The celA Gene, Encoding a Glycosyl Hydrolase Family 3 β-Glucosidase in Azospirillum irakense, Is Required for Optimal Growth on Cellobiosides

Abstract

The CelA β-glucosidase of Azospirillum irakense, belonging to glycosyl hydrolase family 3 (GHF3), preferentially hydrolyzes cellobiose and releases glucose units from the C₃, C₄, and C₅ oligosaccharides. The growth of a ΔcelA mutant on these cellobiosides was affected. In A. irakense, the GHF3 β-glucosidases appear to be functional alternatives for the GHF1 β-glucosidases in the assimilation of β-glucosides by other bacteria.

β-Glucosidases (EC 3.2.1.21) are present in eukaryotic and prokaryotic organisms and catalyze the hydrolysis of cellobiose and chemically related β-glucosides. The β-glucosidases of cellulolytic and noncellulolytic microorganisms are key enzymes for the assimilation of cellobiose, the biodegradation product of cellulose, and other plant-derived β-glucosides, such as arbutin and salicin. These enzymes also modulate the biological activities of different β-glucosides, such as antibiotics (18) or saponins (15). Numerous β-glucosidases have been identified for cellulolytic and noncellulolytic bacteria; nevertheless, their role in plant root colonization by plant growth-promoting bacteria is not understood (14, 19, 22).

Based on their amino acid sequence similarities, β-glucosidases are classified into glycosyl hydrolase families 1 and 3 (GHF1 and GHF3, respectively) (4). The GHF3 β-glucosidases exhibit a modular organization in two domains: an N-terminal catalytic A domain and a C-terminal noncatalytic, but essential, B domain. Among the bacterial enzymes, the AB organization is the most frequent, while the BA and AB′ (exhibiting a truncated C-terminal domain referred to as B′) types have also been described (7, 13, 31). The functional analysis of GHF3 β-glucosidases using genetic approaches has been documented for only a few bacteria (7, 18, 31, 33).

Azospirillum bacteria colonize the rhizosphere of several gramineous plants, such as rice, maize, sorgo, and wheat (17, 28, 29). The ability of Azospirillum irakense to grow on pectins and plant-derived β-glucosides, such as cellobiose, arbutin, and salicin, has been reported (2, 7, 11, 12). The isolation and characterization of the corresponding enzymes could reveal potential commercial applications (e.g., with respect to specific activities, substrate specificities, stability, and so forth). Two GHF3 β-glucosidases, SalA and SalB, are required for the growth of A. irakense on salicin (7). In this work, we describe the characterization of a third GHF3 β-glucosidase, CelA, for A. irakense KBC1. The encoding celA gene is required for optimal growth on cellobiose and cellulose-derived oligosaccharides, emphasizing the importance of the GHF3 enzymes in the assimilation of plant-derived β-glucosides, a feature which is usually assigned to GHF1 enzymes.

Identification of the celA locus.

Approximately 3,000 Escherichia coli clones from a genomic library of A. irakense KBC1 (2) were tested for β-glucosidase (with methylumbelliferyl-β-glucuronide [MUG] as a substrate) and endoglucanase (with carboxymethyl cellulose as a substrate) activities with overlay methods (7, 19). No clone exhibiting endoglucanase activity was obtained. The MUG-positive clones were classified into two families, based on the restriction pattern of their cosmids with EcoRI, HindIII, and BamHI. From each family, one cosmid was retained for further characterization. pFAJ0650, harboring two β-glucosidases, SalA and SalB, has been previously described (7); cosmid pFAJ0649 was analyzed in this work (Table 1). The cosmid insert DNA contained five EcoRI restriction fragments, of 8, 5.5, 4.5, 3.2, and 1.5 kb. Each of these restriction fragments was subcloned into cloning vector pBluescript SK(−). Clones showing MUG activity were obtained only for the 5.5-kb fragment (pFAJ0680). The putative gene encoding β-glucosidase activity was named celA.

TABLE 1.

Strains and plasmids

Strain or plasmid	Relevant characteristic(s)	Reference or source
Strains
A. irakense
KBC1	Wild type; LMG 10653; isolated from rice rhizosphere	11
FAJ0694	KBC1 celA::kan Kmr	This work
E. coli DH5α	F80dlacZΔM15 Δ(lacZYA-argF) recA endA hsdr supE	Gibco BRL
Plasmids
pUC19	Cloning vector; Apr	34
pBluescript SK(−)	Cloning vector; Apr	Stratagene
pSUP202	Mobilizable plasmid; suicide vector for A. irakense; Cmr Tcr Apr	24
pRK2013	Tra+ helper plasmid	9
pLAFR1	IncP1 broad-host-range plasmid;Tcr	10
pLAFR3	pLAFR1 derivative containing the pUC8 multiple cloning site	26
pHP45Ω-Km	Tcr; contains Kmr cassette	8
pFAJ0649	pLAFR1 derivative from genomic library of A. irakense containing celA	This work
pFAJ0680	pBluescript SK(−) with 5.5-kb EcoRI insert from pFAJ0649; expresses CelA	This work

Sequence analysis of the celA gene.

DNA sequencing of pUC18/19 or SK(−) subclones was carried out with an AutoRead sequencing kit (Pharmacia-LKB) and an automated sequencer (ALF; Pharmacia-KLB). Sequence data were processed and analyzed by using the PCGene software package (Intelligenetics). We sequenced, on both strands, 2.5 kb of plasmid pFAJ0680 and found an open reading frame encoding 685 amino acids with a predicted molecular mass of 73 kDa. The N-terminal region of this deduced amino acid sequence exhibited a putative peptide signal 22 residues long: MGALRLLGSISIVALTCGGIHA/STAIAQE (the slash indicates the postulated cleavage site). This prediction of the amino-terminal signal sequence was obtained with the SignalpWWW Server (http://www.cbs.dtu.dk/services/SignalP/) (16).

A comparison of the deduced sequence with sequences in data banks revealed similarity with the GHF3 β-glucosidases (Table 2). The classification of glycosyl hydrolases is available at the following address (4): http://afmb.cnrs-mrs.fr/∼pedro/CAZY/db.html. CelD of Pseudomonas fluorescens showed a particularly high degree of identity with CelA (43.6%), while a lower degree of identity was found between CelD and SalA (24.2%) or SalB (20.5%) and between CelA and SalA (23.4%) or SalB (22.9%). No match was obtained with the GHF1 β-glucosidases. The CelA sequence has all the characteristics of the AB enzymes: an N-terminal A domain with a conserved putative catalytic aspartate residue and a C-terminal B domain showing the typical conserved regions which are also present in the AB or AB′ enzymes (Table 2). A. irakense is the first bacterium for which three orthologous GHF3 β-glucosidases have been described.

TABLE 2.

Comparison of the deduced amino acid sequences of CelA with the sequences of some bacterial GHF3 β-glucosidases

β-Glucosidase	Species	Origin of the sequences (reference)	% Identity with CelAa	Active-site homologyb
CelA	A. irakense	This work	100.0	306 HKQLLTDVLKGQMGFNGFIVGDWNAHDQVPGC
CelD	P. fluorescens	20	43.6	320 HKYLLTDVLKDLSGFDGLVVGDWSGHSFIPGC
SalA	A. irakense	7	23.4	233 NDHLLNKVLKGDWGYKGWVMSDWGAVPATDFA
OleR	Streptomyces antibioticus	18	20.9	193 SDELLNKVLKEQWKFRGWVTSDWLATQSTDAL
Cbg1	Agrobacterium tumefaciens	3	21.6	201 NPWLLTKVLREEWGFDGVVMSDWFGSHSTAET
BglX	E. coli	33	27.5	246 DSWLLKDVLRDQWGFKGITVSDHGAIKELIKH
BgxA	E. chrysanthemi	31	21.5	314 NRFLLTDLLRGQYGFDGVILSDWLITNDCKGD
SalB	A. irakense	7	22.9	313 NKQMLIDLLRGTHKFKGLILSDWAITNDCNES

^aThe putative signal sequences were removed from the analyzed sequences. The percentages of identity were determined by using the ALIGN program of the GENESTREAM network server (http://vega.igh.cnrs.fr) at the Institut de Génétique Humaine Center (Montpellier, France). 

^bThe putative catalytic residue, which was experimentally identified for Aspergillus wentii (1), is underlined in all the sequences. Boldfacing indicates conserved amino acid residues. 

Biochemical characteristics of CelA.

A CelA extract was prepared from E. coli harboring plasmid pFAJ0680. Cells of overnight cultures were collected by centrifugation, washed in phosphate buffer (0.1 M K₂HPO₄-KH₂PO₄ [pH 7.0]), and lysed by sonication. The lysates were cleared by centrifugation, stored at −20°C, and analyzed for β-glucosidase activity. No β-glucosidase activity was detected in E. coli with the SK(−) plasmid. Moreover, the insertion of a Km^r cassette into the celA gene abolished the expressed β-glucosidase activity in E. coli (see below). We also verified that a unique signal was present in the zymogram of the crude CelA extract (data not shown). These features suggested that the β-glucosidase enzyme expressed from plasmid pFAJ0680 is encoded by celA.

The K_m and specific activity of the CelA extract were measured with p-nitrophenyl-β-d-glucoside (PNPG), p-nitrophenyl-β-d-xyloside (PNPX), cellobiose, gentiobiose, and salicin as substrates. The colorimetric methods used have been described previously (7). The optimal conditions retained for β-glucosidase activity were pH 7 and 45°C. The extract from E. coli/pFAJ0680 did not hydrolyze PNPX. The K_m determination revealed a greater affinity for cellobiose than for gentiobiose or the aryl-β-glucosides salicin and PNPG (Table 3). In the presence of 10 mM salicin, gentiobiose, or PNPG, the specific activity ranged from 10 to 35% of the optimal activity obtained with cellobiose. The CelA extract also exhibited the capacity to hydrolyze cellotriose, cellotetraose, and cellopentaose; nevertheless, optimal activity was obtained with cellobiose as a substrate (Table 3). This additional hydrolytic activity of CelA suggests an exo-1,4-β-glucosidase activity (EC 3.2.1.74), which is also exhibited by other GHF3 enzymes, such as CelD of P. fluorescens subsp. cellulosa (20) and CdxA of Prevotella ruminicola (32). Despite this secondary activity, the main characteristic of CelA is the hydrolysis of cellobiose, and the enzyme can therefore be referred to as a cellobiase type. In A. irakense KBC1, the previously characterized adjacent salA and salB genes were implied to be involved in salicin assimilation, and their encoding β-glucosidases exhibited more efficient hydrolysis of aryl-β-glucosides than of cellobiose (7). Moreover, the salCAB operon of A. irakense is specifically induced by aryl-β-glucosides and not by cellobiose (25). These features suggest a specialization of the three GHF3 β-glucosidases in two different assimilatory pathways for either the aryl-β-glucosides or the cellulose-derived oligosaccharides.

TABLE 3.

Substrate specificity of CelA^a

Substrates	Km (mM)	Sp actb	% Relative activity
β-Glucosides
Cellobiose	0.05	386.0	100
Gentiobiose	0.26	82.9	21
Salicin	0.24	38.8	10
PNPG	1.33	138.0	35
Cellobiosides
Cellobiose		449.2	100
Cellotriose		171.7	38
Cellotetraose		168.0	37
Cellopentaose		153.4	34

^aTwo classes of substrates, β-glucosides and cellobiosides, were used; in each set of experiments, cellobiose was used as a reference. Because of differences in experimental procedures, the absolute specific activities between the two experiments cannot be compared. 

^bMicromoles of glucose or p-nitrophenol produced per minute per milligram of protein. 

Construction and phenotype of a ΔcelA A. irakense mutant.

Plasmid pFAJ0680 was digested with BssHII, blunt ended, and ligated with the blunt-ended Km^r cassette isolated from pHP45Ω-Km (Table 1). The resulting plasmid did not express β-glucosidase activity in E. coli transformants. This constructed deletion was subcloned, as an EcoRI fragment, into vector pSUP202. Triparental mating with the pRK2013 helper plasmid allowed the transfer of the pSUP202 derivative into A. irakense KBC1. Km^r A. irakense mutants were isolated and checked by hybridization for double homologous recombination. The constructed A. irakense mutant was named FAJ0694 (Fig. 1), and its growth on MMAB minimal medium (30) supplemented with different sugars as sole carbon sources was compared with that of the wild type. Because only small quantities of C₃, C₄, and C₅ cellobiosides were available, all growth curves were monitored by use of microplates with 200 μl of minimal medium.

FIG. 1.

Construction and phenotype of the ΔcelA A. irakense mutant. E, EcoRI; Bs, BssHII; B, BamHI.

Unlike wild-type A. irakense, FAJ0694 exhibited strongly delayed growth on C₂, C₃, C₄, and C₅ cellobiosides. Wild-type A. irakense cultures reached an optical density at 595 nm of 0.5 within 14 h, while FAJ0694 required 24 h to reach the same optical density at 595 nm for each of the cellobiosides tested. The growth of FAJ0694 on the aryl-β-glucosides arbutin and salicin or on glucose or malate was not affected. This phenotype is in agreement with the substrate specificity of the CelA enzyme (Table 3). While A. irakense uses these oligosaccharides as sole sources of carbon, it does not assimilate carboxymethyl cellulose and cannot be regarded as a truly cellulolytic bacterium. However, because the constructed ΔcelA mutant reached the same cell density as the wild type on cellobiosides after 24 h of culturing and because the salCAB operon was not induced by cellobiose (25), it is likely that as-yet-unidentified alternative β-glucosidases support the growth of A. irakense on cellobiosides.

GHF3 β-glucosidases are alternatives for GHF1 β-glucosidases in A. irakense

The assimilatory pathways of β-glucosides are well characterized for E. coli (23) and Erwinia chrysanthemi (5, 6). All the β-glucosidases of these pathways belong to GHF1 (4), and their homologs have been identified for both gram-positive and gram-negative bacteria (21). A. irakense has been shown to utilize a β-glucoside assimilatory pathway involving at least three GHF3 β-glucosidases, SalA, SalB, and CelA. The GHF3 β-glucosidases seem to be functional alternatives for GHF1 β-glucosidases in the assimilation of β-glucosides, such as salicin and cellobiose, in A. irakense. Moreover, the organization and the regulation of the salCAB operon also suggest that the assimilation of aryl-β-glucosides occurs via an original mechanism which has been not described for other gram-positive or gram-negative bacteria (25).

This work also contributed to the functional analysis of one of the recently described clusters of orthologous genes (COG), COG2091, which matches the GHF3 genes (27). One goal of the COG databases is to facilitate the assignment of a function to the deduced open reading frames of a sequenced genome. Therefore, knowledge of the physiological role of several β-glucosidases of this GHF3-COG2091 cluster is necessary for further predictive investigations. In addition, because three orthologous β-glucosidases, CelA, SalA, and SalB, are present in the same bacterium, A. irakense KBC1, the phylogenetic relationships among the members of this GHF3-COG2091 cluster must be clarified.

Nucleotide sequence accession number.

The sequence obtained in this study was submitted to the GenBank database under accession number AF213463.