Comparison of the Values of a Novel Exolytic and Two Endolytic Alginate Lyases with Mannuronate Preference for Direct Preparation of Oligosaccharides, Action Modes, and Underlying Catalytic Mechanisms

Recent exploration of tool-like alginate lyases has focused on their oligosaccharide products and corresponding substrate action modes, and most were characterized as endolytic lyases with guluronate (G) preference. Herein, we elucidated a novel exolytic lyase, Aly-6, and two typical endolytic lyases, AlgL-Pae and AlgL-Avi, all with mannuronate (M) preferences. AlgL and heparinase_II_III modules play essential roles in determining the similar characteristics of these enzymes, although they are quite different in sequence characteristics. Aly-6 degraded substrates completely by continuously cleaving various monosaccharide units from nonreducing ends and producing various size-dened ΔG-terminated oligosaccharide fractions as intermediate alginate digests, which was inhibited by uorescent labeling of reducing ends. Distinctly, AlgL-Pae and AlgL-Avi varied their action modes toward associated alginate substrates and therefore eventually degraded alginate into various size-dened oligosaccharide products with a specic structure-based succession rule. This study provided new insights into the action modes, associated mechanisms, and enzyme applications of M-preferred lyases.


Introduction
Alginate is the most important polysaccharide component in the Phaeophyceae family of brown algae, which includes such genera as Ascophyllum, Laminaria, and Sargassum. Usually, alginate contributes approximately 40% of seaweed dry weights (1). Alginate is composed of β-(1, 4)-linked uronic residues, i.e., β-D-mannuronate (M) and its C5-epimer α-L-guluronate (G), and thus forms in homogeneous fashion, i.e., poly-M and poly-G blocks, or a heterogeneous fashion, e.g., poly-MG and poly-GM blocks, within the linear molecule (2)(3)(4). Algal alginate, in particular G-enriched polysaccharides, is widely used in the food and pharmaceutical industries due to its excellent gel-forming capability and various associated bene cial effects (5)(6)(7)(8)(9). In contrast, alginate oligosaccharide products have been identi ed with important biological activities, e.g., antibacterial (10), antiobesity (11), antioxidation (12), and antiin ammatory effects (13), which are closely related to the molecular size (degree of poly-Merization, DP), the M/G ratio, and the molecular modi cation type, e.g., acetylation or sulfation. In 2019, a novel drug derived from oligo-mannuronate, GV-971 (14), was permitted for sale in China to treat Alzheimer's disease, which also means an economic value. Thus, direct preparation of sugar chains with designated sugar components and molecular sizes from alginate has urgently become a technical problem.
Distinctly, throughout the alginate-degrading process, exolytic alginate lyases primarily yield the monosaccharide product of the Δ unit, which is further converted into 4-deoxy-L-erythro-5-hexoseulose uronate (DEH) under nonenzymatic conversion (25); thus, exolytic enzymes are essential for the bioconversion of alginate into biofuels. Hence, an increasing number of endolytic alginate lyases have been explored as resources and improved as enzymes with respect to their substrate preferences (M-or Gspeci c, or bifunctional but preferring M or G), oligosaccharide products (e.g., the sizes of oligosaccharide products, the molar ratio of various product fractions, and the structure properties and inner changing rules), and corresponding substrate-degrading modes (e.g., the sizes of the smallest substrates and products, the yielding position of the smallest products, and the endolytic or exolytic patterns) (26), which are essential to clearly and exactly display the utility of enzymes in the direct preparation of targeting unsaturated oligoalginate chains.
For instance, the endolytic guluronate lyase Aly5 (27), a PL7 family member from the polysaccharidedegrading marine bacterium Flammeovirga sp. MY04 is valuable for producing unsaturated disaccharide products primarily consisting of ΔG units, unsaturated trisaccharide products containing ΔG in the presence of ΔM ends, and larger unsaturated products losing ΔG until only ΔM ends remain. Notably, the large size-de ned nal oligosaccharide products, e.g., UDP5, UDP6, and UDP7 fractions produced by rAly5, were identi ed as ΔM-ended and M-enriched unsaturated sugar chains. Therefore, alginate degradation by Aly5 can provide an alternative enzymatic preparation strategy to the chemical or physical strategies of preparing the novel drug GV-971. Two other PL7 family members, Aly1 (28) and Aly2 (29), which are bifunctional and G-preferred endolytic alginate lyases from the same MY04 strain, showed oligosaccharide products similar to those of Aly5, e.g., primarily disaccharide products of ΔG units, whereas for their bifunctional substrate preferences and greater enzyme activities, these two bifunctional enzymes yield only little large oligosaccharide fractions (> UDP4) in their nal alginate digests. Hence, although these two G-preferred bifunctional alginate lyases are endolytic enzymes, they are thought invaluable for the preparation of large and bioactive unsaturated oligoalginates. Notably, although tens of alginate lyases have been reported with G speci city or G preference in recent decades, fewer than ve endolytic enzymes have been well elucidated for oligosaccharide preparation purposes, which is an even lower proportion when compared to the larger number of reported alginate lyases with M speci city or M preference. Therefore, we are interested in the utility and associated catalytic mechanisms of M-preferred lyases and the differences between endolytic and exolytic enzymes.
A number of bacterial strains of the Pseudomonas and Azotobacter genera can secrete extracellular alginate, which contains acetyl modi cation at the O-2 or O-3 positions of sugar rings (30)(31)(32)(33) and are important components in drug-resistant bacterial bio lms. These bacteria encoded the periplasmic AlgL proteins of the PL5 family by the gene algL localized within alginate biosynthesis operons. Through gene knockout and plasmid complementary tests of bacteria, these PL5 alginate lyases were identi ed to have essential roles in splicing large polysaccharide molecules into smaller molecules during alginate secretion. To date, seven AlgL module-containing proteins of the PL5 family have been reported to have crystal structures, key active site residues (e.g., the half-conserved NNHSYW motif), and associated functional roles in catalysis. Moreover, an increasing number of AlgL-conserved PL5 proteins have been widely identi ed by molecular mining of bacterial genomic data. However, relatively little is known about these enzymes' oligosaccharide products and corresponding substrate action modes, except for their Mspeci cities or preferences, which urgently needs to be overcome for direct preparation, structure identi cation, and functional explorations of novel unsaturated oligosaccharide chains by the use of these alginate lyases.
In this study, the wild-type genes of AlgL-Pae and AlgL-Avi were initially codon-optimized, arti cially synthesized, and then cloned and expressed in Escherichia coli strain BL21(DE3) for comparison to the wild-type Aly-6 gene of Flammeovirga sp. strain MY04. The resulting recombinant proteins were individually puri ed for comparative studies on their biochemical characteristics, enzymatic properties (substrate preference, substrate-degrading modes, and the resulting oligosaccharide products), and particularly their structural relationships. Furthermore, gene truncations and site-directed mutations were performed using rational design of homology-based protein structure modeling to determine the enzymatic function changes and associated catalytic mechanisms.
Ltd., USA. Acetylated alginate fractions were purchased from Biosynth Carbosynth Co. Ltd., China. Poly-G blocks, poly-M blocks, and standard size-de ned G-enriched or M-enriched saturated sugar chains (ranging from disaccharide to heptosaccharide in size, with > 95% promised purities) were purchased from Qingdao Biozhi Biotech Co. Ltd, China. Various size-de ned unsaturated oligosaccharide fractions were prepared using the alginate lyases rAly5 and rAly1 as described in previous studies (27)(28) or using recombinant alginate lyases rAly-6, Pae-rAlgL, and Avi-rAlgL in this study.

Gene and protein sequences.
The DNA sequence of ORF2549 from the genome of Flammeovirga sp. strain MY04 was translated into the amino acid sequence of the Aly-6 protein, and the GC content (G + C%) was calculated using BioEdit version 7.2.5. Signal peptides were analyzed using SignalP server 5.0 (http://www.cbs.dtu.dk/services/). Molecular weights and the isoelectric point (pI) of the protein were estimated using the peptide mass tool on the ExPASy server of the Swiss Institute of Bioinformatics (http://swissmodel.expasy.org/). Online similarity searches of the Aly-6 protein sequence were performed using the BLAST algorithm on the National Center for Biotechnology Information server (http://www.ncbi.nlm.nih.gov). Protein modules and domains were identi ed using the Simple Modular Architecture Research Tool (https://en.wikipedia.org/wiki/Simple_Modular_Architecture_Research_Tool), the Pfam database (http://pfam.xfam.org), and the Carbohydrate-Active Enzyme database (http://www.cazy.org). Multiple sequence alignments and phylogenetic analyses were performed using MEGA version 7.2.5 (35). Two wild-type DNA sequences that encode the periplasmic AlgL proteins of P. aeruginosa (under the GenBank nucleotide No. U27829.1/Protein No. AAA91127.1) and A. vinelandii (Nucleotide No. AF027499/Protein No. AAC04567.1) were analyzed as described above.
The wild-type genes of AlgL-Pae and AlgL-Avi were optimized for G + C content and nucleotide codons, arti cially synthesized, and nally directly cloned into the plasmid vector pET-30a (+) between Nde I and Xho I restriction sites, thus forming the recombinant plasmids pE30-AlgL-Pae and pE30-AlgL-Avi to produce the recombinant proteins Pae-rAlgL and Avi-rAlgL, respectively. Markedly, the resulting proteins included no signal peptide at the N-terminus but rather a 6×His tag at the newly formed C-terminus. The integrity of the nucleotide sequences of newly constructed plasmids was con rmed by DNA sequencing.

Heterologous expression and puri cation of recombinant proteins.
Recombinant proteins were expressed and puri ed using the same procedures described for the recombinant alginate lyase rAly5 of Flammeovirga sp. strain MY04 (27). Brie y, each recombinant plasmid was transformed into E. coli BL21(DE3) cells. To initiate protein expression, LB broth was supplemented with isopropyl 1-thio-β-D-galactoside to a nal concentration of 0.05 mM when the A 600 reached 0.8. After a continual 24 h cultivation at 16°C, cells were harvested by centrifugation at 6,000 × g for 10 min, washed twice using ice-cold buffer A (50 mM Tris, 150 mM NaCl, pH 8.0), resuspended in buffer A, and disrupted by sonication (60 repetitions, 5 s). After centrifugation at 15,000 × g for 30 min, the supernatant containing each soluble protein was loaded onto a buffer A-equilibrated Ni-nitrilotriacetic acid agarose (Ni-NTA) column (TaKaRa, Dalian, China). Subsequently, each column was eluted using buffer A that contained imidazole in increasing concentrations, speci cally, 0, 10, 50, and 250 mM.
Fractionated protein samples were analyzed using SDS-PAGE. To obtain active alginate lyases, puri ed protein fractions were dialyzed against buffer B (50 mM Tris, 50 mM NaCl, 5% glycerol (v/v), pH 8.0). SDS-PAGE was performed using 13.2% (w/v) polyacrylamide gels according to the methods of Sambrook and Russell (36). Proteins were detected by staining the gels with Coomassie Brilliant Blue R-250. Protein concentrations were individually determined by the Folin-Lowry method using Folin and Ciocalteu's phenol reagent (Sigma-Aldrich, USA) with bovine serum albumin as the standard.

Enzyme activity assays.
To determine the substrate preference of each puri ed recombinant protein, various polysaccharides and oligosaccharides were individually dissolved in deionized water to prepare stock solutions (3 mg/ml).
Each stock solution (100 µl) was mixed with 30 µl of appropriately diluted enzyme preparation, 100 µl of 150 mM NaAc-HAc buffer (pH 6.0), and 70 µl of water. Each reaction was performed at 40°C for 12 h. Enzyme-treated samples were heated in boiling water for 10 min and subsequently ice-cooled. After centrifugation at 15,000×g for 15 min, the supernatant was collected and analyzed by measuring the absorbance at 235 nm. One unit was de ned as the amount of enzyme required to increase the absorbance at 235 nm by 0.1 per min (23), and samples (1.0 mg/ml) were analyzed by gel ltration on a Superdex peptide 10/300 GL column (GE Healthcare, USA) and had their absorbance monitored at 235 nm using a UV detector. The mobile phase was 0.2 M NH 4 HCO 3 , and the ow rate was 0.4 ml/min. Online monitoring and data analysis were performed using LCsolution version 1.25 software.

Biochemical characterization of recombinant proteins.
To determine the optimal temperature for alginate lyase activities, alginate, poly-G blocks, and poly-M blocks were individually reacted with the protein preparations. Enzymatic reactions were performed in 50 mM NaAc-HAc buffer (pH 6.0) at temperatures ranging from 0 to 80°C for 90 min. After the optimal temperature was determined, the effects of pH on enzyme activity were tested in different buffers, including 50 mM NaAc-HAc buffer (pH 5.0, and 6.0), 50 mM NaH 2 PO 4 -Na 2 HPO 4 buffer (pH 6.0, 7.0, and 8.0), and 50 mM Tris-HCl buffer (pH 7.0, 8.0, 9.0, and 10), each with a total volume of 300 µl. The thermostability was evaluated by measuring the residual enzyme activity of each enzyme preparation after incubation for 0 to 24 h at 0 to 80°C. The effects of pH on enzyme stability were determined by measuring the residual activity of each enzyme after incubation at 4°C at varying pH values (5.0-10) for 2 h. The effects of metal ions and chelating agents on alginate lyase activities were examined by determining the activity of each enzyme in the presence of 1 mM metal ion or 10 mM chelating agent. The effects of NaCl on alginate lyase activities were examined by determining the activity of each enzyme in different concentration (0.0 to 1.0 M). All reactions were performed in triplicate. After each treatment, the enzyme activity was estimated by measuring the absorbance at 235 nm as described previously in 2.5.

Comparison of polysaccharide-degrading patterns.
To compare the polysaccharide degradation patterns of the recombinant proteins, alginate (1.0 mg/ml) digestion by each enzyme (1.0 U/ml) at 40°C was traced over 72 h. Similar experiments were performed at alginate concentrations ranging from 1.0 mg/ml to 10 mg/ml. Aliquots of the digests were removed for time-course analysis. To determine the molar ratios of oligosaccharide fractions in the products, samples (1.0 mg/ml) were analyzed as described in 2.5.
To determine the oligosaccharide compositions of the nal alginate degradation products, 100 mg of alginate (1.0 mg/ml) was initially digested by an excess of each recombinant enzyme (10 U/ml) at 40°C and pH 6.0 for more than 72 h. Moreover, similar operations were performed for rAly-6 and rTF-Aly-6 at pH 7.0 to obtain intermediate alginate digests. To further obtain each size-de ned unsaturated oligosaccharide product fraction, the nal alginate degradation products of Pae-rAlgL and Avi-rAlgL or the intermediated alginate digests by rAly-6 and rTF-Aly-6 were individually gel-ltered through a Superdex peptide 10/300 GL column using the same protocol as described in 2.5. Each fraction was collected and freeze-dried repeatedly to remove NH 4 HCO 3 for further analysis. The molecular mass of each oligosaccharide fraction was determined by matrix-assisted laser desorption/ionization time-of-ight mass spectrometry (AXIMA-CFR plus, Shimadzu, Japan). For 1 H-NMR spectroscopy, each puri ed oligosaccharide fraction (~ 2 mg) was dissolved in 0.3 ml of D 2 O in 5-mm NMR tubes. The spectra were recorded on a JNM-ECP600 (JEOL, Japan) apparatus set at 600 MHz using TMS as the internal standard.

Comparison of oligosaccharide degradation patterns.
To determine the smallest substrate for each recombinant enzyme, unsaturated oligosaccharides with different degrees of poly-Merization (DPs), e.g., UDP2, UDP3, UDP4, UDP5, UDP6, and UDP7 fractions, were rst puri ed from the nal alginate digests by various endolytic alginate lyases, speci cally rAly1 (bifunctional but G-preferred) and rAly5 (G-speci c) of Flammeovirga sp. strain MY04. In addition, unsaturated oligosaccharide fractions were also puri ed from alginate that had been completely digested by the recombinant enzymes Pae-rAlgL and Avi-rAlgL or alginate that had been partially digested by the recombinant enzymes rAly-6 or rTF-Aly-6 as described in 2.7. Then, the obtained oligosaccharide fractions were used as substrates for each recombinant enzyme preparation. To further determine the substrate preference and oligosaccharide degradation patterns of each enzyme, saturated standard sizede ned M-enriched and G-enriched sugar chains were individually reacted with the recombinant enzyme preparations. The tested natural substrates and their enzymatic products (20 µg each) were subjected to the gel ltration assay described in 2.5.
To fully compare the enzymatic degradation patterns of each recombinant enzyme preparation, various size-de ned saturated and unsaturated oligosaccharide fractions were uorescently labeled at their reducing ends using excess 2-aminobenzamide (2-AB, Sigma-Aldrich, USA) (37). The arti cially labeled products (~ 1 µg each) were puri ed by gel-ltration HPLC and further reacted with each enzyme (5 U) in a total volume of 1 ml using the protocol described in 2.7. The above arti cial substrates and their enzymatic products (50 ng each) were subjected to the gel ltration assay described in 2.5 and monitored with a uorescence detector and excitation and emission wavelengths of 330 and 420 nm, respectively.
2.9 Analysis of the catalytic mechanism of the recombinant proteins.

Sequence characteristics of the alginate lyases.
ORF2549 in the genome of Flammeovirga sp. strain MY04 was predicted to encode a candidate polysaccharide lyase, Aly-6 (GenBank Protein Accession No. ANQ49918.2). The gene is 2,238 bp in fulllength with a GC content of 36.8%. The putative protein Aly-6 is composed of 745 amino acid residues with an apparent molecular mass of 84.67 kDa. The predicted pI value is 7.30. SignalP 5.0 analysis indicated that the signal peptide of Aly-6 contains 23 amino acid residues (Met 1 to Ser 23 ) (Fig. 1C).
Analyses using the Carbohydrate-Active Enzyme database and the Simple Modular Architecture Research Tool indicated that the Aly-6 protein contains a putative N-terminal catalytic module (Ala 76 to Phe 320 ) associated with PL5 alginate lyases (AlgL), as well as a putative C-terminal module (Leu 391 to Val 664 ) associated with heparinase II or III (Hep II_III) from Pedobacter heparinus (Fig. 1C). BLASTp searches showed that among characterized enzymes, the whole Aly-6 protein shares the greatest sequence identity (38.93%) with the exo-type alginate lyase Alg17c from Saccharophagus degradans strain 2-40, followed by 38-35% identities with four other reported PL17 family alginate lyases (Fig. 1D): OalS17 of Shewanella sp. strain Kz7, AlgL of Sphingomonas sp. strain MJ-3, OAL of Stenotrophomonas maltophilia, and AlyII of Pseudomonas sp. strain OS-ALG-9. Subsequently, the full-length Aly-6 protein shares very low sequence identities with PL15 and PL5 family alginate lyases (Table. S2). The PL17 proteins described above are all organized in an AlgL and Hep_II_III complex modular architecture similar to that of Aly-6. Notably, the whole Aly-6 protein shares no homology with studied PL6 or PL7 alginate lyases, including Aly1, Aly2, and Aly5 from the same Flammeovirga strain MY04, which contains neither AlgL-like nor Hep_II_III-like modules.
Furthermore, protein sequence alignment showed that the AlgL module of Aly-6 contains one putative catalytic motif, H 195 -N 196 -H 197 -S 198 -T 199 -W 200 , which is half-conserved but different from the conserved NNHSYW catalytic motif of diverse PL5 alginate lyases and the conserved HNHG(A)TW catalytic motif of various PL17 alginate lyases (Fig. S1). Phylogenetic analysis indicated that Aly-6, together with the six genome-predicted alginate lyases that are organized in the same AlgL and Hep_II_III complex modular architecture, i.e., four proteins of other Flammeovirga strains, one of Roseivirga ehrenbergii, and one of R. echinicomitans, are clustered into a novel separate branch within the PL17 superfamily (Fig. 1D).

Production and puri cation of the full-length and truncated recombinant proteins.
The full-length Aly-6 gene was ampli ed directly from the genomic DNA of Flammeovirga sp. strain MY04. The 2.3-kb PCR products produced by the primers E30Aly-6-F and E30Aly-6-R were gel-recovered and enzyme-cloned into the pET30a (+) vector downstream of a T7 promoter. Thus, a His×6 tag was successfully added to the C-terminus of the protein product (rAly-6) encoded by the resulting expression vector (pET30-Aly-6). Similar operations were performed on the PCR products produced by three primer pairs, TF-Aly-6-F and TF-Aly-6-R, TF-Aly-6-Lm-F and TF-Aly-6-Lm-R, and TF-Aly-6-Hpm-F andTF-Aly-6-Hpm-F, to obtain the expression vectors pCTF-Aly-6, pCTF-Aly-6-Lmodule, and pCTF-Aly-6-HPmodule, which added a His×6 tag followed by a TF factor to the N-terminus of the full-length protein product (rTF-Aly-6) and two truncated protein products (rTF-Aly-6-Lmodule and rTF-Aly-6-Hpmodule) individually, thus generating recombinant fusion proteins. To express and purify the recombinant proteins Pae-rAlgL and Avi-rAlgL, similar operations were performed using the pET30a (+)-derived recombinant vectors pE30-AlgL-Pae and pE30-AlgL-Avi, which added a His×6 tag at the C-terminus of the AlgL-Pae and AlgL-Avi proteins, respectively. SDS-PAGE analyses indicated that BL21(DE3) cells harboring each of the above recombinant plasmids produced soluble proteins (Fig. S2), with the correct apparent molecular mass and yields greater than 1.0 g/liter. After sonication and centrifugation, crude enzymes were each extracted from the E. coli cultures. The soluble protein fractions containing rAly-6, rTF-Aly-6, rTF-Aly-6-Lmodule, rTF-Aly-6-Hpmodule, Pae-rAlgL, and Avi-rAlgL were eluted from a Ni-nitrilotriacetic acid (NTA) column using imidazole at concentrations above 50 mM. Further SDS-PAGE analyses indicated that the puri ed soluble proteins each had purities greater than 99% and initial concentrations greater than 1.0 mg/ml (Fig. S2).

Page 10/25
The recombinant proteins rAly-6 and rTF-Aly-6 showed the same substrate spectrum, i.e., they did not digest chondroitin, chondroitin sulfates (A, C, D, and E types), dermatan sulfate B, hyaluronan, heparin, heparin sulfate, pectin, or xanthan, but they e ciently digested alginate, M blocks, and G blocks to produce oligosaccharide products exhibiting strong absorbance at 235 nm. The results suggested that the Aly-6 protein of Flammeovirga sp. strain MY04 is a bifunctional alginate lyase. Furthermore, enzyme activity tests of rAly-6 and rTF-Aly-6 indicated a similar substrate preference for M over G (Table. 1). Quite different from the whole proteins rAly-6 or rTF-Aly-6, the recombinant soluble truncated Aly-6 proteins, i.e., rTF-Aly-6-Lmodule and rTF-Aly-6-HPmodule, showed little degradation activity against any tested polysaccharides, including alginate-associated substrates (Table. 1), indicating that both the AlgL-like and the Hep_II_III-like modules are integral for the alginate lyase activity of the whole Aly-6 enzyme of Flammeovirga sp. strain MY04. Furthermore, the above results demonstrated that in the Aly-6 protein, the Hep_II_III-like module is only a putative element instead of a catalytic module against any tested glycoaminoglycans, including heparin or heparin sulfates. The puri ed protein fractions of Pae-rAlgL and Avi-rAlgL could e ciently degrade only alginate and M blocks to generate unsaturated oligosaccharide products, and they could hardly digest any other tested polysaccharides, including G blocks (Table. 1), demonstrating that they are M-speci c lyases, as reported previously. In addition, none of the above protein preparations degraded acylated alginate from Azotobacter or Pseudomonas bacterial strains to produce detectable oligosaccharide products.
The whole enzyme rAly-6 demonstrated the highest activity at 40°C when alginate, M-enriched blocks, or G-enriched blocks were used as substrates (Fig. S3A). A thermostability assay further showed that the alginate-degrading activity of rAly-6 was stable at 0 to 30°C, and more than 60% activity was retained even if the enzyme was preincubated at 30°C for 24 h (Fig. S3C). The optimal pH, determined at 40°C in both 50 mM sodium acetate-acetic acid (NaAc-HAc) buffer and 50 mM NaH 2 PO 4 -Na 2 HPO 4 buffer, was 6.0 (Fig. S3B). The enzyme retained more than 60% of the highest activity after preincubation for 2 h at pH 5.0 to 8.0 (Fig. S3B). The TF-factor-fused recombinant enzyme rTF-Aly-6 showed similar biochemical characteristics, i.e., the optimal temperature and pH value for catalysis and the enzyme's thermal and pH stabilities, to rAly-6.
The alginate lyase activities of rAly-6 were strongly inhibited by 1.0 or 10 mM Ag + , Cu 2+ , Hg 2+ , sodium dodecyl sulfonate (SDS), 10 mM Pb 2+ , Zn 2+ , Cr 3+ , Fe 3+ , or ethylenediaminetetraacetic acid (EDTA). In contrast, the enzyme activity of rAly-6 was increased to 130 ~ 160% by various concentrations (1.0 or 10 mM) of Co 2+ , Mn 2+ , and Ni 2+ . Chemicals such as glycerol, dithiothreitol (DTT), and the reducing agent βmercaptoethanol (β-ME) weakly increased the activity of Aly-6 (Fig. S3E). Moreover, the enzyme activity of rAly-6 was strongly increased by increasing the NaCl concentration from 0.0 to 0.5 M, reaching approximately 210% at 0.2 M, whereas it was strongly inhibited by increasing the NaCl concentration from 0.5 to 1.0 M, decreasing to approximately 10% at 1.0 M (Fig. S3D). The results indicated that the alginate lyase Aly-6 is active without NaCl while being adapted to a range of NaCl concentrations, which may be associated with the protein origin, Flammeovirga sp. strain MY04, being derived from coastal environments. Similar to that of rAly-6, the enzyme activities of Avi-rAlgL and Pae-rAlgL were both strongly inhibited by 1.0 or 10 mM SDS or 10 mM Cu 2+ , Hg 2+ , Pb 2+ , Zn 2+ , Cr 3+ , or Fe 3+ (Fig. S4A and B). Moreover, their enzyme activities were NaCl independent and increased by increasing the NaCl concentration from 0 to 1.0 M, reaching approximately 130% at over 0.2 M NaCl, while Avi-rAlgL activity reached 160% at 0.6 M ( Fig. S4C and D).
Under optimal conditions (40°C in 50 mM NaAc-HAc buffer, pH 6.0), the enzyme rAly-6 showed speci c activities of 726 ± 2.2, 525 ± 3.5, 196 ± 2.9 U/mg in the degradation of alginate, M blocks, and G blocks. These results demonstrated that the enzyme activity of Aly-6 against M blocks is almost 2.71-fold that against G blocks, and therefore Aly-6 is an M-preferred lyase. Similarly, the reported M-speci c lyases Pae-rAlgL and Avi-rAlgL exhibited activities of 2685/4219, 5704/8085, and 144/98 U/mg in the degradation of the corresponding substrates (Table. 1), respectively.

Polysaccharide degradation patterns and oligosaccharide products.
Unsaturated oligosaccharides with a high degree of poly-Merization (DP) were the main products in the initial stage of the reaction and were then gradually converted into smaller oligosaccharides ( Fig. S5A and B), indicating that Avi-rAlgL and Pae-rAlgL are typical endo-type alginate lyases. The nal products were similar (UDP2 ~ UDP7) but with differing molar ratios for each component (Fig. S5D). These enzymes also degrade unsaturated oligosaccharides endolytically. Unlike the two endolytic agents, rAly-6 exhibited a different degradation behavior (Fig. S5C). To further identify the structures of oligosaccharides produced by rAly-6, Avi-rAlgL, and Pae-rAlgL, six size-de ned oligosaccharide fractions, UDP2 ~ UDP7 (Avi-rAlgL and Pae-rAlgL), were isolated from the nal products, and UDP3 ~ UDP6 (rAly-6) were isolated from intermediate products of alginate digestion as described above. The 1 H-NMR chemical shifts of the protons of the ∆ unit at the nonreducing end of unsaturated alginate oligosaccharides are strongly affected by the properties of the nearest monosaccharide residues and the structure of the residue next to the ∆ unit (29,(38)(39)(40). In the case of UDP3 ~ UDP6 produced by rAly-6, the speci c signal at 5.65 ppm of H-4 of ∆G was strong, but no H-4 of ∆M was observed, indicating that rAly-6 ( Fig. 2C) is currently the rst alginate lyase that produced ∆G-terminal oligosaccharide products.
In addition, the UDP2 ~ UDP7 products from Avi-rAlgL and Pae-rAlgL were also analyzed. The observed characteristic chemical shift value indicated that almost all the UDP2 (5.65 ppm) produced by Avi-rAlgL and Pae-rAlgL was ΔM, while the UDP5 ~ UDP7 fractions produced by Pae-rAlgL and UDP4 ~ UDP7 produced by Avi-rAlgL (5.70 ppm) were ΔG ( Fig. 2A and B). The signals at 5.70 ppm and 5.60 ppm of UDP3 (Avi-rAlgL and Pae-rAlgL) and UDP4 (Pae-rAlgL), respectively, showed that UDP3 was ΔM and UDP4 was ΔG, with molar ratios of 100:104, 134:100, and 100:196. In short, the products produced by Avi-rAlgL and Pae-rAlgL during alginate degradation are similar, and the difference is that only UDP3 produced by Avi-rAlgL, not UDP3 and UDP4 from Avi-rAlgL, has two types of nonreducing ends, ∆M and ∆G. Therefore, although the characteristic succession of the nr end structure of the oligosaccharide end products of the two endo-type alginate lyases is that as the degree of poly-Merization increases, the proportion of nonreducing-end-containing ∆M of the oligosaccharide gradually decreases and the ∆G gradually increases to completely replace ∆M, these two enzymes preferred to produce large oligosaccharides (> UDP5) in which the nr end primarily contained ΔG units.
To investigate the oligosaccharide-degrading properties of the three alginate lyases, size-de ned saturated oligosaccharides (M2 ~ M5, G2 ~ G5) were used as testing substrates. Pae-rAlgL and Avi-rAlgL degraded saturated poly-M oligosaccharides (M3-M5) but did not degrade M2 or G5 (Fig. 3C and D). When the saturated M5 sugars were degraded, the nal products were mainly unsaturated UM2 (disaccharide), with a small amount of unsaturated UM3 (trisaccharide) and unsaturated UM4 (tetrasaccharide). UM2 is mainly produced when saturated M3 is degraded, while UM3 is mainly produced from M4. This indicates that the two enzymes Pae-rAlgL and Avi-rAlgL are M-speci c alginate lyases that can react with the smallest saturated oligosaccharide substrate, M3, and that the smallest product is saturated monosaccharide. In contrast, rAly-6 degrades both M and G series oligosaccharide fragments ( Fig. 3A and B), although the product absorption intensity (area integral) of M is signi cantly higher than that of G (approximately 5:1 ~ 10:1). For example, oligosaccharide yield produced by To exactly determine the substrate degradation orientation of Aly-6 and two endolytic lyases, saturated/unsaturated alginate pentasaccharides (DP5) were labeled with 2-AB at the reducing end and further digested with enzymes in a time-course experiment. A gel ltration assay showed that the digestion of 2-AB-M5/2-AB-UDP5 by Aly-6 yielded a series of 2-AB-labeled oligosaccharide products with high molecular masses at the beginning of the reaction, and then, the larger products 2-AB-UM4/2-AB-UDP4 and 2-AB-UM3/2-AB-UDP3 were gradually converted into the nal product, 2-AB-UM2/2-AB-UDP2 ( Fig. 4A and C). In particular, the nal product 2-AB-UM2 was formed in the initial stage of the reaction, and its yield gradually increased with reaction time. For the same experiment with 2-AB-G5, approximately 45% of the substrate was degraded to produce 2-AB-UG2 in equimolar amounts (Fig. 4B). The results indicated that the recombinant alginate lyase rAly-6 cleaved 2-AB-M5/2-AB-G5/2-AB-UDP5 as a monosaccharide-yielding exo-type enzyme; that is, it gradually cleaved one molecule of a saturated M/G monosaccharide and two unsaturated monosaccharide units (∆) from the nonreducing end of the substrate chain until one molecule of 2-AB-UM2/2-AB-UG2 remained.
As shown in Fig. 4D and E, 2-AB-M5 was produced by Pae-rAlgL or Avi-rAlgL, 2-AB-UM4 and 2-AB-UM3 were rst produced, and nally 2-AB-UM3 and 2-AB-UM2 were the main products. However, when 2-AB-M4 was degraded, 2-AB-UM3 was the main nal product (Fig. S6). In summary, this indicates that when Pae-rAlgL or Avi-rAlgL degrades 2-AB-M5, the main product of 2-AB-UM2 is not produced by continuous exocytosis, which is a variable substrate degradation mode with internal and external mode behavior.
To investigate the degradation patterns of the enzymes reacting against oligosaccharide substrates, each size-de ned unsaturated oligosaccharide chain, i.e., the UDP2, UDP3, UDP4, and UDP5 fractions, was reacted with rAly-6 using the same strategy as described previously. After further enzymatic reaction with rAly-6, the nal digests of each size-de ned oligosaccharide fraction were analyzed via gel ltration HPLC. The results showed that UDP3 and UDP2 were produced by degradation of UDP5 and UDP4, which contained only the ΔG-terminus, and the degradation ratio of UDP3 was approximately 5%, while that of UDP3 was 60% (mix of ΔM/ΔG-termini) (Fig. 5A). Furthermore, characteristic UDP2/M2/G2 fractions were degraded by rAly-6 ( Fig. 5B), showing that the unsaturated disaccharide fragment (ΔM or ΔG) was the smallest oligosaccharide substrate of rAly-6, wherein ΔM was more easily degraded than ΔG, but neither could be completely degraded (Table. 2). Based on the above, the recombinant alginate lyase rAly-6 is an M-preferred alginate lyase that can easily degrade alginate chains larger than trisaccharides or alginate chains containing ΔM termini but cannot easily degrade alginate trisaccharide and disaccharide chains, particularly those containing ΔG, GG or MM termini.
Aly-6 was determined to be composed of α6/α6 barrels and antiparallel β-chains as a sheet (Fig. S7A) with a crack-like catalytic cavity using Alg17c (PDB: 4NEI) as a template (Fig. S7B); this fold is signi cantly different from the barrel or jelly roll folds used by other PL family alginate lyases, with the extra peptide Gly 224 ~ Ala 248 (Fig. 6A). After the extra peptide Gly 224 ~ Ala 248 was truncated, there was little change in enzyme activity (Table. 3), indicating that the extra peptide was not a core essential structure and had a weak effect on the enzyme. Mutants H197A and W200A showed 46.09/144.53 U/mg activity on alginate (Table. 3), indicating that HNH 197 STW 200 is the catalytic motif of Aly-6, which is a representative member of the PL17 subfamily. Moreover, H426A, Y269A, and N144A were approximately completely inactivated. According to similar literature reports, the sensitivity of this mutation inactivation is determined by the characteristics of Aly-6 as an exolytic lyase.

Discussion
In this study, to elucidate and compare the core value of M-speci c and M-preferred alginate lyases in alginate degradation, oligosaccharide preparation, and associated enzymatic mechanisms, we initially expressed, puri ed, and characterized two groups of enzymes: the rst group, i.e., AlgL-Pae and AlgL-Avi, which include only one AlgL catalytic module, and the second group, i.e., Aly-6, which is composed of one AlgL-like module and one Hep_II_III-like module in sequence. As shown in Fig. 1A, B, and C and indicated by the aforementioned sequence alignment, sequence similarities only lower than 30% were shared between the two studied protein groups. Furthermore, Aly-6 of the Flammeovirga sp. strain MY04 was demonstrated to be an M-preferred bifunctional alginate lyase for the rst time, whereas AlgL-Pae and AlgL-Avi were veri ed to be M-speci c enzymes, as previously reported. Therefore, considering the substrate spectrum and enzyme activity comparisons of the truncated proteins, this study demonstrated that the AlgL-like module rather than the Hep_II_III-like module is the key catalytic element of Aly-6, although the two putative modules are integral in alginate degradation. Interestingly, the proteins of the two different groups showed similar biochemical characteristics and were similar affected by metal ions and chemical reagents in terms of their enzyme activities. Therefore, this study also indicated that the AlgL and AlgL-like modules play essential roles in determining the similar enzyme characteristics, although they differ greatly in sequence characteristics. Subsequently, the full-length Aly-6 protein shares very low sequence identities with the PL15 and PL5 families of alginate lyases, containing only a Hep_II_III-like module or an AlgL-like module. Moreover, the whole Aly-6 protein shared very low protein sequence similarities (18%) with heparinase II and heparinase III from Pedobacter heparinus.
Subsequently, we compared the catalytic characteristics of the two enzyme groups, particularly their alginate degradation patterns and oligosaccharide products. As mentioned above, the enzyme Aly-6 is an M-preferred bifunctional exo-type alginate lyase, while the enzymes AlgL-Pae and AlgL-Avi are M-speci c endo-type alginate lyases. Unsaturated disaccharides (∆M or ∆G) are the smallest oligosaccharide substrates, and the ∆M-terminal fractions are more easily degraded by Aly-6, so oligosaccharide products (UDP3-UDP6) obtained from incomplete degradation of alginate indicated that rAly-6 is the rst alginate lyase whose products contain only ΔG-terminal nonreducing ends for preparing a G-rich series of unsaturated oligosaccharide fragments with a ∆G-terminus at the nr end in incomplete degradation of alginate. Therefore, the structural features of oligosaccharide products are closely related to the enzymatic properties, such as the reported G-speci c endolytic Aly5 and G-preferred bifunctional endolytic Aly1 and Aly2 lyases from Flammeovirga sp. MY04, in which the nr end succeeds from ΔG to ΔM as the DP of oligosaccharide products increases. Interestingly, the M-speci c endolytic lyases AlgL-Pae and AlgL-Avi derived from Pseudomonas aeruginosa and Azotobacter vinelandii have conserved structural features with analogous succession rules, from ∆M to ∆G, which is the rst reported oligosaccharide product characteristic of M-speci c endolytic lyases. Furthermore, we con rmed that Aly-6 gradually removes one molecule of a saturated M/G monosaccharide or two molecules of an unsaturated monosaccharide from the nonreducing end in a monosaccharide exonuclease manner, and a 2-AB label at a saturated pentasaccharide inhibited the oligosaccharide degradation, especially that involving ΔM because of the steric hindrance caused by the 2-AB label, while AlgL-Pae and AlgL-Avi have a variable substrate degradation pattern whose key factors are the substrate molecule terminus type, degree of poly-Merization, and M/G ratio. Having a 2-AB label at the nr end had no signi cant effect.
In most cases, compared with endolytic enzymes, exolytic enzymes show very low enzymatic activity toward polysaccharides, possibly because polysaccharides are too large to e ciently bind to exolytic enzymes. The M-preferred monosaccharide-yielding exo-type alginate lyase Aly-6, with 726 U/mg activity, has signi cantly different properties from characterized alginate lyases, is stable in various tested physical and chemical environments and highly active, and has potential industrial applications.
Therefore, Aly-6 may be used for incomplete degradation of alginate to speci cally prepare a series Gblock-rich unsaturated oligosaccharide fractions containing only ∆G-terminal or unsaturated trisaccharide fractions and unsaturated monosaccharide ∆ by complete degradation. Furthermore, ∆ has been deeply and unenzymatically converted into other products that provide an unsaturated monosaccharide carbon source for the growth of microorganisms or further bioconversion into biomass, e.g., ethanol. This is the key genetic basis for the ability of Flammeovirga sp. strain MY04 to grow with the sole carbon source of alginate.
Thus, we further focused on which remaining sites of the AlgL-like modules or of the Aly-6 protein played essential roles in the catalysis of alginate. Notably, Aly-6 possesses a different half-conserved "HNHSTW" catalytic motif, in which His 197 is essential for stabilizing the negative charge on the carboxyl group in catalysis and Ser 198 represents a characteristic active site residue of the Aly-6-containing novel subclass, while Tyr 269 provides protons for the glycosidic bond.
In our study, we compared the suitability of a novel exolytic and two endolytic alginate lyases with mannuronate preference for direct preparation of oligosaccharides and determined the action modes and underlying catalytic mechanisms. These ndings are very important for the direct preparation of speci c sugar chains with designated sugar components and molecular sizes from alginate.

Conclusions
In addition to the previously reported G-speci c endolytic alginate lyase Aly5 and M-preferred bifunctional alginate lyase Aly1, the M-preferred bifunctional Aly-6 of Flammeovirga sp. strain MY04 is a novel member of the PL17 superfamily. Notably, the catalytic modules of AlgL and AlgL-like proteins play essential roles in determining the similar enzyme characteristics although they are quite different in sequence characteristics. The identi cation of M-speci c and M-preferred alginate lyases not only effectively supplements the study of alginate lyases but also provides three useful enzymes. Through studies of the enzymatic characteristics, substrate degradation patterns and oligosaccharide products, the structures of the nal products were clearly elucidated, and the oligosaccharide products were determined by both substrate selectivity and oligosaccharide degradation mode.  1H-NMR analyses of oligosaccharide products. A, 1H-NMR (600 MHz) spectra of the main nal products UDP2~UDP7 obtained from alginate digested by Pae-rAlgL. B, 1H-NMR (600 MHz) spectra of the main nal products UDP2~UDP7 obtained from alginate digested by Avi-rAlgL. C, 1H-NMR (600 MHz) spectra of the main nal products UDP3~UDP6 obtained from alginate digested by rAly-6.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. SupplementaryMaterial.docx