Cytophaga hutchinsonii gldN, Encoding a Core Component of the Type IX Secretion System, Is Essential for Ion Assimilation, Cellulose Degradation, and Cell Motility

The widespread Gram-negative bacterium Cytophaga hutchinsonii uses a novel but poorly understood strategy to utilize crystalline cellulose. Recent studies showed that a T9SS exists in C. hutchinsonii and is involved in cellulose degradation and motility. However, the main components of the C. hutchinsonii T9SS and their functions are still unclear. Our study characterized the function of GldN, which is a core component of the T9SS. GldN was proved to play vital roles in cellulose degradation and cell motility. Notably, GldN is essential for the acquisition of Ca²⁺ and Mg²⁺ ions under Ca²⁺- and Mg²⁺-deficient conditions, revealing a link between the T9SS and the metal ion transport system. The outer membrane abundance of CHU_2807, which is essential for Ca²⁺ and Mg²⁺ uptake in PY6 medium, was affected by the deletion of GldN. This study demonstrated that the C. hutchinsonii T9SS has extensive functions, including cellulose degradation, motility, and metal ion assimilation, and contributes to further understanding of the function of the T9SS in the phylum Bacteroidetes.

ABSTRACT

The type IX secretion system (T9SS), which is involved in pathogenicity, motility, and utilization of complex biopolymers, is a novel protein secretion system confined to the phylum Bacteroidetes. Cytophaga hutchinsonii, a common cellulolytic soil bacterium belonging to the phylum Bacteroidetes, can rapidly digest crystalline cellulose using a novel strategy. In this study, the deletion mutant of chu_0174 (gldN) was obtained using PY6 medium supplemented with Stanier salts. GldN was verified to be a core component of C. hutchinsonii T9SS, and is indispensable for cellulose degradation, motility, and secretion of C-terminal domain (CTD) proteins. Notably, the ΔgldN mutant showed significant growth defects in Ca²⁺- and Mg²⁺-deficient media. These growth defects could be relieved by the addition of Ca²⁺ or Mg²⁺. The intracellular concentrations of Ca²⁺ and Mg²⁺ were markedly reduced in ΔgldN. These results demonstrated that GldN is essential for the acquisition of trace amounts of Ca²⁺ and Mg²⁺, especially for Ca²⁺. Moreover, an outer membrane efflux protein, CHU_2807, which was decreased in abundance on the outer membrane of ΔgldN, is essential for normal growth in PY6 medium. The reduced intracellular accumulation of Ca²⁺ and Mg²⁺ in the Δ2807 mutant indicated that CHU_2807 is involved in the uptake of trace amounts of Ca²⁺ and Mg²⁺. This study provides insights into the role of T9SS in metal ion assimilation in C. hutchinsonii.

IMPORTANCE The widespread Gram-negative bacterium Cytophaga hutchinsonii uses a novel but poorly understood strategy to utilize crystalline cellulose. Recent studies showed that a T9SS exists in C. hutchinsonii and is involved in cellulose degradation and motility. However, the main components of the C. hutchinsonii T9SS and their functions are still unclear. Our study characterized the function of GldN, which is a core component of the T9SS. GldN was proved to play vital roles in cellulose degradation and cell motility. Notably, GldN is essential for the acquisition of Ca²⁺ and Mg²⁺ ions under Ca²⁺- and Mg²⁺-deficient conditions, revealing a link between the T9SS and the metal ion transport system. The outer membrane abundance of CHU_2807, which is essential for Ca²⁺ and Mg²⁺ uptake in PY6 medium, was affected by the deletion of GldN. This study demonstrated that the C. hutchinsonii T9SS has extensive functions, including cellulose degradation, motility, and metal ion assimilation, and contributes to further understanding of the function of the T9SS in the phylum Bacteroidetes.

INTRODUCTION

Nine types of protein secretion systems have been identified that exert crucial physiological functions in bacteria, which involve nutrition acquisition, communication with the environment, attachment to various surfaces, and pathogenicity (1, 2). The type IX secretion system (T9SS) was first discovered in Porphyromonas gingivalis in 2005 (3). It was reported in 2013 that the T9SS is extensively and exclusively distributed in the phylum Bacteroidetes (1, 4). There are N-terminal signal peptides and conserved C-terminal domains (CTDs) in proteins secreted by the T9SS. These proteins are first transported across the cytoplasmic membrane via the Sec transport system, and then are translocated across the outer membrane by the T9SS directed by the CTDs (1, 5). The T9SS has been well studied in the nonmotile bacterium P. gingivalis and in the motile bacterium Flavobacterium johnsoniae (2, 4). There are 35 CTD proteins predicted to be secreted by the P. gingivalis T9SS, including the Arg-specific cysteine proteinases (Rgp) and Lys-specific cysteine proteinase (Kgp), which participate in bacterial pathogenicity (1). F. johnsoniae T9SS is involved in delivery of SprB, RemA, ChiA, and 50 other putative CTD proteins to the cell surface or extracellular medium, which facilitates the bacterium to glide along solid surfaces and digest chitin (6).

Gld and Spr proteins essential for gliding motility are universal in the phylum Bacteroidetes (4); at least 18 of these have been identified to be components of the T9SS (1, 2, 6). Among the identified genes encoding components of the T9SS, gldK, gldL, gldM, and gldN exist consecutively in the genome of at least 37 members of the phylum Bacteroidetes and are transcribed in the same orientation; these are considered to encode the core components of the T9SS (4). It was reported that gldK, gldL, gldM, and gldN form an operon and are essential for the secretion of SprB, RemA, and ChiA in F. johnsoniae (6). P. gingivalis PorP, PorK, PorL, PorM, and PorN, orthologs of F. johnsoniae SprP, GldK, GldL, GldM, and GldN, may form a large secretion apparatus that spans the cell envelope (7, 8). Recently, Lauber et al. reported that SprA is the outer membrane translocon of the F. johnsoniae T9SS, the structure of which has been solved using cryo-electron microscopy. SprA forms an extremely large (36-strand) single-polypeptide transmembrane β barrel, and the pore, at 70 Å in diameter, is large enough to permit the passage of the folded large substrates of the T9SS (9).

C. hutchinsonii is a widespread soil bacterium that can efficiently digest cellulose in a cell contact-dependent manner, a tactic different from those of the free soluble cellulase system and multiprotein cellulosome (10, 11). The mechanism of C. hutchinsonii in degrading cellulose is still enigmatic (12–14). It is speculated that the outer membrane proteins may play significant roles in degradation of cellulose (15–17). Another feature of C. hutchinsonii is its rapid gliding ability along solid surfaces, especially on cellulose fiber, which makes it possible to completely digest cellulose (11). Recent proteomic analyses of C. hutchinsonii cultured on filter paper revealed a highly redundant system during cellulose utilization. Moreover, the study found that a number of T9SS components and T9SS substrates were abundant during growth on filter paper (18). Orthologs of all of the core genes of the T9SS were identified in the genome of C. hutchinsonii (19). We previously made efforts to delete components of C. hutchinsonii T9SS in order to study their functions. However, except for porU and sprP, other genes encoding components of the C. hutchinsonii T9SS could not be deleted. Previous work showed that deletion of chu_3237 (porU), encoding a putative peptidase that cleaves the CTDs of T9SS substrates, leads to defects in cellulose degradation, gliding motility, and secretion of CHU_0344 (a primary protein in the medium secreted by the T9SS) in C. hutchinsonii (20). Inactivation of chu_0170 (sprP) also causes defects in cellulose utilization and gliding motility (21). There are at least 147 proteins with putative CTDs that are predicted to be substrates of C. hutchinsonii T9SS (21, 22). Given the large number of proteins predicted to be secreted by the T9SS in C. hutchinsonii, the T9SS may play various functions in C. hutchinsonii. Functions of the C. hutchinsonii T9SS, in addition to cellulose utilization and gliding motility, need to be further explored.

In this study, the gene deletion mutant of chu_0174 (gldN) was obtained using a modified complex medium. The ΔgldN mutant showed a growth defect in the complex medium but could grow well in the complex medium supplemented with Ca²⁺ and Mg²⁺, indicating that GldN is essential for assimilation of trace amounts of Ca²⁺ and Mg²⁺. The outer membrane localization of CHU_2807, which participates in ion acquisition in the complex medium, was affected in the ΔgldN mutant. The effects of the deletion of gldN on protein secretion, cellulose degradation, and motility were further studied.

RESULTS

Gene target deletion of chu_0174 and chu_2610.

In order to identify and investigate the functions of the core components of the C. hutchinsonii T9SS, we analyzed the genome of C. hutchinsonii and found the conserved gene distribution of chu_0171, chu_0172, chu_0173, and chu_0174. They are transcribed in the same orientation and encode possible orthologs of GldK, GldL, GldM, and GldN of the F. johnsoniae T9SS. Also, the operon structure of chu_0171 to chu_0174 is similar to that of F. johnsoniae gldK, gldL, gldM, and gldN (see Fig. S1 in the supplemental material). Genome analysis found another gene, chu_2610, that also encodes a possible ortholog of F. johnsoniae GldN. Alignments of CHU_0174 and CHU_2610 with F. johnsoniae GldN, respectively, are shown in Fig. S2 and S3 in the supplemental material. CHU_0174 exhibits 59% identity to CHU_2610 over 253 amino acids, implying that CHU_0174 and CHU_2610 may be functionally complementary proteins.

To investigate the functions of CHU_0174 and CHU_2610 in C. hutchinsonii, they were deleted according to a method previously described (20). The deletion processes are illustrated in Fig. S4 in the supplemental material. In the deletion process of chu_2610, the transformants were abundant on PY6 plates. In contrast, no transformants of the Δ0174 mutant were obtained on PY6 plates. We tried repeatedly but failed to delete chu_0174. Traditionally, two media, PY6 (the complex medium) and Stanier medium (the minimal medium), were used to culture C. hutchinsonii. PY6 plates were often used to cultivate the transformants after electroporation. Recently, it was reported that the type VI secretion system (T6SS) is involved in zinc, iron, and manganese assimilation (23–25), so we speculated that the deletion of chu_0174 might also result in defects in ion acquisition in the complex medium with scarce metal ion nutrients. Therefore, PY6 medium supplemented with Stanier salts was used to culture C. hutchinsonii after electroporation, and chu_0174 was successfully deleted. The selected transformants were tested by PCR, and all had the expected band sizes (data not shown).

GldN is involved in secretion of CHU_0344.

CHU_0344 is one of the dominant extracellular proteins, and it was verified to be secreted by the T9SS (20, 26). Therefore, CHU_0344 can be used as a reporter protein to determine the function of the T9SS in protein secretion (20). The effect of the deletion of chu_0174 and chu_2610 on the secretion of CHU_0344 was determined. The extracellular protein profiles of the wild type (WT) and mutants are shown in Fig. 1A. As identified by liquid chromatography-tandem mass spectrometry (LC-MS/MS), CHU_0344 could be efficiently secreted to the extracellular medium in the wild type and in the Δ2610 mutant, but it was absent from the extracellular proteins extracted from the Δ0174 mutant. Moreover, abundant CHU_0344 was detected in the extracellular proteins extracted from the wild type and the Δ2610 mutant via Western blotting using an antibody against CHU_0344. CHU_0344 was absent from the extracellular proteins extracted from the Δ0174 mutant (Fig. 1B). Notably, abundant proprotein of CHU_0344 was detected in the periplasmic space of the Δ0174 mutant (Fig. 1B). These results demonstrated the secretion defect of CHU_0344 in the Δ0174 mutant and the accumulation of the proprotein of CHU_0344 in the periplasmic space of the Δ0174 mutant, suggesting that CHU_0174 is a component of the C. huchinsonii T9SS. However, CHU_2610 does not appear to participate in protein secretion. Complementation of the Δ0174 mutant with pHB0174, which carries the wild-type chu_0174 with its native promoter, restored the ability of the Δ0174 mutant to secrete CHU_0344 (Fig. 1B), confirming the role of CHU_0174 in the secretion of CHU_0344. Here, chu_0174 is designated gldN.

FIG 1.

Secretion of CHU_0344 in the wild type (WT) and mutants. (A) SDS-PAGE of the extracellular proteins of the WT and mutants cultured in PYT medium to the mid-log phase. The arrow indicates CHU_0344 identified by liquid chromatography-tandem mass spectrometry (LC-MS/MS). (B) Western blot analysis detected the extracellular and periplasmic CHU_0344 using an antibody against CHU_0344 (the Δ0344 mutant was used as a negative control). Loading samples were normalized by equal biomass. WT, wild-type strain; Δ2610, chu_2610 deletion mutant; Δ0174, chu_0174 deletion mutant; C0174, Δ0174 deletion mutant complemented with pHB0174; Δ0344, chu_0344 deletion mutant. All measurements were carried out in triplicate with the same results.

GldN is essential for cellulose utilization.

The ΔgldN mutant grew as well as the wild type in liquid Stanier medium with glucose as the sole carbon and energy source, so Stanier medium was used to examine the ability of the ΔgldN mutant to digest cellulose. The wild type could grow rapidly on plates of Stanier agar covered with Whatman filter paper as the sole carbon and energy source, while the ΔgldN mutant failed to utilize filter paper even after incubation for 15 days (Fig. 2A). Moreover, cells of the ΔgldN mutant failed to utilize 0.2% Avicel (Fig. 2B) and 0.2% regenerated amorphous cellulose (RAC) (Fig. 2C) in liquid Stanier medium. Complementation of the ΔgldN mutant with pHB0174 restored the ability of the mutant to digest filter paper.

FIG 2.

Cellulose degradation abilities of the WT and the ΔgldN mutant. (A) Filter paper degradation assay. Strains were cultured in Stanier medium to the mid-log phase, washed with Stanier without a carbon source and adjusted to an optical density (OD) of 1.0. Aliquots (3 μl) of WT, ΔgldN mutant, and C0174 cells were spotted on Whatman filter paper on Stanier agar, followed by incubation at 30°C for 15 days and recorded with a Canon camera. (B) 0.2% Avicel. (C) 0.2% regenerated amorphous cellulose (RAC) utilization ability. Mean values and standard deviations (SDs) from at least three replicates are shown. (D) Analysis of cellulase activity. The WT and the ΔgldN mutant were cultured in Stanier medium to the mid-log phase, the intact cell and intracellular cellulase activity were determined using sodium carboxymethyl cellulose (CMC-Na) as the substrate, and the reducing end concentration was measured using the 3,5-dinitrosalicylic acid procedure. The data shown are the averages and SDs from three independent experiments. *, P < 0.05. WT, wild-type strain; ΔgldN, chu_0174 deletion mutant; C0174, ΔgldN deletion mutant complemented with pHB0174. The filter paper degradation assays were performed in triplicate using three independent transformants (with the same results), and one representative result is shown.

Recent studies revealed that various cellulases are associated with cellulose degradation, and the two periplasmic endoglucanases, CHU_1280 and CHU_2103, are the key endoglucanases that are required for efficient cellulose utilization (18, 27). Because a number of cellulases contain conserved CTDs and are supposed to be secreted and anchored on the outer membrane by the T9SS (21), the effect of deletion of gldN on cellulase activity was determined. The intact cell and intracellular cellulase activity of mid-log-phase cells were detected using sodium carboxymethyl cellulose (CMC-Na) as the substrate. As shown in Fig. 2D, the intact cell cellulase activity of the ΔgldN mutant decreased by 55% compared to that of the wild type. Also, the intracellular cellulase activity of the ΔgldN mutant increased by 35% compared to that of the wild type.

Cellulose adhesion ability and motility of the ΔgldN mutant.

In the process of cellulose degradation, cells of C. hutchinsonii arrange regularly along the cellulose fiber, which is important for efficient cellulose digestion (17). To determine whether the deletion of gldN affects cellulose adhesion ability, the cellulose adhesion percentages of the wild type and the ΔgldN mutant were measured. The majority of cells of the wild type could adhere to Avicel (Fig. 3A), while cells of the ΔgldN mutant could hardly adhere to Avicel (Fig. 3A). To further investigate the cell colonization behavior of the ΔgldN mutant on cellulose fiber, cells of the wild type and the ΔgldN mutant were cultivated on filter paper and observed by scanning electron microscopy. As shown in Fig. 3B, cells of the wild type arranged regularly along the long axis of the cellulose fiber. In contrast, only rare cells of the ΔgldN mutant were detected on the cellulose fiber (Fig. 3C).

FIG 3.

Cellulose adhesion ability and motility of the ΔgldN mutant. (A) Adhesion percentage to Avicel of the WT and the ΔgldN mutant. Strains were cultured in PYT medium to the mid-log phase, and the relative adhesion rate to Avicel was measured by a turbidity-based method. The mean values and SDs from at least three replicates are shown. ***, P < 0.001. Colonization behavior of (B) WT and (C) the ΔgldN mutant on Whatman filter paper fiber. Strains were cultured in PYT medium to the mid-log phase, then washed with Stanier medium without a carbon source and adjusted to an optical density (OD) of 1.0, then equal amounts of cells of the WT and the ΔgldN mutant were spotted on Whatman filter paper, cultured for 48 h, an then detected by scanning electron microscopy. Bars, 10 μm. (D) Colony spreading of the WT and the ΔgldN mutant on PY2T hard agar cultured at 30°C for 4 days. (E) Colony spreading of the WT and the ΔgldN mutant on PY2T soft agar cultured at 30°C for 10 days. WT, wild-type strain; ΔgldN, chu_0174 deletion mutant. The colonization behavior on filter paper fiber and motility assays were performed in triplicate using three independent transformants (with the same results), and one representative result is shown.

In order to determine the effect of the deletion of gldN on motility, colony spreading of the ΔgldN mutant was studied on PY2T plates. Cells of the ΔgldN mutant formed nonspreading colonies on both PY2T hard agar (Fig. 3D) and soft agar (Fig. 3E). Individual cell motility on glass was also determined. Individual cells of the wild type exhibited vigorous gliding motility (see Movie S1 in the supplemental material), whereas a few cells of the ΔgldN mutant could adhere to glass, and the adhered cells lost the ability to glide (see Movie S2 in the supplemental material).

Secretion defect of CHU_3220 and the increased transcription level of degQ in the ΔgldN mutant.

CHU_3220 is a large cell surface protein with a putative conserved CTD, and it plays a vital role in degradation of crystalline cellulose (28). To determine whether the deletion of gldN affects the localization of CHU_3220, a Western blot analysis using an antibody against CHU_3320 was performed. As shown in Fig. 4A, CHU_3220 was detected in the outer membrane proteins extracted from the wild type but was almost undetectable in the outer membrane proteins extracted from the ΔgldN mutant. In addition, the accumulation of CHU_3220 in the periplasmic space of the ΔgldN mutant was detected (Fig. 4B). chu_0052 (degQ) encodes a homologous protein of HtrA, which is important for protein quality control in the periplasmic space (29). The accumulation of CHU_0344 and CHU_3220 in the periplasmic space of the ΔgldN mutant prompted us to determine the transcription level of degQ in the ΔgldN mutant. We found that the transcription level of degQ was increased by 5 times in the ΔgldN mutant (Fig. 4C). Moreover, Western blotting using an antibody against DegQ also indicated the significantly increased protein level of DegQ in the periplasmic space of the ΔgldN mutant (Fig. 4D).

FIG 4.

Secretion defect of CHU_3220 and the increased transcription level of chu_0052 and protein level of CHU_0052 in the ΔgldN mutant. Western blot detection of CHU_3220 in the outer membrane proteins (A) and periplasmic space proteins (B) using an antibody against CHU_3220 (the Δ3220 mutant was used as a negative control). Strains were cultured in Stanier medium to an OD of 0.6. Loading samples were normalized by equal biomass. (C) Transcription levels of chu_0052 in the wild type and the ΔgldN mutant, detected by qPCR. Strains were cultured in Stanier medium to an OD of 0.6. The mean values and SDs from at least three replicates are shown. ***, P < 0.001. (D) Western blot detection of periplasmic CHU_0052 in the wild type and the ΔgldN mutant cultured in Stanier medium to the mid-log phase; the Δ0052 mutant was used as a negative control. WT, wild-type strain; ΔgldN, chu_0174 deletion mutant; Δ3220, chu_3220 deletion mutant; Δ0052, chu_0052 deletion mutant. Loading samples were normalized by equal biomass. All measurements were carried out in triplicate with the same results.

The ΔgldN mutant was defective in ion assimilation.

Because of the different growth status of the ΔgldN mutant on plates of PY6 and PY6 supplemented with five Stanier salts, we speculated that GldN might play important roles in the assimilation of some ions in PY6 medium. To investigate the growth state of the ΔgldN mutant under different culture conditions, the growth curves were measured. The results indicated that the ΔgldN mutant showed a large growth defect in PY6 medium, with a lag phase of more than 3 days and a reduced maximum biomass compared with that of the wild type (Fig. 5A). Then, the growth status of the ΔgldN mutant in PY6 medium supplemented with CaCl₂ (0.9 mM), MgSO₄ (0.8 mM), FeCl₃ (0.07 mM), and K₂HPO₄ (4.4 mM), respectively, was determined. It was found that the addition of CaCl₂ (0.9 mM) could significantly shorten the lag phase and increase the maximum biomass of the ΔgldN mutant (Fig. 5B). The addition of MgSO₄ (0.8 mM) could also shorten the lag phase and increase the maximum biomass of the ΔgldN mutant (Fig. 5C). However, the addition of FeCl₃ (0.07 mM) or K₂HPO₄ (4.4 mM) was not beneficial for the growth of the ΔgldN mutant (see Fig. S5A and B in the supplemental material). Compared with PY6 medium cultivation, the ΔgldN mutant grew almost normally in Stanier medium (Fig. 5D). Notably, in Stanier medium without the addition of CaCl₂, the wild type could grow. However, the ΔgldN mutant could not grow at all during a cultivation of 4 days (Fig. 5E). Similarly, in Stanier medium with the concentration of MgSO₄ reduced from 0.8 mM to 0.1 mM, the wild type grew almost normally, whereas the ΔgldN mutant showed a significant growth defect with a lag phase of more than 3 days (Fig. 5F). PY6 medium supplemented with 0.9 mM CaCl₂ and 0.8 mM MgSO₄, which was designated PYT medium, was used to culture the ΔgldN mutant, and it could grow with a largely shortened lag phase and increased maximum biomass (Fig. S5C). PYT medium was used to culture the ΔgldN mutant from then on unless specially stated. Complementation of the ΔgldN mutant with pHB0174 restored the ability of the ΔgldN mutant to grow normally in PY6 medium (Fig. S5D).

FIG 5.

Growth analysis of the WT and the ΔgldN mutant in different media. (A) PY6 medium, (B) PY6 medium with CaCl₂ (0.9 mM), (C) PY6 medium with MgSO₄ (0.8 mM), (D) Stanier medium, (E) Stanier medium without the addition of CaCl₂, and (F) Stanier medium with a reduced concentration of MgSO₄ (0.1 mM). Growth was monitored by the absorbance at 600 nm. Abs, absorbance; WT, wild-type strain; ΔgldN, chu_0174 deletion mutant. The mean values and SDs from at least three replicates are shown.

The outer membrane proteins of the ΔgldN mutant significantly changed.

It has been speculated that outer membrane proteins are important for cellulose degradation and ion assimilation (15, 16, 23). To determine changes in the outer membrane proteins of the ΔgldN mutant, proteins were extracted as described in Materials and Methods, then separated by SDS-PAGE. The results showed that the outer membrane protein profile of the ΔgldN mutant significantly changed compared with that of the wild type (Fig. 6), suggesting that GldN is crucial for the cell surface attachment of the outer membrane proteins. The differential bands between the wild type and the ΔgldN mutant were identified by LC-MS/MS. As shown in Table 1, six proteins decreased in abundance or disappeared, and three proteins increased in abundance in the outer membrane proteins of the ΔgldN mutant. Among these proteins with reduced abundance on the outer membrane of the ΔgldN mutant, CHU_0007, CHU_3384, and CHU_3732 are annotated as hypothetical proteins or possible outer membrane proteins. The genes encoding these three proteins were deleted to study their functions. The results showed that deletion of chu_0007, chu_3384, and chu_3732 did not cause defects in ion acquisition, cellulose degradation, or motility (data not shown). The functions of the other differentially expressed proteins in C. hutchinsonii are under study.

FIG 6.

Outer membrane protein profiles of the WT and the ΔgldN mutant. Strains were cultured in PY6 medium to the mid-log phase, and the outer membrane proteins were extracted, then separated by SDS-PAGE. Loading samples were normalized by equal biomass. Differential protein bands between the WT and ΔgldN were marked by black arrows and identified by LC-MS/MS. WT, wild-type strain; ΔgldN, chu_0174 deletion mutant. This measurement was carried out in triplicate with the same results.

TABLE 1.

Identification of the outer membrane proteins differentially expressed between the wild type and ΔgldN

Band	Locus	MW (kDa)a	Predicted protein function	CTD	Abundanceb
1	CHU_3441	296.3	GH8 endoglucanase	TIGR04131	—
2	CHU_1075	274.2	GH8 endoglucanase	TIGR04183	—
3	CHU_1230	110.6	Zinc protease	—	↑
4	CHU_1417	65.4	Conserved hypothetical protein	—	↑
5	CHU_2807	49.9	Outer membrane efflux protein	—	↓
6	CHU_3732	44.1	Possible outer membrane protein	—	↓
7	CHU_3238	42.8	Conserved hypothetical protein	—	↑
8	CHU_3384	33.8	Hypothetical protein	—	↓
9	CHU_0007	27.4	Hypothetical protein	—	↓

^aMW, molecular mass of the primary product of translation, including predicted signal peptide.

^bAbundance change: —, missing; ↑, increased in abundance; ↓, decreased in abundance.

CHU_2807 is involved in ion assimilation.

Because of the decreased abundance of CHU_2807 on the outer membrane of the ΔgldN mutant, we investigated the function of CHU_2807 in C. hutchinsonii. CHU_2807 is annotated as an outer membrane efflux protein and contains a TolC domain that may participate in the type I secretion system (11, 30). The transcription level of chu_2807 was reduced by 66% in the ΔgldN mutant cultured in PY6 medium (see Fig. S6A in the supplemental material). Deletion of chu_2807 also caused a large growth defect in PY6 medium, and the Δ2807 mutant did not grow during a cultivation of 4 days in PY6 medium (Fig. 7A). Similarly to the ΔgldN mutant, the addition of CaCl₂ (0.9 mM) and MgSO₄ (0.8 mM), respectively, to PY6 medium could shorten the lag phase of the Δ2807 mutant (Fig. 7B and C), and the Δ2807 mutant grew almost normally in PYT medium (Fig. 7D). Unlike the ΔgldN mutant, the Δ2807 mutant could grow well in Stanier medium, Stanier medium without the addition of CaCl₂, and Stanier medium with the concentration of MgSO₄ reduced from 0.8 mM to 0.1 mM (see Fig. S6B to D in the supplemental material). Complementation of the Δ2807 mutant with pHB2807, which carries the wild-type chu_2807 with the constitutive promoter, restored the ability of the Δ2807 mutant to grow normally in PY6 medium (Fig. S6E). These results suggested that CHU_2807 is involved in the acquisition of Ca²⁺ and Mg²⁺ in the complex medium. Therefore, the intracellular concentrations of Ca²⁺ and Mg²⁺ in the Δ2807 and ΔgldN mutants were determined by inductively coupled plasma (ICP). Notably, the intracellular concentration of Ca²⁺ was reduced by 70% and 17% in the ΔgldN and Δ2807 mutants, respectively (Fig. 7E). The intracellular concentration of Mg²⁺ was reduced by 30% and 13% in the ΔgldN and Δ2807 mutants, respectively (Fig. 7F). Deletion of chu _2807 did not cause defects in cellulose utilization or soft agar spreading (see Fig. S7 in the supplemental material).

FIG 7.

CHU_2807 is involved in ion assimilation. Growth curves of the Δ2807 mutant in (A) PY6 medium, (B) PY6 medium with 0.9 mM CaCl₂, (C) PY6 medium with 0.8 mM MgSO₄, and (D) PYT medium. Intracellular concentrations of (E) Ca²⁺ and (F) Mg²⁺ were detected by inductively coupled plasma (ICP). Strains were cultured in PY6 medium to the mid-log phase, cells were washed twice with PIPES buffer and broken by ultrasonication, and the ion concentration was detected by ICP. WT, wild-type strain; ΔgldN, chu_0174 deletion mutant; Δ2807, chu_2807 deletion mutant. The mean values and SDs from at least three replicates are shown. ***, P < 0.001; Abs, absorbance.

DISCUSSION

Protein secretion systems play various physiological functions in Gram-negative bacteria, including those related to nutrient acquisition, adaptability to the environment, motility, and secretion of virulence factors (1). A T9SS was recently discovered in the phylum Bacteroidetes, where it plays important roles in pathogenicity, motility, and degradation of complex biopolymers (3, 19, 31). The widely distributed bacterium C. hutchinsonii, which is a member of the phylum Bacteroidetes, can efficiently digest cellulose using a unique strategy (11, 14). Although several proteins were identified as playing significant roles in cellulose degradation over the past decade, the cellulose degradation mechanism remains enigmatic (14–16, 28). It was reported that C. hutchinsonii has homologous proteins of PorU and SprP, which are components of the T9SS and are involved in cellulose degradation and motility (20, 21). However, other components of the C. hutchinsonii T9SS and their functions have not been reported. In this study, it was found that the deletion of the encoding gene of the T9SS GldN caused a significant growth defect in PY6 medium, which may be the reason for the difficulty in obtaining the ΔgldN mutant and the limited study on the core components of the T9SS in C. hutchinsonii. We found that the addition of Ca²⁺ or Mg²⁺ to PY6 medium could greatly relieve the growth defects of the ΔgldN mutant, and gldN could be successfully deleted using PYT medium. Furthermore, the genes encoding SprA and SprT, which are important components of the T9SS (9, 19), were also successfully deleted using PYT medium. The mutants also showed growth defects under Ca²⁺- and Mg²⁺-deficient conditions, and the deletion mutants could not be obtained using PY6 medium (unpublished data). We tried many times to delete gldK, gldL, and gldM using PYT medium, but failed, possibly because these genes are essential for survival of C. hutchinsonii in PYT medium. This study provided a modified complex medium to delete other encoding genes of the T9SS in C. hutchinsonii and other bacteria, in which the T9SS is essential for normal growth in complex medium.

As reported previously, most metal ions can pass through the outer membrane by passive diffusion through porins. However, when the concentration of extracellular metal ions is low, diffusion is not effective (32, 33). Bacteria have developed many effective means to obtain scarce metal ions in an energy-dependent manner (34). Recently, it is reported that the T6SS is involved in zinc, manganese, and iron acquisition, revealing a link between the protein secretion system and metal ion uptake (23–25). The ΔgldN mutant showed significant growth defects in media with limited Ca²⁺ and Mg²⁺ compared with the wild type, which could be alleviated by the addition of Ca²⁺ (0.9 mM) or Mg²⁺ (0.8 mM). The markedly reduced intracellular contents of Ca²⁺ and Mg²⁺ in the ΔgldN mutant further demonstrated that GldN, directly or indirectly, participates in the acquisition of trace amounts of Ca²⁺ and Mg²⁺. CHU_2807 is a conserved protein containing an N-terminal signal peptide and no transmembrane structure, and it is predicted to be located on the outer membrane. The outer membrane abundance of CHU_2807 was decreased in the ΔgldN mutant. Deletion of chu_2807 also caused significant growth defects in PY6 medium, which could be relieved by the addition of Ca²⁺ (0.9 mM) or Mg²⁺ (0.8 mM). The reduced intracellular contents of Ca²⁺ and Mg²⁺ in the Δ2807 mutant demonstrated that CHU_2807 participates in the uptake of trace amounts of Ca²⁺ and Mg²⁺ in PY6 medium. Although CHU_2807 is not a substrate protein of the T9SS, the reduced abundance of CHU_2807 on the outer membrane of the ΔgldN mutant may account for the growth defect of the ΔgldN mutant in PY6 medium to some degree. A protein BLAST search revealed that CHU_2807 contains a TolC domain. It was reported previously that TolC could form a trimeric channel structure, which is involved in type I protein secretion and export of a variety of substrates, especially multidrug substrates (30, 35). Therefore, it can be speculated that CHU_2807 may function as a transporter on the outer membrane of C. hutchinsonii. Recently, it was found that the TolC domain is necessary for importing colicin (36). However, no study has reported that the TolC domain plays a role in ion assimilation. Compared with the Δ2807 mutant, the ΔgldN mutant was more defective in Ca²⁺ and Mg²⁺ accumulation, suggesting that the T9SS may be involved in Ca²⁺ and Mg²⁺ acquisition through the secretion of relative substrates to the outer membrane or extracellular medium, which play important roles in ion chelation and assimilation. In contrast to deletion of gldN, deletion of porU and sprP did not cause significant growth defect in PY6 medium (see Fig. S8 in the supplemental material), demonstrating that different components of the C. hutchinsonii T9SS play different roles in ion assimilation. No study has reported that inactivation of T9SS components affects the growth of F. johnsoniae, P. gingivalis and other bacteria in the phylum Bacteroidetes in complex medium (3), indicating that the T9SS of C. hutchinsonii has a more extensive function.

The ΔgldN mutant completely lost the ability to utilize cellulose, including crystalline and amorphous cellulose, and cells of the ΔgldN mutant could neither arrange regularly along the cellulose fiber nor grow on filter paper. In contrast, the ΔporU and ΔsprP mutants were only partially defective in cellulose degradation. Deletion of porU did not affect the cell arrangement on the cellulose fiber, and cells of the ΔporU mutant could still grow on filter paper (20), whereas the ΔsprP mutant retains the ability to digest amorphous cellulose (21). These results demonstrated that GldN is more important than PorU and SprP in cellulose degradation. Our study showed that many proteins were decreased in abundance on the outer membrane of the ΔgldN mutant, including endoglucanases CHU_1075 and CHU_3441, and several proteins with unknown functions, such as CHU_0007, CHU_3384, and CHU_3732. Proteomic study of C. hutchinsonii found that these proteins were abundant during growth on filter paper, indicating an overlooked important role of these proteins during cellulose metabolism (18). Although single deletion of each of their encoding genes did not obviously affect cellulose degradation (data not shown), they may participate in the redundant cellulose degradation system. Our previous study showed that CHU_0344 is one of the dominant extracellular proteins, and it was verified to be secreted by the T9SS (20). CHU_3220 is a large cell surface protein with a CHU_C domain, which shares similarity to the CTD and is important for the localization of CHU_3220 (28). The secretion defects of CHU_0344 and CHU_3220 and their accumulation in the periplasmic space of the ΔgldN mutant demonstrated that GldN is important for the secretion of these CTD proteins. In C. hutchinsonii, there are at least 147 proteins with putative CTDs, including a number of endoglucanases (21, 22). These proteins are supposed to be translocated from the periplasmic space to the outer membrane by the T9SS. The reduced intact cell cellulase activity and increased intracellular cellulase activity of the ΔgldN mutant likely resulted from the secretion defect of the cellulases, which would be accumulated in the periplasmic space of the ΔgldN mutant. The aggregation of proteins is detrimental for cell viability and may cause increased stress in protein folding. HtrA proteins play crucial roles in degradation and refolding of the aberrant proteins (37, 38). A Western blot experiment demonstrated that DegQ is located in the periplasmic space in C. hutchinsonii cells (unpublished data). The accumulation of cell surface proteins in the periplasmic space of the ΔgldN mutant may be the reason for the increased transcription level of degQ in the ΔgldN mutant. Deletion of gldN resulted in complete defects in motility and decreased adhesion ability to cellulose and glass surfaces. Gliding motility has been well studied in F. johnsoniae, in which gliding relies on the rapid movement of cell surface adhesions, such as SprB, the secretion of which depends on the T9SS (39, 40). We speculated that GldN may be essential for the secretion of similar cell surface adhesions in C. hutchinsonii. Taken together, we confirmed that GldN is crucial for protein secretion by the T9SS, although the exact function of GldN in the T9SS is still unclear.

This study identified GldN as a core component of the C. hutchinsonii T9SS, where it is essential for not only cellulose degradation and motility but also for Ca²⁺ and Mg²⁺ acquisition, revealing a link between the T9SS and metal ion uptake in C. hutchinsonii. The blocked T9SS substrates in the ΔgldN mutant may be involved in these processes. Further studies will focus on specific T9SS substrates that play key roles in ion acquisition, cellulose degradation, and motility to uncover the mystery of these mechanisms in C. hutchinsonii.

MATERIALS AND METHODS

Bacterial strains, plasmids, and growth conditions.

All strains and plasmids used in this study are listed in Table 2. C. hutchinsonii ATCC 33406 was used as the wild-type strain. Strains were cultured in PYT medium (6 g/liter peptone, 0.5 g/liter yeast extract, 0.9 mM CaCl₂, 0.8 mM MgSO₄, and 4 g/liter glucose [pH 7.3]), modified from PY6 medium (6 g/liter peptone, 0.5 g/liter yeast extract, and 4 g/liter glucose [pH 7.3]), PY2T medium (2 g/liter peptone, 0.5 g/liter yeast extract, 0.9 mM CaCl₂, and 0.8 mM MgSO₄ [pH 7.3]) with 2 g/liter glucose and 5 g/liter agar in soft agar or 0.5 g/liter glucose and 12 g/liter agar in hard agar, Stanier medium (10 mM KNO₃, 4.4 mM K₂HPO₄, 0.8 mM MgSO₄, 0.07 mM FeCl₃, 0.9 mM CaCl₂, and 2 g/liter glucose [pH 7.3]) at 30°C. Liquid cultures were incubated with shaking at 160 rpm. Solid media with 10 g/liter agar (unless otherwise indicated) were cultured in a constant-temperature incubator. Escherichia coli strains were cultivated in Luria-Bertani medium at 37°C with shaking at 170 rpm or cultured in an incubator at 37°C. Antibiotics were used at the following concentrations when required: ampicillin, 100 μg/ml; erythromycin, 30 μg/ml; and cefoxitin, 15 μg/ml. The primers used in this study are listed in Table 3.

TABLE 2.

Strains and plasmids used in this study

Strain, mutant, or plasmid	Descriptiona	Reference or source
E. coli strain
DH5α	Strain used for gene cloning	TaKaRa
C. hutchinsonii strains
ATCC 33406	Wild type	ATCC
ΔgldN	chu_0174 deleted	This study
Δ2610	chu_2610 deleted	This study
C0174	Complementation of ΔgldN with pHB0174	This study
Δ2807	chu_2807 deleted	This study
Δ0344	chu_0344 deleted	20
Δ0052	chu_0052 deleted	Unpublished data
Δ0007	chu_0007 deleted	This study
Δ3384	chu_3384 deleted	This study
Δ3732	chu_3732 deleted	41
Plasmids
pCFX	Gene deletion template plasmid carrying cfxA flanked by two MCS; Apr (Cfxr); derived from pCFXSK with removal of the FRT sites	This study
pTSKQ	Gene deletion template plasmid carrying ermF flanked by two MCS; Apr (Emr); derived from pTSK with removal of the FRT sites	This study
pTKS	Gene deletion template plasmid carrying ermF flanked by two MCS; Apr (Emr)	20
pSJHS	Gene deletion template plasmid carrying ermF flanked by two MCS; Apr (Emr)	20
pSKSO8TG	Used for complementation of the mutants in C. hutchinsonii with oriC; Apr (Emr)	20
pHB0174	Plasmid constructed from pSKSO8TG used for complementation of ΔgldN; Apr (Emr)	This study
pHB2807	Plasmid constructed from pSKSO8TG used for complementation of Δ2807; Apr (Emr)	This study

^aMCS, multiple cloning site(s); FRT, FLP recombination target; Ap^r, ampicillin resistance; Cfx^r, cefoxitin resistance; Em^r, erythromycin resistance. Phenotypes in parentheses are expressed in C. hutchinsonii, and phenotypes not in parentheses are expressed in E. coli.

TABLE 3.

Sequences of primers used in this study

Primer	Sequence
0174H1F	TTGGGATCCGTGGAGGAAAGGATAAAGCTG
0174H1R	AGGGGTACCAAGACCTGCAAGAGTAGCATC
0174H2F	GGTGTCGACATAAACAACGTTCAAGACAAT
0174H2R	ACTGAGCTCGGATGATAAGACATACCCGTT
0174UF	CTCAATCATTCAAAGAAGAATTAGG
0174UR	CTTCATTTTGGAACATCATCAATGTATC
CFXR	CTACAGCTGATATATGCGCAAC
C0174F	ACAGAGCTCCACCACTGAAATGGCTAA
C0174R	TTGTCGACAACAGGGTTATCGTATCT
2610 H1F	CTCAGGATCCGAGTGGATTAAGGTATATGATC
2610 H1R	CGTTGGTACCGTTACAGACATAATGGAGTC
2610 H2F	GTGCGTCGACATGGAATACGAACATAATCTG
2610 H2R	CTGCGAGCTCAAGATTTCTTTGCTATTGC
2610UF	AATTTACACAGGGCCAGACG
2610UR	TTCCAAAAACAGGTCAACCG
EMR	TTTCTGGGAGGTTCCATTGTCCT
0007H1F	TCTGCTGCAGGATTTGTTGACTGTGATCGGTG
0007H1R	AGAAGGTACCGCTAAAAGCGAAGCAATAG
0007H2F	CGCCGTCGACTCGTTTAGGTTTCATTTTCTAG
0007H2R	AGACGAGCTCTGTATTATCAGAAGTGGGCTC
0007UF	TGAATCAGCCACTTTTGTC
0007UR	AACCGAAAGAAGCGTTATC
3384H1F	ATATGGTACCGAGTAGCTGGCGTTTCTTCATC
3384H1R	ATGTGGATCCTTCATTGCCACTACTTCATCCG
3384H2F	ATGCGTCGACTTGTTGCTAAACTTAGAGGTGG
3384H2R	ATTAGAGCTCACGTAGAAAGTAACTGGCACCC
3384UF	TTCGGATCGTAGCCATACAG
3384UR	AAACGCACCTGAGTTTGGTC
2807H1F	ATTGGATCCTCTTCTCCCTGAGTGCGATTGGT
2807H1R	AGACATATGTGTTCGGACCGTTCGCTTTGATG
2807H2F	TCAGTCGACCCTTTGACCGATGGTTTGATTAC
2807H2R	TCTGAGCTCCGTTTCTTCGCATTCTGGCTTAC
2807UF	TGGAGGACCTTGGTGACAAATAC
2807UR	TGGATAAAGATTCCTGAAGTGTCAT
P1284F	GAAGTCGACGATCGGAAATCCCTGAGGTTC
P1284R	CACAAGGAGTAAATTCTCATAAATAAAACTTTTCTTATTTC
C2807F	GAAATAAGAAAAGTTTTATTTATGAGAATTTACTCCTTGTG
C2807R	ATGGAGCTCTTATAGTTTAGTTGCAAATGGTATA
16S-F	AAGGGTGAAACTCAAAGGA
16S-R	CTCGCTGGCAACTAAAGAT
q0052F	GGTGCGTATCGGTGAGTGG
q0052R	TATGCTCCCGTAGGACTTG
q2807F	ATTCAGCCCAACAGAAACAAGAC
q2807R	TAATTGTGAGGATACGGATACGTTC

Genetic constructs.

chu_0174, chu_2610, and chu_2807 deletion mutants were constructed by the linear DNA double crossover deletion method, as previously described (20). Briefly, a 1.7-kbp fragment containing the first 267 bp of chu_0174 and the region directly upstream was amplified from the genome of C. hutchinsonii using the primers 0174HIF and 0174H1R, referred to together as 0174H1. 0174H1 was digested with BamHI and KpnI and cloned into plasmid pCFX digested with the same enzymes to generate pCFX0174H1F. A 1.9-kbp fragment that included the last 332 bp of chu_0174 and the region directly downstream was amplified from the genome of C. hutchinsonii with the primers 0174H2F and 0174H2R, referred to together as 0174H2, digested with SalI and SacI, and then was ligated into the corresponding sites of pCFX0174H1F digested with the same enzymes. The gene-targeting cassette was amplified with the primers 0174H1F and 0174H2R and purified with a Cycle Pure kit (Omega, Norcross, GA). A total of 1.5 μg PCR product was transformed into 100 μl of competent cells of C. hutchinsonii by electroporation as previously described (20), and then the transformants were cultured on PYT plates containing cefoxitin at 30°C for 10 to 15 days. The transformants were verified by PCR with two sets of primers, 0174UF/0174UR and 0174UF/CFXR, and the PCR amplicons were then verified by sequencing. The in-frame deletions of chu_2610 and chu_2807 were constructed by a similar approach using the oligonucleotides specified in Table 3. In the deletion process of chu_2610, the deletion template plasmid was pTSKQ, and the culture medium for the transformants was a PY6 plate with 30 μg/ml erythromycin. In the deletion process of chu_2807, the deletion template plasmid was pCFX, and the culture medium was a PYT plate with 15 μg/ml cefoxitin. The deletion process of chu_3732 was reported previously by Wang et al. (41). The in-frame deletion of chu_0007 and chu_3384 was constructed in a similar approach using the oligonucleotides specified in Table 3. The template plasmids pTKS and pSJHS were used to delete chu_0007 and chu_3384, respectively, and the transformants were cultured on PY6 plates with 30 μg/ml erythromycin.

Complementation of the ΔgldN and Δ2807 mutants.

The ΔgldN mutant was complemented using the replicative plasmid pSKSO8TG. The plasmid was digested with SacI and SalI to excise the green fluorescent protein (GFP) gene. A fragment spanning chu_0174, containing 172 bp upstream of the start codon and 24 bp downstream of the stop codon, was amplified with the primers C0174F and C0174R. The fragment was digested with SacI and SalI and ligated into the linearized pSKSO8TG plasmid to generate pHB0174. Plasmid pHB0174 was electroporated into the ΔgldN mutant, and the transformants were selected by erythromycin resistance in PY6 plates. C0174 refers to the complemented strain of the ΔgldN mutant with pHB0174. Complementation of the Δ2807 mutant was constructed using the same method and the same plasmid. The promoter of chu_1284 was amplified with the primers P1284F and P1284R, and the PCR product was fused to the chu_2807 gene, which was amplified with the primers C2807F and C2807R. The fusion product was digest with SacI and SalI and ligated into the linearized pSKSO8TG plasmid to generate pHB2807.

Growth analysis in different cultures.

C. hutchinsonii strains were precultured in PYT medium to the mid-exponential phase, and cells were harvested and washed with PY6 medium containing no carbon source, then adjusted to the same cell density for inoculation. When glucose was used as the sole carbon source, the growth status was monitored with the Bioscreen C analyzer (Oy Growth Curves Ab Ltd., Finland). Strains were inoculated into 200 μl of Stanier; PY6; PY6 with 4.4 mM K₂HPO₄, 0.8 mM MgSO₄, 0.07 mM FeCl₃, 0.9 mM CaCl₂, respectively; and PYT media with an inoculum concentration of 3% (vol/vol) in sample plate. The plate was incubated at 30°C with shaking at medium speed, and the growth status was monitored by the absorbance at 600 nm per 3 h. When 0.2% Avicel and 0.2% regenerated amorphous cellulose (RAC) were used as the sole carbon source, the growth status was monitored by measurement of the cellular protein concentration at set intervals as previously described (42). The protein concentration was quantified as described by Bradford (43). For the growth analysis on filter paper, equivalent amounts of cells of the WT and mutants were spotted on Stanier plates covered with Whatman filter paper. The plates were incubated at 30°C, and the degradation of filter papers was recorded with a Canon camera at set intervals.

Measurement of colony spreading and microscopic observation of individual cell motility.

Colony spreading on soft and hard agar were observed as previously described (44). Briefly, cells of the WT and mutants were cultured in PYT medium to the mid-exponential phase, harvested and washed with PY2T medium without any carbon sources, and then resuspended with the same medium to an optical density (OD) of 1.0. Equivalent amounts of cells of the WT and mutants were spotted on PY2T soft agar and hard agar, followed by incubation at 30°C for about 4 to 10 days. Soft agar spreading was recorded with a Canon camera, and hard agar spreading was observed and recorded with an IX51 phase-contrast microscope (Olympus, Tokyo, Japan). Individual cell motility over a glass surface was observed as described by Ji et al. (45). Briefly, WT and ΔgldN mutant cells were cultured on PY2T agar with 2.0 g/liter glucose at 30°C for 4 days. Tunnel slides were prepared as previously described (46), using double-sided tape to hold a glass cover slip over a glass slide. Cells were suspended in TC buffer (10 mM Tris and 8 mM CaCl2 [pH 7.3]), then introduced into tunnel slides and incubated for 5 min. The motility of individual cells that adhered to the glass cover slip was observed and recorded using an Olympus phase-contrast microscope with a heated stage at 30°C.

Cellulase activity assay.

Cells incubated in Stanier medium to the mid-exponential phase were collected by centrifugation at 5,100 × g for 5 min. Sodium carboxymethyl cellulose (CMC-Na) was used as the substrate to determine endoglucanase activity. For intact cell samples, cell pellets were washed with Na₂HPO₄-KH₂PO₄ buffer (100 mM, pH 6.8) and resuspended in the same buffer. For cell extract samples, cell pellets were washed with Na₂HPO₄-KH₂PO₄ buffer, then resuspended with Na₂HPO₄-KH₂PO₄ buffer containing 2% (vol/vol) Triton X-100 for about 4 h at 4°C, 0.5 mg/ml phenylmethylsulfonyl fluoride (PMSF) was added to deactivate the proteases. To measure cellulase activity, a mixture of 500 μl of resuspended cell samples and 500 μl of 1% (wt/vol) sodium carboxymethyl cellulose (CMC-Na) in distilled water was incubated for 30 min at 30°C. The reducing ends were measured using 3,5-dinitrosalicylic acid as previously described (44, 47). The protein concentration was quantified as described by Bradford (43). The intracellular cellulase activity was calculated by taking the cell extract cellulase activity and subtracting the intact cell cellulase activity. All measurements were carried out in triplicate.

Assay of bacterial adhesion to cellulose.

Strains of the WT and the ΔgldN mutant were cultured in PYT medium to the mid-exponential phase, and cells were harvested by centrifugation at 5,100 × g for 5 min. Relative bacterial adhesion to Avicel PH-101 was measured by the turbidity-based method described by Ji et al. (47). All measurements were carried out in triplicate.

Cells arrangement on cellulose fiber detected by scanning electron microscopy.

Samples for scanning electron microscopy were prepared as described by Xie et al. (11), with some modifications. Briefly, cells were cultured on Whatman no.1 fiber paper as the sole carbon and energy source on Stanier agar for 48 h. Samples were fixed with 2.5% glutaraldehyde in phosphate-buffered saline (PBS) buffer (pH 7.3）for 12 h at 4°C. Fixed cells were washed twice in PBS buffer, dehydrated with ethanol, and dried in a glass desiccator. Samples were processed according to a standard procedure and viewed by a scanning electron microscope with a JEOL JSM-7600F field emission scanning electron microscope.

Cell fractionation and Western blot analysis.

Outer membrane proteins were extracted as previously described by Wang et al. (28). Briefly, strains were cultivated in Stanier or PY6 medium to an OD of 0.6, and cells of equal culture volume were harvested by centrifugation at 5,000 × g for 10 min at 4°C. The pellets were washed with 50 mM piperazine-N,N′-bis(2-ethanesulfonic acid) (PIPES) buffer (pH 6.8), resuspended in PIPES buffer with 0.5 M NaCl, and then incubated at 4°C for 30 min with shaking at 150 rpm. Cells were removed by centrifugation at 12,000 × g for 10 min at 4°C, and the supernatant containing the buffer-washed proteins was ultracentrifuged (100,000 × g, 30 min, 4°C). The sediment was resuspended in PIPES buffer as outer membrane proteins. The extracellular proteins were isolated as described by Wang et al. (20), and the periplasmic proteins were extracted as described by Soares et al. (48). Strains were cultured in PYT medium to an OD of 0.6, and samples of equal culture volume were harvested by centrifugation at 5,000 × g for 10 min at 4°C; the supernatants were used to isolate extracellular proteins, and the cell pellets were used to extract periplasmic proteins. The protein concentrations were measured by the Bradford method according to the manufacture’s instruction. Outer membrane proteins with equal biomass were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) stained by Coomassie brilliant blue R-250. Differential bands between the WT and the ΔgldN mutant were excised, digested with trypsin, and identified by LC-MS/MS. The Western blot procedure was performed as described by Wang et al. (20). Extracellular proteins, outer membrane proteins, and periplasmic space proteins with equal biomasses were separated by SDS-PAGE, and then transferred onto a polyvinylidene difluoride (PVDF) membrane (Immobilon-P; Millipore, MA) using a semidry electrophoretic transfer cell (Bio-Rad, CA, USA) according to the manufacture’s instruction. For detection of CHU_0344 and CHU_3220, antibodies were the same as previously reported (20, 28). The antibody against CHU_0052 (DegQ) was prepared by heterologous expression of CHU_0052 in Escherichia coli (unpublished data).

Real-time quantitative PCR analysis.

The WT and the ΔgldN mutant were cultured in Stanier medium or PY6 medium to the mid-log phase, and cell pellets of 4 ml of the culture were collected by centrifugation at 5,000 × g for 5 min at 4°C. RNA extraction and quantitative PCR (qPCR) were performed as previously described (28, 44). qPCR was carried out on a LightCycler 480 System with SYBR green supermix (TaKaRa, Dalian, China). Data analysis was carried out by the relative quantitation/comparative threshold cycle (ΔΔC_T) method (49) and was normalized to an endogenous control (16S rRNA gene). Three biological repeats were set for all assays.

Detection of intracellular ion content.

The method for detection of intracellular magnesium and calcium ions was modified from Si et al. (23). Strains of the wild type and the ΔgldN and Δ2807 mutants were cultured in 150 ml PY6 medium to the mid-log phase, cells were collected and washed twice by 10 ml 50 mM PIPES buffer, the cell pellets were resuspended in 3 ml PIPES buffer, and cells were broken by ultrasonication. The total protein concentration was detected by the Bradford method following the manufacturer’s instructions. Each sample was diluted 25-fold in 1% (vol/vol) superior grade nitric acid to a total volume of 25 ml. The concentrations of Mg²⁺ and Ca²⁺ were detected by an inductively coupled plasma analyzer (ICP-7200; Thermo Scientific). The results were corrected using the appropriate buffers for reference and dilution factors. Triplicate cultures of each strain were analyzed during a single experiment, and the experiment was repeated at least three times.

Statistical analysis.

Statistical analysis was performed using Student’s t test. Three biological replicates were undertaken for each analysis. Reported results and errors are means and standard deviations, respectively, for these replicates.