Trehalose Degradation by Cellvibrio japonicus Exhibits No Functional Redundancy and Is Solely Dependent on the Tre37A Enzyme

The metabolism of trehalose is becoming increasingly important due to the inclusion of this α-diglucoside in a number of foods and its prevalence in the environment. Bacteria able to utilize trehalose in the human gut possess a competitive advantage, as do saprophytic microbes in terrestrial environments. While the biochemical mechanism of trehalose degradation is well understood, what is less clear is how bacteria acquire this metabolite from the environment. The significance of this report is that by using the model saprophyte Cellvibrio japonicus, we were able to functionally characterize the two predicted trehalase enzymes that the bacterium possesses and determined that the two enzymes are not equivalent and are not functionally redundant. The results and approaches used to understand the complex physiology of α-diglucoside metabolism from this study can be applied broadly to other polysaccharide-degrading bacteria.

ABSTRACT

The α-diglucoside trehalose has historically been known as a component of the bacterial stress response, though it more recently has been studied for its relevance in human gut health and biotechnology development. The utilization of trehalose as a nutrient source by bacteria relies on carbohydrate-active enzymes, specifically those of the glycoside hydrolase family 37 (GH37), to degrade the disaccharide into substituent glucose moieties for entry into metabolism. Environmental bacteria using oligosaccharides for nutrients often possess multiple carbohydrate-active enzymes predicted to have the same biochemical activity and therefore are thought to be functionally redundant. In this study, we characterized trehalose degradation by the biotechnologically important saprophytic bacterium Cellvibrio japonicus. This bacterium possesses two predicted α-α-trehalase genes, tre37A and tre37B, and our investigation using mutational analysis found that only the former is essential for trehalose utilization by C. japonicus. Heterologous expression experiments found that only the expression of the C. japonicus tre37A gene in an Escherichia coli treA mutant strain allowed for full utilization of trehalose. Biochemical characterization of C. japonicus GH37 activity determined that the tre37A gene product is solely responsible for cleaving trehalose and is an acidic α-α-trehalase. Bioinformatic and mutational analyses indicate that Tre37A directly cleaves trehalose to glucose in the periplasm, as C. japonicus does not possess a phosphotransferase system. This study facilitates the development of a comprehensive metabolic model for α-linked disaccharides in C. japonicus and more broadly expands our understanding of the strategies that saprophytic bacteria employ to capture diverse carbohydrates from the environment.

IMPORTANCE The metabolism of trehalose is becoming increasingly important due to the inclusion of this α-diglucoside in a number of foods and its prevalence in the environment. Bacteria able to utilize trehalose in the human gut possess a competitive advantage, as do saprophytic microbes in terrestrial environments. While the biochemical mechanism of trehalose degradation is well understood, what is less clear is how bacteria acquire this metabolite from the environment. The significance of this report is that by using the model saprophyte Cellvibrio japonicus, we were able to functionally characterize the two predicted trehalase enzymes that the bacterium possesses and determined that the two enzymes are not equivalent and are not functionally redundant. The results and approaches used to understand the complex physiology of α-diglucoside metabolism from this study can be applied broadly to other polysaccharide-degrading bacteria.

INTRODUCTION

Understanding the mechanisms of competition for nutrients by bacterial communities is broadly important for the improvement of biotechnology and human health. In particular, how bacteria acquire energy from sugar oligosaccharides in a crowded community heavily influences both species composition and the metabolic potential of the surrounding environment. For example, the versatile usage of carbohydrates confers competitive advantage amid the diverse gut microflora, which partially explains some diseases and guides therapeutic approaches that exploit prebiotics and/or probiotics (1, 2). Additionally, by engineering novel carbon source usage into biotechnology platforms, such as Escherichia coli, feedstock conversion efficiency will increase while simultaneously providing increased metabolic versatility for synthetic biology applications (3). Consequently, the characterization of microbial systems that can degrade uncommon, complex, or recalcitrant carbon sources found in nature provides insights into physiological functions that are important in both ecological and industrial contexts (4).

Trehalose is an environmentally ubiquitous α-diglucoside and has an α-1,1 bond, which makes it a nonreducing sugar (5–7). This disaccharide is a component in many pharmaceuticals, foods, and cosmetics due to its properties as an odor and color stabilizer (8, 9). A recent study of Clostridium difficile infection suggested that trehalose was an important nutritional contributor to virulence, and that the increased use of this α-diglucoside in food may have unintended consequences on the human gut microbiome (1). From an ecological context, trehalose is an osmoprotectant synthesized as part of bacterial stress responses to maintain an isotonic cytoplasm during periods of dehydration or freezing (10). When synthesized by bacteria, trehalose can be derived from UDP-glucose and glucose 6-phosphate by using the enzymes trehalose 6-phosphate synthase (E.C. 2.4.1.15) and trehalose 6-phosphate phosphatase (E.C. 3.1.3.12) via a two-step process (2, 5). Alternatively, bacteria may produce trehalose in a one-step process using ADP-glucose and glucose with the enzyme trehalose synthase (E.C. 2.4.1.245) (2, 11).

Bacterial degradation of trehalose can occur by two mechanisms, with the first being simultaneous transport and cleavage of the diglucoside via the phophotransferase system (PTS) (3, 12), which results in the production of glucose and glucose-6-phosphate. The second mechanism is transport of trehalose into the periplasm and cleavage into two glucose molecules. For this second mechanism, the direct cleavage of trehalose is mediated by carbohydrate-active enzymes (CAZymes), specifically those from the glycoside hydrolase family 37 (GH37; E.C. 3.2.1.28) (13–16). This GH family is broadly present in bacteria and the CAZy database (cazy.org) has cataloged 178 bacterial genera to date that have at least one GH37 gene in their genome (13, 14). Interestingly, there are several examples of bacteria, especially those of industrial interest such as E. coli, having both the CAZyme and PTS mechanisms of trehalose degradation, which has led to some predictions that these are redundant mechanisms of α-diglucoside utilization (10, 17).

Another industrially relevant bacterium able to metabolize trehalose is Cellvibrio japonicus, which is a Gram-negative saprophyte emerging as a model organism to study the degradation of complex and recalcitrant polysaccharides. Previous analyses of the C. japonicus genome revealed a broad complement of CAZymes, including those for the catabolism of α-diglucosides (18, 19); however, it is unknown how well this bacterium is actually able to use these substrates as nutrient sources. Previous functional characterization of C. japonicus CAZyme genes yielded insights into the mechanisms of lignocellulose degradation (20–22). Interestingly, several reports indicated that C. japonicus CAZymes predicted to be biochemically redundant are not physiologically redundant in this bacterium (23–25). These studies were significant because they shed light on how to probe for the physiologically relevant CAZymes that would be most important for nutrient generation in bacterial communities and as predictors of enzymes valuable for biotechnology applications.

In this study, we evaluate the ability of C. japonicus to utilize trehalose as a sole carbon source and functionally characterize its two GH37 genes, tre37A and tre37B. Mutational analysis determined that the tre37A and tre37B gene products are not redundant and that only a Δtre37A mutant was deficient in trehalose utilization. Complementation of both the Δtre37A single mutant and the Δtre37A Δtre37B double mutant with a functional copy of tre37A confirmed an essential role in trehalose metabolism. Heterologous expression experiments found that only the C. japonicus tre37A gene functionally complemented an E. coli treA trehalase mutant. Enzymatic analyses of both trehalases determined that the tre37A gene product was capable of cleaving trehalose, while the tre37B gene product had no trehalose degradation activity. These results argued against trehalase functional redundancy in C. japonicus. Finally, a combination of bioinformatic and mutational analyses indicated that Tre37A is localized to the periplasm, which is in agreement with what is known for other bacterial trehalases. We argue that a better understanding of α-diglucoside metabolism has the potential to accelerate synthetic biology applications in the biotechnology and health industries. Specifically, this work brings us close to a comprehensive model for trehalose utilization in C. japonicus and is likely to be predictive for other Gram-negative bacteria, especially pseudomonads.

RESULTS

Tre37A is essential for growth using trehalose as a sole carbon source in C. japonicus.

Previous bioinformatic analysis of the C. japonicus genome found two predicted trehalase genes: tre37A (CJA_0284) and tre37B (CJA_1999) (19). Both genes are predicted to be transcribed alone (i.e., not in an operon) and are distant from one another on the chromosome (∼2 Mb). To determine if one or both genes were required for trehalose utilization, in-frame deletion mutant strains were constructed for each gene (Fig. S1 in the supplemental material) and grown on either glucose or trehalose medium. As expected, both mutant strains were able to grow similarly to wild type when using glucose, but the Δtre37A mutant was completely unable to grow when trehalose was the sole carbon source. The Δtre37B mutant grew similarly to wild type when using trehalose (Fig. 1).

FIG 1.

Growth analysis of C. japonicus GH37 mutant strains. Cells were grown using MOPS defined medium containing either 0.2% (w/v) glucose (A) or 0.5% (w/v) trehalose (B). Experiments were performed in biological triplicate with error bars representing standard deviation. Data regarding growth rate and maximum optical density are summarized in Table S1A.

To validate the importance of the tre37A gene product for trehalose utilization, we constructed complementation strains with a functional copy of the tre37A gene using allelic exchange. A previous study used the cel3B locus for the complementation of C. japonicus mutants (23), and here we used the same strategy. A Δtre37A Δcel3B double mutant and a Δtre37A Δtre37B Δcel3B triple mutant were constructed, as well as the Δtre37A Δcel3B::tre37A⁺ and Δtre37A Δtre37B Δcel3B::tre37A⁺ complementation strains. The Δcel3B mutation does not influence growth on trehalose, but these strains all had the diagnostic growth defect on cellobiose (Fig. S2 and Table S1B), as was shown previously (22, 24). The Δtre37A and Δtre37A Δtre37B mutants and the complementation strains were then grown using either glucose or trehalose as a sole carbon source (Fig. 2). Ectopic expression of tre37A at the cel3B locus restored growth using trehalose for strains that had Δtre37A; however, it was not to wild-type levels in terms of growth rate or maximum optical density (Table S1A). This result was unsurprising as tre37A gene expression was driven by the cel3B promoter, which is known to be constitutive (24) but might not be expressed as highly as the native tre37A promoter.

FIG 2.

Complementation analysis of C. japonicus GH37 mutants. The wild type, GH37 deletion strains, and the complementation strains were grown using MOPS minimal medium containing either 0.2% (w/v) glucose (A) or 0.5% (w/v) trehalose (B). Error bars represent standard deviation from biological triplicate experiments. Growth rates and maximum growth attained for all strains are summarized in Table S1A.

Bioinformatic analyses highlight key differences between C. japonicus Tre37A and Tre37B.

As our mutational analysis indicated that the tre37A and tre37B gene products did not have redundant function in C. japonicus, we next used a series of bioinformatics analyses to identify the differences between the two encoded proteins. We used the EMBOSS Alignment Tool (26) to compare the amino acid sequences of Tre37A and Tre37B and determined that the enzymes share 44% identity and 61% similarity. A Pfam (27) amino acid sequence alignment identified a trehalase domain within each of the C. japonicus GH37 proteins that is structurally identical to the previously characterized periplasmic trehalases from E. coli and Xanthomonas citri (28, 29). The trehalase domain identified within these four proteins was nearly identical in size (the E. coli TreA domain is 478 amino acids [aa], the X. citri TreA domain is 476 aa, the C. japonicus Tre37A is 478 aa, and the C. japonicus Tre37B is 479 aa) with similar positioning within the entire full-length protein. Furthermore, the alignment also identified that the catalytic residues previously described for E. coli TreA (30) are conserved in C. japonicus Tre37A and Tre37B (Fig. 3A). Overall, C. japonicus Tre37A shared 45% identity and 57% similarity with E. coli TreA. An amino acid alignment comparing C. japonicus Tre37B to E. coli TreA yielded similar results (43% identity and 53% similarity).

FIG 3.

Amino acid sequences and structural comparisons between full-length E. coli TreA, Xanthomonas citri TreA, and C. japonicus Tre37A and Tre37B. (A) Trehalase alignment of C. japonicus Tre37A (GenBank: ACE82971.1), Tre37B (GenBank: ACE85234.1), E. coli TreA (GenBank: WP_086585925.1, PDB: 2JG0), and X. citri TreA (GenBank: AAM35493.1) using Clustal Omega analysis. Underneath the alignment a period designates a position with a low level of similarity, a colon designates a position with a high level of similarity, and an asterisk designates identical amino acids. The catalytic residues previously identified in E. coli TreA are conserved in the X. citri and C. japonicus GH37 enzymes and are lettered in red. The hydrophilic bonding residues that are conserved in the X. citri and C. japonicus enzymes are shown in blue lettering. Among these hydrophilic bonding residues there is a single amino acid substitution in the C. japonicus Tre37B enzyme that is highlighted in red. Similarly, residues identified as being involved in substrate recognition in a previously characterized E. cloacae TreA enzyme are highlighted in yellow and are conserved in the C. japonicus trehalases. Trehalase motifs (TM1, TM2) are shown in boxes, and a single black line indicates the lid loop region. Many of the residues in the TM and lid loop regions are identical between the C. japonicus trehalases; however, there are a few variations in the Tre37B sequence. (B) Predicted tertiary structures of the C. japonicus GH37 enzymes show great similarity to the solved structure of E. coli TreA. Amino acid sequences were used to generate simulated tertiary structures for the C. japonicus and X. citri GH37 enzymes using Phyre2 analysis. Simulated structures of C. japonicus Tre37A and Tre37B had 99% coverage and 100% confidence when modeled according to E. coli TreA. The X. citri simulated structure had 89% coverage and 100% confidence compared to the E. coli TreA. The structure of the E. coli TreA trehalase domain was previously solved (PDB ID: 2JF4) and the representation shown here is adapted from the 3D viewer available on the Protein Data Bank website (http://www.rcsb.org/structure/2JF4).

In addition to the catalytic residues, three other domains have been identified as important determinants of substrate binding and enzyme activity, termed the trehalase signature motif, the hydrophilic binding pocket, and the lid loop region (15, 30–32). The residues that form the hydrophilic binding pocket in bacterial trehalases have been previously described (15, 29) and were largely conserved within the C. japonicus GH37 enzymes, with the exception of a single amino acid change at position 516 (E516N) in Tre37B. Comparison of the trehalase signature motif 1, trehalase signature motif 2, and the lid loop region from the solved structure of Enterobacter cloacae TreA (15, 30) identified a Y161F variance in the signature motif 1 region of C. japonicus Tre37B. While the lid loop region of bacterial trehalases does not possess a high level of conservation, we found A/G511K differences in the C. japonicus Tre37B compared to C. japonicus Tre37A, E. coli TreA, and X. citri TreA (15, 30). Finally, we used models generated by Phyre2 (33) to compare the predicted structures of the C. japonicus GH37 CAZymes to the experimentally determined structure of E. coli TreA. For both Tre37A and Tre37B from C. japonicus, the models indicated a structural similarity with a (α/α)₆ barrel fold that is diagnostic of GH37 family enzymes (Fig. 3B) (15, 29).

C. japonicus Tre37A enzyme is periplasmic.

A previous genomic analysis of C. japonicus suggested that Tre37A had a cleavable signal sequence and was secreted into the extracellular environment, while Tre37B had a noncleavable signal sequence and was transported to the periplasm (19). Due to the >50% similarity to E. coli TreA, which was previously shown to be periplasmic, we reevaluated the cellular location of Tre37A using a combination of bioinformatic and mutational analyses.

The presence of a signal peptide on Tre37A was confirmed with the program SignalP; however, the analysis indicated a lipoprotein signal peptide that would target the enzyme to the outer membrane due to the presence of a serine at the +2 site (34–36). This result contradicted the previous model for Tre37A, which predicted it was a secreted enzyme. Therefore, to experimentally determine the presence of secreted trehalases, we used a previously characterized C. japonicus mutant that is deficient in enzyme secretion (37). This mutant has the entire general secretory pathway (gsp) deleted and is completely unable to export CAZymes into the extracellular environment; however, periplasmic and membrane-trafficked proteins are unaffected. The Δgsp mutant grew similarly using either glucose or trehalose, suggesting that there are not any secreted C. japonicus CAZymes that have physiologically relevant trehalase activity (Fig. S3).

To confirm both the presence of periplasmic trehalase activity and that Tre37A was the essential enzyme for trehalose metabolism in C. japonicus, we tested cell-free extracts (CFEs) from wild type (WT), the Δtre37A mutant, and Δtre37B mutant strains. These CFEs were incubated with trehalose and we then analyzed the degradation products using thin-layer chromatography (TLC). Wild-type and Δtre37B CFEs produced glucose from trehalose, while no glucose was detectable from the Δtre37A CFE (Fig. 4). These in vitro results corroborate the in vivo experiments indicating that the tre37A gene product is the sole trehalase in C. japonicus. Combined with our in silico analyses, in particular the high similarity to the well characterized E. coli TreA, we conclude that Tre37A from C. japonicus is located in the periplasm.

FIG 4.

Thin layer chromatography of trehalose degradation products using C. japonicus GH37 mutant cell-free extracts (CFEs). Lanes 1 to 5 corresponded to standards and negative controls, specifically: (1) glucose standard; (2) trehalose standard; (3) glucose and trehalose mixed standard; (4) trehalose standard boiled for 15 min; and (5) reaction buffer only negative control. Lanes 6 to 11 contain CFEs incubated with 10 mM trehalose, specifically: (6) Δtre37A CFE at T₀; (7) Δtre37A CFE at T₂₄; (8) Δtre37B CFE at T₀; (9) Δtre37B CFE at T₂₄; (10) WT CFE at T₀; and (11) WT CFE at T₂₄. Lanes 12 to 14 are controls to show the amount of glucose background in the CFEs, specifically; (12) Δtre37A CFE; (13) Δtre37B CFE; and (14) WT CFE. Lanes 15–20 contain heat-denatured CFEs incubated with 10 mM trehalose in the same order as lanes 6–11. The plate was developed utilizing an eluent of 2-propanol:ethanol:water (5:3:1) for 4 h and visualized with a solution of 5% H₂SO₄:95% MeOH with subsequent charring in an oven set at 180°C for 10 min. The arrows indicate where glucose (Glc), trehalose (Tre), and the origin (Org) are in relation to one another on the silica plate.

As our analyses indicated that Tre37A is periplasmic, a mode of trehalose import is required in C. japonicus. Genomic analysis of C. japonicus indicated the presence of a predicted major facilitator family (MFS) transport gene (CJA_0281) located 2,818 bp from the tre37A locus. This gene shares less than 20% similarity with known trehalose MFS transport genes and does not reside in a characteristic MFS operon, suggesting that it is unlikely to be involved in trehalose import (38–42). An additional transporter gene predicted to encode a TonB-dependent receptor (CJA_0283) is located 154 bp from the tre37A locus. To determine if this gene played a role in trehalose transport, we generated an in-frame deletion of CJA_0281. As shown in Fig. S5 and Table S1D, the ΔCJA_0283 mutant grew similarly to WT when using trehalose, indicating that it is not involved in trehalose import. Consequently, the method of trehalose transport into the periplasm is yet undetermined.

Heterologous expression of C. japonicus tre37A is sufficient to rescue E. coli Δtre mutant phenotypes.

Mutational analysis of the predicted GH37 genes from C. japonicus established the essential role for the tre37A gene product and suggested that the tre37B gene product was not involved in the degradation of α-1,1 diglucosides. To probe if the tre37A gene product was sufficient to confer trehalose-degrading capability in other bacteria, we attempted to rescue E. coli trehalase mutants. Enzymes encoded by four genes, treA, treB, treC, and treF, mediate trehalose metabolism in E. coli. The treA and treF genes encode GH37 trehalases, with the former being located in the periplasm and the latter in the cytoplasm (17, 43–45). The treB and treC gene products encode diglucoside-specific modules of the E. coli PTS. Specifically treB encodes the trehalose-specific IIB component of the PTS (E.C. 2.7.1.201), and treC encodes trehalose-6-phosphate hydrolase (E.C. 3.2.1.93) (10, 44). Growth analysis of E. coli tre mutant strains indicated that only absence of the treA gene product resulted in the inability to use trehalose as a sole carbon source (Fig. S4). As expected, all E. coli tre mutant strains grew like wild type when glucose was used as the sole carbon source.

The full-length C. japonicus tre37A gene was cloned into pBBRMCS5 and tested for the ability to rescue the E. coli Δtre mutants when using trehalose as the carbon source. Expression of the C. japonicus tre37A gene allowed the E. coli treA mutant to grow at a much higher rate than wild type when using trehalose (Fig. 5). Strikingly, when C. japonicus tre37A was expressed in wild-type E. coli, the growth rate was enhanced as well (Table S1C). Expression of C. japonicus tre37B gene did not rescue the E. coli treA mutant and did not confer any improved growth to the wild-type E. coli.

FIG 5.

Heterologous expression of C. japonicus GH37 genes in E. coli. Strains were grown in M9 defined medium containing either 0.2% (w/v) glucose (A) or 0.5% (w/v) trehalose (B). The E. coli strain BW25113 is the parent of the treA mutant strains. All growth experiments were performed in biological triplicate with error bars representing standard deviation, but most are unable to be observed due to error being less than 10%. Table S1C summarizes the growth rates and maximum optical densities of these strains.

C. japonicus Tre37A is an acidic α-α-trehalase.

Mutational analyses performed in C. japonicus and heterologous expression in E. coli established a role for the tre37A gene product in trehalose degradation; however, the specific biochemical properties of the Tre37A and Tre37B proteins were still unknown. Using a commercial vendor to generate recombinant Tre37A and Tre37B proteins, we determined enzyme activity by measuring glucose production over time with three diglucosides (trehalose, maltose, and isomaltose). There was no significant enzymatic activity observed with Tre37B for any pH, temperature, or substrate tested. In contrast, we observed that the Tre37A enzyme exhibited a maximum specific activity for trehalose of 7.3 × 10³ ± 3.5 × 10³ μmol min⁻¹mg⁻¹, with optimal activity between 10°C and 30°C and at pH 6.0 (Fig. 6). The specific activity of C. japonicus Tre37A is similar to that of the previously characterized Enterobacter cloacae trehalase (Table 1) (29, 46), and the slightly acidic pH and moderate temperature preference (e.g., 25 to 37°C) observed is characteristic for a number of other bacterial trehalases (29, 30, 32, 47). The Tre37A enzyme was specific for trehalose, as there was no detectable activity when maltose or isomaltose was used as a substrate (data not shown). These results align with previous studies of GH37 CAZymes, which are characterized as possessing exclusively α-trehalase activity (15, 29, 30).

FIG 6.

Optimal temperature and pH for C. japonicus Tre37A activity. Purified Tre37A was incubated at temperatures ranging from 10°C to 70°C in pH 6 buffer (A) and buffers ranging from pH 3 to pH 10 at 20°C (B) to determine optimum enzymatic conditions. For each pH or temperature tested, the assays were performed in biological triplicate. The average values and standard deviations are shown.

TABLE 1.

Comparison of bacterial GH37 trehalase specific activities

Enzyme	Specific activity (μmol min−1 mg−1)	pH	Temp (°C)	Reference/source
Cellvibrio japonicus Tre37A	7.3 × 103	6	20	This study
Enterobacter cloacae TreA	1.4 × 103	5	55	(30)
Escherichia coli TreA	6.6 × 101	7.2	25a	(29)
Escherichia coli TreF	6.0 × 101	6	25a	(44)
Mycobacterium smegmatis TreA	3.7 × 101	7.1	37	(46)
Nostoc punctiforme TreH	2.5 × 10−2	7.5	37	(70)
Rhodothermus marinus TreH	1.3 × 101	6.5	88	(47)
Streptomyces hygroscopicus TreA	1.6 × 102	6.5	37	(53)
Xanthomonas citri TreA	5.5 × 101	8	25	(28)
Zunongwangia sp. TreZ	2.6 × 102	6	50	(32)

^aReported as room temperature.

DISCUSSION

Trehalose is ubiquitous in soils and a common metabolite in bacteria, fungi, insects, and plants (7). As such, for microbes that are able to cleave the α,1-1 bond of this diglucoside, there is substantial carbon to be captured. While several bacterial GH37 CAZymes have been characterized biochemically (15, 29, 30), it is still poorly understood how these CAZymes work in a physiologically relevant manner in vivo. A major complicating factor is that often there are multiple GH37 genes in bacteria, raising the question whether they are functionally redundant. To better understand the physiological importance of bacterial trehalases, we characterized the two GH37 genes from the saprophytic bacterium Cellvibrio japonicus (tre37A and tre37B) to determine essential biological functions, cellular location, and enzymatic activities. The synthesis of in vivo, in vitro, and in silico approaches provided a robust perspective on bacterial GH37 function, and can serve as a template for future studies of bacterial polysaccharide degradation.

The two predicted GH37 enzymes from C. japonicus are not functionally redundant.

Several bacterial saprophytes and human gut symbionts contain an overabundance of CAZymes, which suggests a lifestyle focused on polysaccharide utilization (48). Many of these organisms contain multiple enzymes from certain GH classes, giving the impression of extensive functional redundancy. However, for C. japonicus, previous studies have indicated that these categorically similar enzymes are in fact not redundant and perform physiologically distinct functions in the cell (23–25). This report provides another example that care must be taken when using exclusively computational methods to make assessments about bacterial physiology.

To begin our assessment of the two predicted C. japonicus GH37 enzymes, we generated a series of mutant strains and subsequently tested both growth and trehalose-degradation capability. Our mutational analysis of the two C. japonicus GH37 genes indicated that only the tre37A gene product was essential for trehalose metabolism (Fig. 1 and Fig. 2). Furthermore, TLC analysis of C. japonicus wild-type GH37 mutant CFEs resulted in trehalose hydrolysis by only wild-type and Δtre37B extracts, and not the Δtre37A extract (Fig. 4). Our interpretation of these results is that the Δtre37A mutant is completely unable to cleave trehalose, while the Δtre37B mutant still has wild-type levels of trehalase function. Consequently, only the tre37A gene product is required for trehalose degradation in C. japonicus.

To determine if C. japonicus Tre37A is solely sufficient for trehalose degradation, we heterologously expressed the C. japonicus tre37A gene in a series of E. coli tre mutants. Despite possessing four genes that play a role in trehalose degradation (10, 44, 49), only an E. coli treA mutant demonstrated trehalose-dependent growth defects (Fig. S4) (43). The full complementation of the E. coli ΔtreA stain using C. japonicus tre37A, in addition to the increased growth rate of wild-type E. coli expressing C. japonicus tre37A, suggested enhanced trehalose degradation (Fig. 5). We speculate these phenotypes are due to the comparatively high specific activity of C. japonicus Tre37A compared to the E. coli TreA (Table 1). As expected, expression of the C. japonicus tre37B gene in E. coli did not rescue any of the E. coli tre mutant growth defects. While it is possible that the lack of grow defect rescue via C. japonicus tre37B was due to complications in heterologous expression of the C. japonicus tre37B gene, we interpret the E. coli sufficiency experiments as corroborating the results observed from the C. japonicus mutant growth analysis.

Cellular location and enzyme activity of C. japonicus Tre37A is characteristic for bacterial trehalases.

The CFE data indicated that the tre37A gene product was either cytoplasmic or periplasmic, which is analogous with the trehalose metabolic systems of E. coli and X. citri (17, 28, 43). Previous annotation of the C. japonicus GH37 enzymes suggested they were located in either the periplasm or secreted. While secreted trehalase enzymes have been reported in fungi (50), in bacterial systems trehalose is often imported via a PTS (51). Similar to other pseudomonads, C. japonicus does not have a PTS (3), and the secretion of trehalases is a poor nutrient acquisition strategy given the product is glucose. Therefore, we reevaluated the cellular location of C. japonicus Tre37A.

A CAZyme secretion-deficient mutant was still able to grow like wild type on trehalose (Fig. S3), which strongly suggested that trehalases are not secreted by C. japonicus. Bioinformatic analysis using SignalP and LipoP identified lipoprotein signals for C. japonicus Tre37A (34, 35). Furthermore, amino acid sequence analysis identified a serine at the +2 site of Tre37A, and this residue is implicated in lipoprotein sorting to the outer membrane (34, 36). Given the >50% amino acid conservation between the trehalases from C. japonicus, E. coli, and X. citri (Fig. 3A), our in silico analysis suggested that Tre37A was located in the periplasm. Moreover, the restoration of trehalose degradation in an E. coli ΔtreA mutant with full-length C. japonicus tre37A (Fig. 5) is consistent with C. japonicus Tre37A being located in the periplasm.

Enzyme activity assays using purified recombinant enzyme demonstrated that Tre37A possesses high trehalase activity in slightly acidic conditions (Fig. 6). The pH and temperature optima for Tre37A are as expected for a mesophillic soil bacterium, and similar results have been reported (Table 1) (29, 30, 32, 44, 52, 53). Interestingly, the specific activity for Tre37A is higher than for several other reported bacterial GH37 enzymes (30, 32, 46, 47), which might be advantageous if used in synthetic biology or biotechnology applications. Amino acid sequence analysis indicated that several of the residues involved in the hydrophilic binding pocket, trehalase motif 1, and the lid loop deviate from those observed in E. coli TreA and X. citri TreA (Fig. 3A) and might contribute to the increased specific activity of C. japonicus Tre37A (15, 29, 30). Similar to other GH37 family CAZymes, Tre37A was highly specific for trehalose and did not cleave maltose or isomaltose (14, 15).

A robust model for α-diglucoside metabolism in C. japonicus will require identification of the disaccharide transporters and function of Tre37B.

Our results indicated that the C. japonicus Tre37A CAZyme cleaves trehalose in the periplasm; however, no trehalose-specific transporter has been identified in this bacterium. In addition to the PTS, the major facilitator family (MFS) transporters have been previously proposed as bacterial trehalose importers (38–42). A putative MFS transport gene (locus tag CJA_0281) is located 2,818 bp from the C. japonicus tre37A locus, however, the amino acid sequence alignment had less than 20% sequence similarity to MFS proteins known to interact with trehalose (38–42). Additionally, MFS genes involved in trehalose transport are typically located within an operon consisting of other genes that form an MFS complex, and we found no evidence that CJA_0281 was in an operon with other MFS genes. A second possible transporter gene adjacent to C. japonicus tre37A gene is predicted to encode a TonB-dependent receptor (locus tag CJA_0283). These receptors are the outer membrane pore of multiprotein complexes that span both membranes of Gram-negative bacteria and are involved in transporting large, complex, or rare nutrients from the environment (54). Deletion of CJA_0283 did not elicit a phenotype distinct from wild type when grown using trehalose as a sole carbon source (Fig. S5 and Table S1D). Additional bioinformatic analysis of the C. japonicus genome revealed the presence of 44 predicted TonB-dependent receptors and 14 MFS transporters; however, a comprehensive mutational analysis of these potential targets is outside the scope of the current study. Consequently, the identity of the trehalose transporter in C. japonicus is currently unknown, but ongoing work by our group is examining α-diglucoside transport with a focus on MFS- and TonB-dependent mechanisms.

Although the physiological function of C. japonicus Tre37B remains unclear, our in silico analyses suggest that this enzyme is periplasmic due to the presence of glutamate at the +2 site and arginine at the +3 site in the signal sequence. We hypothesize that the physiological function of Tre37B is unrelated to trehalose degradation, despite Tre37B having 43% amino acid identity to Tre37A. This level of homology is similar to that found in E. coli, where the TreA and TreF proteins share 49% amino acid identity but only the loss of TreA leads to growth defects during growth on trehalose (Fig. S4) (10, 44). Interestingly, E. coli TreF does have trehalase activity; however, we were unable to detect any analogous trehalase degradation from the Tre37B. This result leads us to hypothesize that Tre37B may play a role in the degradation of rare α-diglucosides, such as kojibiose (α-1,2 diglucoside) or nigerose (α-1,3 diglucoside), as Tre37B had no activity against more common α-diglucosides. Future work by our group will assess if Tre37B is able to degrade rare diglucosides; however, it will likely take transcriptome studies of diglucoside metabolism to find the conditions where the tre37B gene is highly expressed to find likely targets.

While the model of C. japonicus trehalose metabolism is not complete, the fundamental data generated from this report expands our understanding of bacterial carbohydrate degradation systems important for biomedical and biotechnology applications. For example, trehalose metabolism has been identified as a key factor for the generation of the mycoplasmic cell envelope in Mycobacterium tuberculosis (55). Additionally, probiotics have gained interest as one possible method for reducing C. difficile infections, however, most common probiotic strains do not possess trehalases, which would still leave the disaccharide available to the pathogen (56, 57). Finally, bacterial trehalose metabolism is also being studied to improve crop stress response, and predictive models for rhizosphere bacteria possessing trehalases would be useful in determining how bacterial trehalases contribute to crop health (58).

MATERIALS AND METHODS

Strains and growth conditions.

A complete list of strains used in this study is available in Table S2A. C. japonicus and E. coli strains were cultured using MOPS (morpholinepropanesulfonic acid) defined medium (TekNova) or M9 defined medium, respectively, as previously described (59–61). Carbon sources were used at the following concentrations: glucose (TekNova; 0.2% w/v); trehalose (Sigma; 0.5% w/v); or cellobiose (Acros; 0.5% w/v). When appropriate, gentamicin (7.5 μg/ml) was added to growth media. All bacterial strains were grown at 30°C. Liquid cultures were incubated with shaking (200 rpm) for aeration. Growth experiments were performed in either 96-well microtiter plates or 18-mm test tubes. Optical density at 600 nm (OD₆₀₀) was measured using an M200Pro microplate reader (TECAN), an Epoch 2 microplate spectrophotometer (BioTek), or a Spec20D+ spectrophotometer (Thermo Scientific). All growth experiments were performed in biological triplicate. Statistical analysis was performed with the GraphPad Prism 6 software. Lag phase was defined as the time required for a 25% increase in OD₆₀₀ from the initial OD₆₀₀, as was described previously (24).

Bioinformatic analyses.

Previous analysis of the C. japonicus genome identified trehalase genes tre37A (CJA_0284) and tre37B (CJA_1999) (19). The amino acid sequences encoded by these genes were analyzed for the presence and location of signal peptides using SignalP 5.0, LipoP 1.0, TOPCONS, and Busca (34, 35, 62, 63). The homology between the amino acid sequences of both C. japonicus GH37 enzymes was analyzed using the EMBOSS Needle Alignment Tool (26). Domain regions were identified using Pfam 32.0 (27). Structural models of the trehalases were generated using Phyre2 software (33). Additional sequence alignment analysis was performed utilizing the MUSCLE program (26). Default Gram-negative or bacterial settings were used for all programs when applicable.

Genetic manipulation of C. japonicus and E. coli.

In-frame deletion mutant generation of C. japonicus strains and complementation by ectopic expression via allelic exchange was accomplished as described previously (23). C. japonicus complementation strain construction was verified by PCR (Fig. S1), and the full lists of primers and plasmids used in this study are found in Tables S2B and C. The pBBRMCS5 plasmid was used for all heterologous expression experiments (64). C. japonicus GH37 genes were cloned into this vector by Gibson assembly (65). The full-length tre37A and tre37B genes, including the sequences for the predicted signal peptides, were cloned into the pBBRMCS5 vector. Heat shock or electroporation were used interchangeably for the transformation of plasmids into E. coli (66).

Thin-layer chromatography.

C. japonicus wild-type, Δtre37A, and Δtre37B strains were cultured in MOPS medium containing 0.2% (w/v) glucose. Cells were harvested through centrifugation at 27,000 × g for 20 min at 4°C. The cells were then resuspended with Complete Bacterial Protein Extraction reagent (Thermo Scientific) before being frozen at –80°C and subsequently thawed and broken via French pressure cell at 10,000 lb/in². Cell debris was removed by centrifugation under the same conditions as before and the cell extract then filtered through a 0.45-μm membrane. Protein concentration was diluted to 20 mg/ml using Complete Bacterial Protein Extraction reagent. Control samples of cell extract were heat denatured at 100°C for 15 min prior to incubation with trehalose. All reaction mixtures were incubated with 10 mM trehalose and mixed at 500 rpm in a Fisher 5350 ThermoMixer for 24 h. Reactions were stopped by boiling the samples for 5 min prior to spotting on a silica gel TLC plate (Merck). Plates were developed using 2-propanol:ethanol:water (5:3:1) for 4 h, and visualized with a solution of 5% H₂SO₄:95% MeOH followed by charring in an 180°C oven for 10 min.

Biochemical characterization.

Recombinant proteins were purchased from Allotropic Tech, LLC. (Halethorpe, Maryland), which used a Trichoplusia ni and baculovirus system (67–69). Enzymes were eluted in sodium phosphate buffer (pH 6) and used for activity studies at a purity of >80% for Tre37A and >50% for Tre37B. Activity assays used 75 ng of enzyme per reaction and were initiated by the addition of 10 mM substrate. Enzyme activity was measured over a pH range from 3 to 10. Buffers consisted of glycine-HCl, sodium acetate-acetic acid, sodium phosphate, or glycine-NaOH buffers (50 mM each) at pH values of 3, 4 to 6, 6 to 9, and 10, respectively. Optimum temperature was determined at pH 6 in sodium phosphate buffer at temperatures ranging from 10°C to 70°C. Carbohydrate specificity assays used 10 mM trehalose, 10 mM maltose, or 10 mM isomaltose in sodium phosphate buffer (pH 6.0) at 20°C. Assays using Tre37A were incubated with substrate for 15 s, while assays using Tre37B were incubated for 1 min. Trehalase enzyme reactions were stopped by incubating at 100°C for 10 min. Solutions containing trehalose without enzyme were incubated and boiled under identical conditions to reaction mixtures containing enzyme. Trehalase activity was measured by observing the quantity of glucose produced from a 50-μl reaction with a glucose oxidase/peroxidase kit used according to the manufacturer’s instructions (Sigma-Aldrich). The oxidase/peroxidase reaction was halted with the addition of 100 μl of 16 M sulfuric acid and the absorbance read at 540 nm using an M200Pro microplate reader (TECAN). Values obtained from samples containing trehalose without enzyme were background subtracted from all experimental samples. Assays were performed in biological triplicate, and statistical analysis was performed using GraphPad Prism 6 software.