18,19]. The potential regions

АА Tyurin et al. / Biochimica et Biophysica Acta 1854 (2015) 10–19

permitting integration are of a paramount importance for successful construction of such fusion proteins. We have assumed that the regions permitting circular permutations in a protein molecule, that is, the regions where new N and C termini of the protein are formed without any significant loss in the major protein properties, may be potentially used also as the regions for insertion fusions with protein sequences. Glycosyl hydrolases are used as models when studying the folding of protein molecules and the mechanisms rendering them thermostable [20,21]. In addition, a practical application has been found for glycosyl hydrolases [13,22], in particular, for constructing bifunctional proteins [14,19]. In this work, the Clostridium thermocellum lichenase (endo-β1,3;1,4-glucan-D-glycosyl hydrolase, EC 3.2.1.73, P29716) was used as a model. Similar to the majority of glycosyl hydrolases, lichenase is a single subunit protein comprising (i) a signaling peptide (positions 1–27); (ii) a catalytic module, homologous to the full-sized Bacillus lichenase (positions 28–251); (iii) a Thr–Pro box (positions 252–261); and (iv) a cellulosome-binding (dockerin) domain (positions 262–334) [22]. It has been earlier demonstrated that the lichenase deletion variant carrying the catalytic domain and part of Thr–Pro box as well as the lichenase catalytic domain alone retains a high specific activity and thermostability [23]. A compact structure of the lichenase molecule, its high thermostability, and specific activity are attractive, in particular, for constructing hybrid proteins. However, it is necessary to clarify which particular regions in the lichenase catalytic domain permit internal integration without a dramatic loss in its activity and thermostability. Note that the precise information about the role of structural elements (modules) constituting the lichenase catalytic domain in manifestation of its major enzymatic properties is currently rather limited [7,24]. In this work, we have utilized in silico analysis of the lichenase sequences, in particular, by predicting potential circular permutation sites, to construct three CP lichenase variants and experimentally determine their properties (activity and thermostability). New open ends in the N-terminal region (amino acid residues 53 or 99), unlike the open ends in the C-terminal region (residue 140), had no effect on the activity or thermal tolerance of the lichenase CP variants. We have constructed and tested activity and thermostability of the modified lichenase variants (M variants) with peptide insertion fusions and found out that the CP regions are also suitable for integrating peptides. It has been clearly demonstrated that the N-terminal regions of lichenase catalytic domain that permit CP with retention of the enzyme activity and thermostability also permit internal integration of small peptides, including multiple integration. Moreover, it has been shown that the lichenase CP variants that retained their activity and thermostability also permit insertion fusions with small peptides as internal modules between the C and N termini of the native lichenase without any dramatic decrease in their thermostability and activity. As a result of this study, potential sites that can be used to integrate other proteins as internal modules have been detected in the lichenase and its CP variants, namely, three variants of the catalytic domains and two variants of the CP variants of this enzyme carrying the integration sites and retaining high thermostability and activity are proposed, which have a potential for further designing of bifunctional proteins. 2. Materials and methods 2.1. In silico analysis Homologs of thermostable lichenase were identified using the Blast software [25]. BLAST was performed to search the potential templates available in the PDB database [26]. The homology model of lichenase was built with Phyre [27], allowing for 3D structure prediction for the proteins with a known PDB structure used as a template pattern. Of all the proteins that Phyre suggested as a pattern, we selected the 3I4I protein as displaying the maximal homology to LicBM3 and obtained the

11

3D structure for lichenase. The program Muscle [28] was used to construct pairwise alignments of amino acid sequences. The circularly permutated lichenase homologs were searched for using iSARST web server [29] and the Blast software in PDB [26]. Potential CP sites in proteins were predicted using the CPred web server [30]. 2.2. Recombinant DNA techniques Standard molecular cloning procedures were used in the work as well as Evrogen (Russia) primers and Promega (United States), Fermentas (Lithuania), QIAexpress (United States), and Novagene (United States) enzymes and reagent kits. Table S1 lists the primers used in cloning procedures. The sequence of licBM3 deletion mutant was determined by PCR with the plasmid pQE-licBM2-KM2-Mys25 [31] as a template and the primers L-Forw and L-Rev. The BamHI–PstI PCR fragment was cloned into the plasmid pQE30 (Qiagen, United States) hydrolyzed with the BamHI and PstI restriction endonucleases to get the plasmid pQE-licBM3. 2.2.1. Constructing the hybrid genes encoding circularly permutated (CP) lichenase variants Initially, licBM3 gene fragments, designated C53, C99, and C140, were produced by PCR using pQE-licBM3 as a template and the primers C53-For/C-Rev, C99-For/C-Rev, and C140-For/C-Rev, respectively, as well as licBM3 gene fragments, designated as N53, N99, and N140, using the primers N-For/N53-Rev, N-For/N99-Rev, and N1-For/N140Rev, respectively. Synthesized PCR fragments C53, C99, and C140 were hydrolyzed with BamHI restriction endonuclease and PCR fragments N53, N99, and N140, with PstI. Then the fragment pairs BamHI-C53/ PstI-N53, BamHI-C99/PstI-N99, and BamHI-C140/PstI-N140 were ligated to clone each pair into the pQE30 hydrolyzed with BamHI and PstI. The resulting expression vectors were designated pQE-CN-53, pQE-CN-99, and pQE-CN-140. 2.2.2. Constructing the hybrid genes encoding the lichenase catalytic module with integrated small peptides The plasmid pQE-NC-L-53 was constructed in several stages. The licBM3 gene fragment designed N-53-1 was obtained by PCR with pQE-licBM3 as a template and F1/R1-53 primer pair. At the next stage, the PCR fragment designated N-53 was produced using N-53-1 PCR fragment as a template and F1/R2-53 primer pair and the pQE-licBM3 gene fragment designated C-53 using pQE-licBM3 as a template and F53/R1 primer pair. N-53 and C-53 PCR fragments were hydrolyzed with ApaI restriction endonuclease and ligated. The ligation product was further used as a template for PCR with F1 and R1 primers. The PCR fragment designated NC-L-53 was hydrolyzed with SphI and PstI and cloned into the plasmid pQE30 hydrolyzed with SphI and PstI. The plasmid pQE-NC-L-53-99 was also constructed in several stages. The licBM3 gene fragment designated N-99-1 was produced by PCR using pQE-NC-L-53 as a template and F1/R1-99 primer pair. Then, the PCR fragment designated N-99 was obtained with N-99-1 PCR fragment as a template and F1/R2-99 primer pair and the licBM3 gene fragment designated C-99 was obtained using pQE-licBM3 as a template and F99/R1 primer pair. N-99 and C-99 PCR fragments were hydrolyzed with SmaI and ligated. The ligation product was used as a template for PCR with F1 and R1 primers. The resulting PCR fragment, designated NC-L-53-99, was hydrolyzed with SphI and PstI and cloned into the pQE30 hydrolyzed with the same restriction endonucleases. The plasmid pQE-NC-L-99 was constructed by cloning ApaI–PstI fragment of the plasmid pQE-NC-L-53 into the pQE-NC-L-53-99 hydrolyzed with ApaI and PstI restriction endonucleases. The plasmid pQE-NC-L-140 was constructed in several stages. The licBM3 fragment designated N-140-1 was produced by PCR using pQE-licBM3 as a template and F1/R1-140 primer pair. Then the PCR fragment designated N-140 was synthesized with N-140-1 PCR fragment as a template and F1/R2-140 primer pair and the licBM3 gene

12

АА Tyurin et al. / Biochimica et Biophysica Acta 1854 (2015) 10–19

fragment designated C-140 was produced with pQE-licBM3 as a template and F-140/R1 primer pair. N-140 and C-140 PCR fragments were hydrolyzed with SmaI and ligated. The ligation product was used as a template for PCR with F1 and R1 primers. The resulting PCR fragment, designated NC-L-140, was hydrolyzed with SphI and PstI restriction endonucleases and cloned into the pQE30 plasmid hydrolyzed with the same restriction endonucleases. 2.2.3. Constructing the hybrid genes encoding the circularly permutated (CP) lichenase variants with integrated small peptides Initially, the licBM3 gene fragments designated С53-L1 and С99-L1 were produced by PCR using the C53 and C99 PCR fragments as a template and the primers C53 forw/R-CP and C99 forw/R-CP, respectively, as well as the licBM3 fragments designated N53-L1 and N99-L1 using the primers F1-CP/N53 rev and F1-CP/N99 rev, respectively, were obtained. Then the PCR fragments designated N53-L2 and N99L2 using the primers F2-CP/N53 rev and F2-CP/N99 rev, respectively. The resulting PCR fragments С53-L1 and N53-L2 as well as С99-L1 and N99-L2 were hydrolyzed with SmaI restriction endonuclease and ligated in a pairwise manner. In the next PCR, the ligation products, designated CN-L-53-1 and CN-L-99-1, were used in PCR as a template with the primers C53 forw/N53 rev and C99 forw/N99 rev, respectively, to give the fragments designated CN-L-53 and CN-L-99. The last fragments were hydrolyzed with the BamHI and PstI restriction endonucleases and cloned in the pQE30 vector preliminary hydrolyzed with BamHI and PstI. The final expression vectors were designated pQE-CN-L-53 and pQE-CN-L-99. 2.3. Expression and purification of proteins Escherichia coli XL1-Blue (Stratagene, United States) harboring pQElicBM3, pQE-CN-53, pQE-CN-99, pQE-CN-140, pQE-NC-L-53, pQE-NC-L99, pQE-NC-L-140, pQE-NC-L-53-99, pQE-CN-L-53 and pQE-CN-L-99 was grown to OD600 = 0.6 at 37 °C in LB medium and was then induced with 0.5 mM isopropyl β-D-1-thiogalactoside (IPTG) at 37 °C for 16 h. Cells were harvested, washed twice, and suspended in 50 mM Tris– HCl, pH 8.0. The cells were lysed using a French pressure cell press (Aminco, SLM Instruments Inc., American Instrument Company) and cleared by centrifugation (crude protein lysate). The supernatant was incubated at 70 °C for 1 h and cleared by centrifugation (heated cell lysate). In addition, the crude protein lysate was loaded on a Ni2+-NTAagarose column (Qiagen, Hilden, Germany) equilibrated with 50 mM Tris–HCl (pH 8.0) and eluted with the same buffer containing a linear imidazole gradient (20–200 mM). The eluted enzyme was gradually dialyzed against 50 mM Tris–HCl (pH 8.0). SDS-PAGE was conducted in 12% acrylamide gel stained with Coomassie brilliant blue R-250 (Sigma Aldrich, St. Louis, MO, United States) [32]. Molecular weights were estimated using a Prestained Protein Ladder (Fermentas, Lithuania). Protein concentration was determined according to Bradford with bovine serum albumin as a standard [33]. 2.4. Enzyme activity assays Plate assay was performed as described previously [24]. Zymograms were obtained by staining the gel after protein separation by SDS-PAGE (10%) in the presence of 0.1% lichenan as described earlier [24]. After electrophoresis, the gels were incubated at 70 °C for 1 h. Enzyme activities were determined by staining with 0.5% Congo red solution followed by washing in 1 M NaCl. The lichenase activity was determined using lichenan (Megazyme, Ireland) as a substrate. Reducing sugars were determined using the 3,5-dinitrosalicylic acid (DNS) method and colorimetrically monitored at 540 nm [34] in a CLARIOstar multimode microplate reader (BMG LABTECH, Germany). One activity unit was defined as the quantity of enzyme necessary to release reducing sugars per 1 min under standard conditions. The standard assay was conducted for 10 min in 50 mM Tris–HCl (pH 8.0) at 70 °C. Thermostability of

the enzymes incubated in 50 mM Tris–HCl (pH 8.0) at 70 °C for 60, 120, and 240 min was estimated according to their activity determined by a standard assay. 2.5. Statistical analysis All experiments were performed independently at least three times, and the results were expressed as mean ± standard deviation. 3. Results and discussion 3.1. In silico analysis Initially, we analyzed the lichenase catalytic domain, designated LicBM3. The modeling was initiated by submitting the protein sequence of the LicBM3, and BLAST was performed to search for the potential template available in PDB. PDB was searched for LicBM3 homologs [26] selecting the sequences that met the following criteria: the sequences displaying the maximal homology to LicBM3 and the sequences partitioned into two regions each homologous to LicBM3 but with inverted positions (in other words, circularly permuted; hereinafter, CP) relative to the LicBM3 amino acid sequence. This means that the region located close to the LicBM3 C-terminus corresponds to the region located closer to the N-terminus and vice versa, the region located closer to the LicBM3 N terminus corresponds to the region located closer to its C terminus. The circularly permuted lichenase homologs were selected using the program iSARST [29] (option CPSARST [35], Circular Permutation Search Aided by Ramachandran Sequential Transformation) and 3I4I as a template, displaying the maximal homology to LicBM3 and manually tested using the Blast web server. The two templates were selected from the search that met the first condition (PDB ID: 3I4I and 1BYH) and five proteins that met the second one (PDB ID: 1CPN, 1CPM, 1AXK, 1AJK, and 1AJO) (Table 1). When aligning LicBM3 with each of the inverted proteins (1CPN, 1CPM, 1AXK, 1AJK, and 1AJO), the homologous regions were manually moved so that they would correspond to the homologous LicBM3 region. The positions of α-helices and β-sheets were specified according to PDB. Then, the secondary structures of the proteins from the final sample aligned in a pairwise manner were mapped into LicBM3 amino acid sequence. We assumed that the protein structure should be retained in a group of closely related proteins. Correspondingly, the regions forming a globular structure in a pair of aligned proteins should with a high accuracy map into one another (α-helices into α-helices and β-sheets into β-sheets). It should be emphasized that when mapping the secondary structures of the proteins forming the final sample into alignment, we allowed no more than one insertion/deletion within α-helix and β-sheet due to the conservation of globular structure. Otherwise, we considered a larger part as an α-helix or a β-sheet and the remaining smaller part was regarded as belonging to a loop. The next stage of in silico analysis was conducted with the help of the CPred server, providing prediction of potential CP sites in the proteins with a known PDB structure [30]. For this purpose, we selected the 3I4I protein, displaying the maximal linear homology to LicBM3 (Table 1). The probability of a successful CP was estimated for each amino acid residue of this protein. Then the positions of 3I4I protein favorable for CP were mapped using pairwise alignments in the LicBM3 amino acid sequence (Fig. 1). In this procedure, we selected the potential CPred-predicted CP regions with a length of no less than seven amino acid residues. Note that the location of these regions within an α-helix was allowed but it should not overlap a β-sheet more than by one-third of its length. The LicBM3 secondary structure maps constructed using the proteins of the final sample were compared (Fig. 1). As is evident from this figure, most α-helixes retain their positions: their boundaries differ by one amino acid residue only in one case and strictly coincide in all the

АА Tyurin et al. / Biochimica et Biophysica Acta 1854 (2015) 10–19

13

Table 1 Proteins homologous to LicBM3. No.

PDB

Protein name

Source

Query cover, %

ID, %

Citation

0 0 1 1 1 2 3

3I4I 1BYH 1CPM 1CPN 1AXK 1AJK 1AJO

1,3-1,4-beta-glucanase 1,3-1,4-beta-glucanase Circularly permuted jellyroll protein Circularly permuted jellyroll protein Circularly permuted 1,3-1,4-beta-glucanase Circularly permuted 1,3-1,4-beta-glucanase Circularly permuted 1,3-1,4-beta-glucanase

Prokaryote, mouse hindgut metagenome Bacillus Paenibacillus macerans Paenibacillus macerans Bacillus subtilis Bacillus macerans Paenibacillus macerans

96 95 90 89 88 95 95

60 59 66 66 66 69 63

[36] [37] [11] [11] [38] [12] [12]

Notes: 0, the proteins corresponding to LicBM3; 1–3, proteins corresponding to the lichenase variants permuted at positions 53, 99, and 140, respectively; ID, identity; and PDB, the protein accession numbers in Protein Data Bank.

remaining cases. All the positions of β-sheets are also preserved with the maximally observed differences between the boundaries of three amino acid residues. Note that the proteins of the final sample permuted relative to LicBM3 fell into three groups according to the position of the inversion breakpoint. Later, this fact allowed us to select the potential regions for CP in the LicBM3 amino acid sequence. The first group included the 1CPN, 1CPM, and 1AXK proteins, displaying a high homology to each other (Table 1). Comparison of LicBM3 to 1CPN and 1CPM demonstrated that the inversion breakpoint is at position 58 (between amino acid residues 58 and 59) beyond any β-sheets (the nearest β-sheets are at positions 45–49 and 60–66). Comparison of LicBM3 to 1AXK showed that one of the two regions in the 1AXK sequence ended at LicBM3 position 49 and the other region started from position 59. Correspondingly, the inversion breakpoint also falls within the region between these β-sheets (positions 45–49 and 60–66, respectively). Note that positions 50–58 were predicted by CPred as potential CP sites. Based on this protein sample, the LicBM3 amino acid residue at position 53 was selected for constructing one of the CP lichenase variants. The second group contained one protein, 1AJK. Analysis of LicBM3 and 1AJK demonstrated that the inversion breakpoint fell between two β-sheets: one of the 1AJK regions ended to position 83 in the LicBM3 sequence and the other started immediately after it, from position 84. The positions of the neighboring β-sheets at both sides of the breakpoint are 73–80 and 87–95, respectively. Note that Asn84 is the only amino acid predicted by CPred as a potential CP site. This attracted our attention to the region between the next two β-sheets (positions 87–95 and 104–111). Since CPred characterized positions 97–103, falling in between these β-sheets, as favorable for CP, we chose LicBM3 amino acid residue 99 for constructing the second CP lichenase variant. It is necessary to emphasize that the β-sheet located at positions 104–111 is highly conserved and houses the protein reaction center (Fig. 1). The third region also contains only one protein, 1AJO. Comparison of LicBM3 and 1AJO sequences demonstrated that the inversion breakpoint in 1AJO fell between two neighboring β-sheets: the first region ended at LicBM3 amino acid residue 126 and the second one started from position 127. Note that CPred characterized region 125–145 as favorable for potential CP (Fig. 1). The coordinates of the β-sheets located at both sides are 117–123 and 132–135, respectively. This suggested us to focus on the region between the next two βsheets (positions 132–135 and 145–152). As is mentioned above, this region was predicted by CPred as favorable for CP. Finally, we selected the LicBM3 residue at position 140 for constructing the third lichenase CP variant. It is necessary to emphasize that residue 140 falls between two β-sheets in the models constructed based on non-CP proteins 3I4I and 1BYH, displaying a high homology to LicBM3; in the case of 1CPN, 1CPM, 1AXK, 1AJK, and 1AJO proteins, with a lesser homology to LicBM3 and circularly permuted relative to its sequence, it coincides with the first amino acid residue of α-helix (coordinates, 140–142 amino acid residues) (Fig. 1). This suggested that the presence of an

α-helix at this position would not considerably influence the major lichenase properties. We have also modeled LicBM3 3D structures with Phyre [27] using 3I4I as a potential template (Supplementary Materials, Fig. S1). According to the constructed 3D structure of thermostable lichenase predicted by Phyre, the N- and C-terminal regions of the protein are spatially close (Supplementary Materials, Fig. S1). In addition, the regions of amino acid residues 53, 99, and 140 are localized to loops. That is why it is also possible to attempt a CP of lichenase in the selected regions [9]. Thus, three potential regions were selected for CP in the LicBM3 amino acid sequence namely, regions housing 53, 99, and 140 amino acid residues. 3.2. Properties of CP variants of thermostable lichenase Recombinant CP genes of thermostable lichenase were constructed for experimental verification of in silico predictions. The following designations were accepted for convenience in description of CP in lichenase: N module spans from the starting methionine to the first catalytic amino acid residue (1–104 amino acid residues); C module, after the second catalytic amino acid residue to the terminal amino acid residue (110–210 residues); and Cat module (catalytic module), 105–109 residues (Fig. 2A). Initially, the thermostable lichenase gene variant encoding only the enzyme catalytic domain (LicBM3) was constructed as well as three genes encoding CP LicBM3 variants (designated CN-53, CN-99, and CN-140). In the modified CN-53 and CN-99 variants, 53 and 99 amino acid residues from the N module, respectively, were “transferred” to the C module and in the CN-140 variant, 70 amino acid residues from the C module were “transferred” to the N module (Fig. 2A). Note that the modified CN-53 and CN-99 have a “truncated” lichenase N module (to six amino acid residues for CN-99), while CN-140 variant has a shortened C module with a complete preservation of both the qualitative and quantitative amino acid compositions in all three CP protein variants as compared to LicBM3. The constructed recombinant genes were transferred into pQE expression vectors to obtain the bacterial transformants, designated LicBM3, CN-53, CN-99, and CN-140 according to the names of the corresponding recombinant genes. Analysis of the bacterial transformants by plate test has shown that only three of the four transformants are active at a temperature of 70 °C, namely, LicBM3, CN-53 and CN-99, as demonstrated by lightened spots around colonies (Fig. 2B). Presumably, the absence of activity in the case of CN-140 in plate test is the result of a decrease in its thermostability or activity. Comparative analysis of specific activities of the LicBM3 protein and CP variants CN-53, CN-99, and CN-140 estimated at 70 °C demonstrates that the CN-140 specific activity is dramatically decreased to 4% as compared with LicBM3. CN-53 and CN-99 retain high specific activities at a level of 85 and 95%, respectively, as compared with LicBM3 (Fig. 2A). The question arises here whether the C module is important for lichenase activity or its thermostability.

14

АА Tyurin et al. / Biochimica et Biophysica Acta 1854 (2015) 10–19

with stability of the protein molecule in the case of CN-140 CP variant. To test this assumption, protein lysates of LicBM3 and the CP variants CN-53, CN-99, and CN-140 before and after heating (70 °C for 1 h) were analyzed by SDS-PAGE. As is evident from the described data, CN-140 is actually unstable at 70 °C (Fig. 2E). Note here that the temperature optimums of LicBM3 and all the tested CP variants (CN-53, CN-99, and CN-140) do not change (data not shown). Zymograms have allowed for characterization of the proteins according to their ability to hydrolyze the substrate after PAGE separation. Unlike CN-140, LicBM3 as well as CP variants CN-53 and CN-99 retained this ability (Fig. 2D). The absence of activity in the CN-140 zymogram is explainable by the loss of its ability to refold to an active enzyme conformation in the gel after denaturing electrophoresis as compared with the other variants. However, a critical decrease in thermostability at 70 °C also cannot be excluded for CN-140, as suggested by the data on thermostability estimation and results of protein lysate separation before and after heating at 70 °C. Thus, the results suggest that (i) part of C module containing the breakpoint is probably important for lichenase thermostability and, as a consequence, cannot be used for modifications; (ii) the N module is appropriate for modifications without a loss in the major enzyme properties (activity and thermostability); (iii) the length of N module does not influence the lichenase activity and thermostability (as demonstrated by its absence in CN-99); and the N-module regions involved in circular permutations (53 and 99 amino acid residues) are likely sites for integrating peptides.

3.3. Properties of the variants of thermostable lichenase catalytic domain carrying integrated small peptides as internal modules

Fig. 1. LicBM3 structure obtained based on different PDB structures. The first character denotes the number of construct: 1, the proteins corresponding to LicBM3 and 2–4 (see Table 1), the proteins corresponding to permuted constructs CN-53, CN-99, and CN-140, respectively. P2, p3 and p4 are the breakpoints relative to which homologous regions were manually transposed when constructing LicBM3 pairwise alignments with the PDB sequences similar to constructs 2–4, respectively. β-Sheets are shown with yellow Roman type letters; α-helices, with pink italic; and catalytic amino acid residue, with red bold underlined letters. Cpred, results of analysis of circular permutations; (o) the predicted regions of permitted circular permutations; (−) predicted region with prohibited circular permutations; and (=) regions of circular permutations selected for analysis and further experiment.

To assess thermostability of LicBM3 and the CP variants CN-53, CN99, and CN-140, the residual activity of the proteins was calculated using the incubation of unheated protein lysates at 70 °C. Comparative analysis has shown that CN-99 is comparable in its thermostability to LicBM3; the thermostability of CN-53 after 4-h incubation decreases to approximately 40%; and the CN-140 thermostability drops by 80% as early as after 1-h incubation disappearing at all after 2 h of incubation at 70 °C (Fig. 2C). These data suggest that a high temperature interferes

In order to verify the above assumptions, we have constructed recombinant genes that encode the variants of thermostable lichenase catalytic domain (M variants) with internally integrated small peptides, namely, Gly-Pro-Met-Leu-Gly-Ser (GPMLGS) and Pro-Gly-Met-Leu-LysLeu (PGMLKL), designated P1 and P2, respectively. The amino acid sequences in these peptides were selected based on the following considerations. First, it is known that except for some specific cases, Pro is disfavored in both helices and sheets because it has no backbone N–H group to participate in hydrogen bonding. Gly is also less commonly found in helices and sheets, in part because it lacks a side chain and, therefore, can adopt a much wider range of phi/psi torsion angles in peptides. However, these two residues are strongly associated with βturns, and the sequences, such as Pro-Gly and Gly-Pro are sometimes considered diagnostic for turns [39]. Correspondingly, the dipeptides Gly-Pro and Pro-Gly, respectively, were included into P1 and P2. It should be also emphasized that the proline residue located in some appropriate loop sites or random folds is believed to render proteins more thermostable [40]. Second, the amino acid residues potentially stabilizing 3D structure were included into P1 and P2, namely M and S, carrying hydrophobic side chains and able to form H\H bond, or K, a positively charged amino acid residue with a hydrophilic side chain (in the case of P2). In the M variants NС-L-53 and NС-L-99, P1 and P2 peptides were integrated in 53- or 99-residue region, respectively; in NС-L-53-99, these peptides (P1 and P2) were integrated into the 53- or 99-residue region of the N module; and in NC-L-140, P2 was integrated into the 140residue of C module (Fig. 3A). The constructed recombinant genes were transferred into pQE expression vectors to produce bacterial transformants, designated NC-L53, NC-L-99, NС-L-53-99, and NC-L-140 according to the names of the recombinant genes. Analysis of the bacterial transformants by plate method has demonstrated that all the transformants (NC-L-53, NC-L99, NС-L-53-99, and NC-L-140) display the activity at 70 °C. However, the lightened spot, denoting a lichenase activity, around the colony of NC-L-140 transformant was considerably smaller as compared with LicBM3 and the remaining M variants (NC-L-53, NC-L-99, and NС-L-

АА Tyurin et al. / Biochimica et Biophysica Acta 1854 (2015) 10–19

15

Fig. 2. Structure and properties of LicBM3 and its CP variants CN-53, CN-99, and CN-140. (A) Scheme of proteins and their activities (green block, N module; blue block, C module; and orange, Cat module); (B) plate test of bacterial transformants; (C) thermostability of LicBM3 and its CP variants CN-53, CN-99, and CN-140; (D) zymograms of protein lysates of LicBM3 and the CP variants CN-53, CN-99, and CN-140, M — marker Mm; and (E) electrophoretic pattern of protein lysates before (LicBM3, CN-53, CN-99, and CN-140) and after heating (LicBM3(h), CN-53(h), CN-99(h), and CN-140(h)) (70 °C for 1 h), M — marker Mm.

53-99) (Fig. 3B). This may result from a decrease in NC-L-140 thermostability and/or activity. Comparative analysis of specific activities of the purified LicBM3 and the M variants NC-L-53, NC-L-99, NС-L-53-99, and NC-L-140 estimated at 70 °C has shown that the activity of NC-L-140 amounts to 20% relative to LicBM3. NC-L-53 and NC-L-99 retained a high specific activity, amounting to 90 and 98% as compared with LicBM3 and NС-L-53-99, about 85% of the LicBM3 activity (Fig. 3A). In order to find out whether the M variant NC-L-140, similar to the CP variant CN140, displays a decreased thermostability, the thermostability of LicBM3 and all the M variants (NC-L-53, NC-L-99, NС-L-53-99, and NC-L-140) was analyzed. The comparative analysis confirmed that NC-L-140 actually had a decreased thermostability: its activity reduced by 80% as early as after 1-h incubation at 70 °C and further reduced to only 5% after 4-h incubation (Fig. 3C). These data suggest that that the M variant NC-L-140, similar to the CP variant CN-140, may display a decreased protein stability at a high temperature. Analysis of protein lysates of the M variants NC-L-99 and NC-L-140 before and after heating at 70 °C for 1 h by denaturing SDS-PAGE has shown that the NC-L-140 protein product is undetectable in the heated protein lysate (Fig. 3E), unlike the M variants NC-L-53, NC-L-99, and NСL-53-99. These results reliably suggest that the region of residue 140 in the C module with a circular permutation (CP variant CN-140) or integration of a small peptide (M variant NC-L-140) is likely to be important

for lichenase stability. To find out the particular factors causing a loss in stability of these lichenase variants is of interest for further studies. Unlike NC-L-140, the thermostability of NC-L-53, NC-L-99, and NСL-53-99 was significantly higher after 3 h of incubation at 70°, remaining comparable to that of LicBM3 after the same incubation. A reliable increase in thermostability of NC-L-53, NC-L-99, and NС-L-53-99 after a 3-h incubation at 70 °C may be determined by the amino acid compositions of the integrated peptides, P1 and/or P2. These peptides contain amino acids with nonpolar (hydrophobic) R groups (M and S), able to form additional H\H bonds, or the amino acid residue with a polar (hydrophilic) R group (K), which can be involved in hydrophobic interactions between the nonpolar (hydrophobic) R groups of the amino acids contained in thermostable lichenase. Note that these bonds can stabilize its 3D structure and increase the thermostability of M variants NC-L-53, NC-L-99, and NС-L-53-99 as compared with the native LicBM3. Note also that proline residue may also have a positive effect on thermostabilities of these lichenase M variants [40]. In order to estimate the ability of lichenase M variants to renature to an active conformation in gel after denaturing electrophoresis, the protein lysates were analyzed by zymogram method. As has been shown, the three M variants, namely, NC-L-53, NC-L-99, and NС-L-5399, retained this ability, whereas NC-L-140 did not (Fig. 3D). The absence of NC-L-140 activity in the zymogram is also explainable either by a loss of the ability to refold to an active enzyme in the gel after

16

АА Tyurin et al. / Biochimica et Biophysica Acta 1854 (2015) 10–19

Fig. 3. Structure and properties of LicBM3 and its modified variants NC-L-53, NC-L-99, NС-L-53-99, and NC-L-140. (A) Scheme of proteins and their activities (green block, N module; blue block, C module; orange, Cat module; and red, integrated amino acids); (B) plate test; (C) thermostability; (D) zymograms of protein lysates of LicBM3 and the modified variants, M — marker Mm; and (E) electrophoretic pattern of protein lysates before (1 — LicBM3, 2 — NC-L-53, 3 — NC-L-99, 4 — NС-L-53-99, and 5 — NC-L-140) and after heating (1(h) — LicBM3, 2(h) — NC-L-53, 3(h) — NC-L-99(h), 4(h) — NС-L-53-99(h), and 5(h) NC-L-140) (70 °C for 1 h), M — marker Mm.

denaturing electrophoresis as compared with the other M variants or by a decrease in it thermostability at 70 °C. Thus, it has been confirmed that the regions of lichenase catalytic domain that permit introduction of circular permutations with retention of the enzyme activity and thermostability also permit internal integration of small peptides, including multiple integration. Detection of the internal regions in protein molecules that permit insertion fusion with other protein sequences is of interest for protein engineering, in particular, construction of multifunctional enzymes. Here, the strategy for creating chimeric proteins with the help of insertion fusion, implying that a gene is inserted within the other one, may be beneficial for solving such problems [12,18,19]. Our results and data of other researchers suggest that the C. thermocellum thermostable lichenase has a potential for constructing new bifunctional proteins using insertion fusion in the region of residues 53 and/or 99 of its catalytic domain. Note that the codons encoding the integrated peptides are selected so that they form the

unique restriction sites at the level of nucleotide sequences, namely, BamHI and ApaI (for P1) and HindIII and SmaI (for P2). In future, this will allow them to be used for insertion fusions and design of multifunctional proteins. 3.4. Properties of the thermostable lichenase CP variants carrying integrated small peptides as internal modules Note that cyclically permuted proteins are also used for constructing bifunctional enzymes [12,18]. In particular, it has been demonstrated that a bifunctional enzyme, GluXyn-1, resulting from insertion of the entire xylanase domain into a surface loop of cpMAC-57 (СР variant of a β-glucanase), displays both enzymatic activities at wild-type levels [12]. Correspondingly, we have attempted to clarify whether it is possible to integrate small peptides into CP variant of thermostable lichenase and how such an integration will influence their thermostability and activity. For this purpose, we constructed the recombinant genes based on

АА Tyurin et al. / Biochimica et Biophysica Acta 1854 (2015) 10–19

the genes encoding the CP variants CN-53 and СN-99. These genes were selected because their protein products retained high activity and thermostability (Fig. 2A, C). The constructed chimeric genes code for CP variants with peptide P2 (Pro-Gly-Met-Leu-Lys-Leu) integrated into their junction region (the C- and N-terminal regions of the native lichenase) (Fig. 4A). Note that an insertion fusion of the peptide for both variants fell into the same region, namely, between the C and N termini of the native lichenase and they differ only in the “cutting” site, residues 53 or 99, where the new N and C termini of the molecule were created. Analysis of the bacterial transformants by a plate test has demonstrated that the CN-L-53 and CN-L-99 transformants display a high activity at 70 °C (Fig. 4B), which amounted to about 94% of the CN-53 (without the integrated peptide) activity for CN-L-53 and 97% activity as compared with CN-99 (without the integrated peptide) for CN-L-99 (Fig. 4A). Note here that characteristic of the CP-L variants is an increase in thermostability: CN-L-53 and CN-L-99 display the thermostability elevated by 10 and 6%, respectively, as compared with the CP variants CN-53 and CN-99, which may be determined by the composition of integrated peptide (for discussion, see Section 3.3). The capability of CN-L-53 and CN-L-99 lichenase variants of renaturing to an active enzyme form in gel after denaturing electrophoresis was

17

confirmed by zymogram method (Fig. 4D). As for the stability estimation of CN-L-53 and CN-L-99 protein lysates by assaying these lysates before and after heating (70 °C for 1 h) by SDS-PAGE, it has shown that the integration of small peptides has no significant effect on their stability as compared to both LicBM3 (Fig. 4E) and the CP variants CN53 and CN-99 (Fig. 2E). Thus, we have confirmed that the circularly permuted variants of thermostable lichenase CN-53 and CN-99 as well as the region of N module in the LicBM3 catalytic domain permit internal integration of small peptides and have a potential for constructing bifunctional proteins. 4. Concluding remarks In this work, the regions for construction of CP lichenase variants were predicted with the help of CPred (Circular Permutation Site Predictor); selection of these regions was theoretically justified based on additional in silico analysis and enhanced construction of three CP lichenase variants. Involvement of in silico analysis into CP prediction helped in the selection of circular permutations in lichenase sequence [10,30, 41]. New open termini in two CP lichenase variants (CN-53 and CN99) had no effect on both the activity and thermal tolerance unlike

Fig. 4. Structure and properties of LicBM3 and its CP variants with integrated small peptides CN-L-53 and CN-L-99. (A) Scheme of proteins and their activities (green block, N module; blue block, C module; orange, Cat module; and red, integrated amino acids); (B) plate test; (C) thermostability; (D) zymograms of protein lysates of LicBM3 and the CP variants with integrated small peptides, M — marker Mm; and (E) electrophoretic pattern of protein lysates before (LicBM3, CN-L-53 and CN-L-99) and after heating (LicBM3(h), CN-L-53(h) and CN-L-99(h)) (70 °C for 1 h), M — marker Mm.

18

АА Tyurin et al. / Biochimica et Biophysica Acta 1854 (2015) 10–19

another CP variant, CN-140, which displayed a dramatic decrease in the activity and thermostability, presumably associated with a loss in stability at a high temperature (Fig. 2A, B, C). Note that the CP variants CN-53 and CN-99 are localized to the N module of the catalytic domain and CN140, to the C module (Fig. 2A). We have clarified that the regions of circular permutations are also suitable for integrating small peptides. Construction of modified lichenase variants (M variants) carrying integrated small peptides (six amino acid residues) as internal modules (NC-L-53, NC-L-99, NС-L-5399 и NC-L-140) in the regions preferable for CP (Fig. 3A) and subsequent testing of their activity and thermostability (Fig. 3A, C) allowed for a grounded conclusion that the regions of the lichenase catalytic domain that permit circular permutation with retention of its activity and thermostability also permit internal integration of small peptides, including multiple integration. This conclusion is based on the results of thermostability testing; in particular, the thermostability of lichenase M variants (NC-L-53, NC-L-99, and NС-L-53-99) is in general even better as compared with LicBM3 (catalytic domain) and their activity is similar to LicBM3 (Fig. 3A, C); as for NC-L-140 variant, it displayed a dramatically decreased activity and thermostability (Fig. 3A, C), similar to the analogous CP variant, CN-140 (Fig. 2A, C). Thus, the prediction of CP regions with subsequent experimental verification of the properties retained by CP variants makes it possible to predict also the regions potentially appropriate for integration of peptides as internal modules. Moreover, it has been also demonstrated that the lichenase CP variants that retained their activity and thermostability also permit insertion fusions of small peptides as internal modules between the C and N termini of native lichenase without any dramatic changes in their thermostability and activity (Fig. 4A, C). Note that retention of the ability of M variants (NC-L-53, NC-L-99, and NС-L-53-99) as well as CP variants (СN-L-53 and CN-L-99) to renature and hydrolyze the substrate after denaturing SDS-PAGE (Fig. 3E and E) similar to the intact lichenase, will allow for estimation and control of the integrity of hybrid proteins that could be constructed based on M and СР variants. This work demonstrated the utility of CP method in protein engineering [9 30, 41], namely, that (i) CP predictions can be used for searching protein molecules for the regions permitting internal integration, since it has been shown that the lichenase N-terminal regions permitting circular permutations are more likely to permit integration of another protein, and (ii) active and thermostable lichenase CP variants may be used for insertion fusions with small peptides without any significant loss in the enzyme activity and thermostability. Thus, our results form the background for further enzyme engineering of the C. thermocellum thermostable lichenase aimed at the creation of new bifunctional proteins with the help of insertion fusion, implying insertion of a gene into another one. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.bbapap.2014.10.012.

[4]

[5]

[6]

[7]

[8]

[9] [10]

[11]

[12]

[13]

[14]

[15] [16]

[17]

[18]

[19]

[20]

[21]

[22]

[23]

Acknowledgments The work was supported by the Ministry of Science and Education of the Russian Federation under federal target program (state contract no. 16.512.11.2268; IVG-P).

[24]

[25]

References [26] [1] J.P. Bustamante, A. Bonamore, A.D. Nadra, N. Sciamanna, A. Boffi, D.A. Estrin, L. Boechi, Molecular basis of thermal stability in truncated (2/2) hemoglobins, Biochim. Biophys. Acta 1840 (2014) 2281–2288, http://dx.doi.org/10.1016/j.bbagen.2014.03. 018. [2] M. Mortazavi, S. Hosseinkhani, Design of thermostable luciferases through arginine saturation in solvent-exposed loops, Protein Eng. Des. Sel. 24 (2011) 893–903, http://dx.doi.org/10.1093/protein/gzr051. [3] J.-W. Huang, C.-C. Chen, C.-H. Huang, T.-Y. Huang, T.-H. Wu, Y.-S. Cheng, T.-P. Ko, C.-Y. Lin, J.-R. Liu, R.-T. Guo, Improving the specific activity of β-mannanase from Aspergillus

[27]

[28] [29]

niger BK01 by structure-based rational design, Biochim. Biophys. Acta 1844 (2014) 663–669, http://dx.doi.org/10.1016/j.bbapap.2014.01.011. C.A. Mchugh, R.F. Tammariello, C.B. Millard, J.H. Carra, Improved stability of a protein vaccine through elimination of a partially unfolded state, Protein Sci. 13 (2004) 2736–2743. G.A. Kotzia, N.E. Labrou, Engineering thermal stability of L-asparaginase by in vitro directed evolution, FEBS J. 276 (2009) 1750–1761, http://dx.doi.org/10.1111/j. 1742-4658.2009.06910.x. A. Urvoas, M. Valerio-Lepiniec, P. Minard, Artificial proteins from combinatorial approaches, Trends Biotechnol. 30 (2012) 512–520, http://dx.doi.org/10.1016/j. tibtech.2012.06.001. D. Niu, X-x. Zhou, T-y. Yuan, Z-w. Lin, H. Ruan, W-f. Li, Effect of the C-terminal domains and terminal residues of catalytic domain on enzymatic activity and thermostability of lichenase from Clostridium thermocellum, Biotechnol. Lett. 32 (2010) 963–967, http://dx.doi.org/10.1007/s10529-010-0241-9. W.C. Liu, Y.S. Lin, W.Y. Jeng, J.H. Chen, A.H. Wang, L.F. Shyur, Engineering of dualfunctional hybrid glucanases, Protein Eng. Des. Sel. 25 (2012) 771–780, http://dx. doi.org/10.1093/protein/gzs083. S. Bliven, A. Prlic, Circular permutation in proteins, PLoS Comput. Biol. 8 (2012) e1002445, http://dx.doi.org/10.1371/journal.pcbi.1002445. Y.T. Lee, T.H. Su, W.C. Lo, P.C. Lyu, S.C. Sue, Circular permutation prediction reveals a viable backbone disconnection for split proteins: an approach in identifying a new functional split intein, PLoS One 7 (2012) e43820, http://dx.doi.org/10.1371/journal. pone.0043820. M. Hahn, K. Piotukh, R. Borriss, U. Heinemann, Native-like in vivo folding of a circularly permuted jellyroll protein shown by crystal structure analysis, Proc. Natl. Acad. Sci. U. S. A. 91 (1994) 10417–10421. J. Aÿ, M. Hahn, K. Decanniere, K. Piotukh, R. Borriss, U. Heinemann, Crystal structures and properties of de novo circularly permuted 1,3-1,4-beta-glucanases, Proteins 30 (1998) 155–167. O. Olsen, K.K. Thomsen, J. Weber, J.O. Duus, I. Svendsen, C. Wegener, D. von Wettstein, Transplanting two unique beta-glucanase catalytic activities into one multienzyme, which forms glucose, Biotechnology (N Y) 14 (1996) 71–76. P. Lu, M.-G. Feng, Bifunctional enhancement of a β-glucanase–xylanase fusion enzyme by optimization of peptide linkers, Appl. Microbiol. Biotechnol. 79 (2008) 579–587, http://dx.doi.org/10.1007/s00253-008-1468-4. L. Liu, L. Wang, Z. Zhang, X. Guo, X. Li, H. Chen, Domain-swapping of mesophilic xylanase with hyper-thermophilic glucanase, BMC Biotechnol. 12 (2012) 28. J. Cota, L.C. Oliveira, A.R.L. Damásio, A.P. Citadini, Z.B. Hoffmam, T.M. Alvarez, C.A. Codima, V.B.P. Leite, G. Pastore, M. de Oliveira-Neto, M.T. Murakami, R. Ruller, F.M. Squina, Assembling a xylanase–lichenase chimera through all-atom molecular dynamics simulations, Biochim. Biophys. Acta 1834 (2013) 1492–1500. G.J. Kim, Y.H. Cheon, H.S. Kim, Directed evolution of a novel N-carbamylase/Dhydantoinase fusion enzyme for functional expression with enhanced stability, Biotechnol. Bioeng. 20 (2000) 211–217. G.S. Baird, D.A. Zacharias, R.Y. Tsien, Circular permutation and receptor insertion within green fluorescent proteins, Proc. Natl. Acad. Sci. U. S. A. 96 (1999) 11241–11246. G.P. Furtado, L.F. Ribeiro, M.R. Lourenzoni, R.J. Ward, A designed bifunctional laccase/β-1,3-1,4-glucanase enzyme shows synergistic sugar release from milled sugarcane bagasse, Protein Eng. Des. Sel. 25 (2012) 1–9, http://dx.doi.org/10. 1093/protein/gzs057. C. Dumon, A. Varvak, M.A. Wall, J.E. Flint, R.J. Lewis, J.H. Lakey, C. Morland, P. Luginbühl, S. Healey, T. Todaro, G. De Santis, M. Sun, L. Parra-Gessert, X. Tan, D.P. Weiner, H.J. Gilbert, Engineering hyperthermostability into a GH11 xylanase is mediated by subtle changes to protein structure, J. Biol. Chem. 283 (2008) 22557–22564, http://dx.doi.org/10.1074/jbc.M800936200. Y.-S. Cheng, C.-H. Huang, C.-C. Chen, T.-Y. Huang, T.-P. Ko, J.-W. Huang, T.-H. Wu, J.-R. Liu, R.-T. Guo, Structural and mutagenetic analyses of a 1,3-1,4-β-glucanase from Paecilomyces thermophile, Biochim. Biophys. Acta 1844 (2014) 366–373, http://dx.doi.org/10.1016/j.bbapap.2013.11.005. V.V. Zverlov, W.H. Schwarz, Bacterial cellulose hydrolysis in anaerobic environmental subsystems—Clostridium thermocellum and Clostridium stercorarium, thermophilic plant-fiber degraders, Ann. N. Y. Acad. Sci. 1125 (2008) 298–307, http://dx.doi. org/10.1196/annals.1419.008. S. Schimming, W.H. Schwarz, W.L. Staudenbauer, Structure of the Clostridium thermocellum gene licB and the encoded β-1,3-1,4-glucanase. A catalytic region homologous to Bacillus lichenases joined to the reiterated domain of clostridial cellulases, Eur. J. Biochem. 204 (1992) 13–19. E.S. Piruzian, I.V. Goldenkova, K.A. Musiychuk, N.S. Kobets, I.P. Arman, I.V. Bobrysheva, I.A. Chekhuta, D. Glazkova, A reporter system for prokaryotic and eukaryotic cells based on the thermostable lichenase from Clostridium thermocellum, Mol. Genet. Genomics 266 (2002) 778–786. S.F. Altschul, T.L. Madden, A.A. Schaffer, J. Zhang, Z. Zhang, W. Miller, D.J. Lipman, Gapped BLAST and PSI-BLAST: a new generation of protein database search programs, Nucleic Acids Res. 25 (1997) 3389–3402. H.M. Berman, J. Westbrook, Z. Feng, G. Gilliland, T.N. Bhat, H. Weissig, I.N. Shindyalov, P.E. Bourne, The Protein Data Bank, Nucleic Acids Res. 28 (2000) 235–242. L.A. Kelley, M.J. Sternberg, Protein structure prediction on the Web: a case study using the Phyre server, Nat. Protoc. 4 (2009) 363–371, http://dx.doi.org/10.1038/ nprot.2009.2. R.C. Edgar, MUSCLE: multiple sequence alignment with high accuracy and high throughput, Nucleic Acids Res. 32 (2004) 1792–1797. W.-C. Lo, C.-Y. Lee, C.-C. Lee, P.-C. Lyu, iSARST: an integrated SARST web server for rapid protein structural similarity searches, Nucleic Acids Res. 37 (2009) W545–W551, http://dx.doi.org/10.1093/nar/gkp291 (Web Server issue).

АА Tyurin et al. / Biochimica et Biophysica Acta 1854 (2015) 10–19 [30] W.-C. Lo, L.-F. Wang, Y.-Y. Liu, T. Dai, J.-K. Hwang, P.-C. Lyu, CPred: a web server for predicting viable circular permutations in proteins, Nucleic Acids Res. 40 (2012) W232–W237, http://dx.doi.org/10.1093/nar/gks529. [31] K.A. Musiychuk, I.V. Goldenkova, R.M. Abdeev, N.S. Kobets, E.S. Piruzian, Preparation and properties of Clostridium thermocellum lichenase deletion variants and their use for construction of bifunctional hybrid proteins, Biochemistry (Mosc.) 65 (2000) 1397–1402. [32] U.K. Laemmli, Cleavage of structural proteins during the assembly of the head of bacteriophage T4, Nature 227 (1970) 680–685. [33] M.M. Bradford, Rapid and sensitive method for quantitation of microgram quantities of protein utilizing principle of protein-dye binding, Anal. Biochem. 72 (1976) 248–254. [34] G.L. Miller, Use of dinitrosalicylic acid reagent for determination of reducing sugar, Anal. Chem. 31 (1959) 426–428. [35] W.C. Lo, P.C. Lyu, CPSARST: an efficient circular permutation search tool applied to the detection of novel protein structural relationships, Genome Biol. 9 (2008) R11, http://dx.doi.org/10.1186/gb-2008-9-1-r11.

19

[36] Y. Nakatani, T.D. Nalder, G.W. Tannock, J.F. Cutfield, R.W. Jack, A. Carne, Crystal Structure of a Prokaryotic Beta-1,3-1,4-glucanase (lichenase) Derived From a Mouse Hindgut Metagenome, 2014. [37] T. Keitel, O. Simon, R. Borriss, U. Heinemann, Molecular and active-site structure of a Bacillus 1,3-1,4-beta-glucanase, Proc. Natl. Acad. Sci. U. S. A. 90 (1993) 5287–5291. [38] J. Aÿ, F. Götz, R. Borriss, U. Heinemann, Structure and function of the Bacillus hybrid enzyme GluXyn-1: native-like jellyroll fold preserved after insertion of autonomous globular domain, Proc. Natl. Acad. Sci. U. S. A. 95 (1998) 6613–6618. [39] R.W. Williams, A. Chang, D. Juretić, S. Loughran, Secondary structure predictions and medium range interactions, Biochim. Biophys. Acta 916 (1982) 200–204. [40] C. Zhou, Y. Xue, Y. Ma, Enhancing the thermostability of α-glucosidase from Thermoanaerobacter tengcongensis MB4 by single proline substitution, J. Biosci. Bioeng. 110 (2010) 12–17. [41] W.C. Lo, C.C. Lee, C.Y. Lee, P.C. Lyu, CPDB: a database of circular permutation in proteins, Nucleic Acids Res. 37 (2009) D328–D332, http://dx.doi.org/10.1093/nar/ gkn679.