Molecular basis of cellulose biosynthesis disappearance in submerged culture of Acetobacter xylinum

Acetobacter xylinum strains are known as very efficient producers of bacterial cellulose which, due to its unique properties, has great application potential. One of the most important problems faced during cellulose synthesis by these bacteria is generation of cellulose non-producing cells, which can appear under submerged culture conditions. The reasons of this remain unknow. These studies have been undertaken to compare at the molecular level wild-type, cellulose producing (Cel+) A. xylinum strains with Cel– forms of cellulose-negative phenotype. Comparison of protein profiles of both forms of A. xylinum by 2D electrophoresis allowed for the isolation of proteins which were produced exclusively by either Cel+ or Cel– cells. Sequences of peptides derived from these proteins were aligned with those of proteins deposited in databases. This analysis revealed that Cel– cells lacked two enzymes: phosphoglucomutase and glucose-1-phosphate uridylyltransferase, which generates UDP-glucose being the substrate for cellulose synthase. DNA was analyzed by ligation-mediated PCR carried out at low denaturation temperature (PCRMP). Two DNA fragments of different thermal stability (218 and 217 bp) were obtained from the DNA of Cel+ and Cel– forms, respectively. The only difference between these Cel– and Cel+ DNA fragments is deletion of one T residue. Alignment of those two sequences with those deposited in the GenBank database revealed that similar fragments are present in the genomes of some bacterial cellulose producers and are located downstream from open reading frames (ORF) encoding phosphoglucomutase. The meaning of this observation is discussed.


2005
A. Krystynowicz and others lose on the surface of liquid and solid culture media.Gelatinuous, leather-like mats formed on the surface of liquid culture media under stationary culture conditions contain bacterial cells entrapped in a network of cellulose fibers.Under agitated culture conditions, deposition of pellicle is disrupted and cellulose forms irregular granules stellate and fibrous strands (Bielecki et al., 2001).On agar media A. xylinum forms colonies with equal or undulate edges, transparent or white, smooth or rough, flat or convex.
A. xylinum strains are prone to spontaneous mutations yielding cellulose non-producing cells, which is one of the major problems facing commercial exploitation of bacterial cellulose biosynthesis.The appearance of Cel -forms in agitated cultures was first described by Schramm and Hestrin (1954) who isolated three different types of A. xylinum cells distinguished by morphology of colonies and efficiency of cellulose biosynthesis : -Type I: wild-type, cellulose-producing (Cel + ) cells; -Type II: cellulose-nonproducing forms (Cel -) capable of reverting through passages; -Type III: non-reverting cellulose-nonproducing forms (Cel -).The morphology of Cel + and Cel -colonies is different.Colonies of Cel -cells are rough, flat, slimy with undulate edges, in contrast to those of the Cel + type, which are spherical with smooth edges, gelatinous and convex.
The frequency of Cel + to Cel -conversion depends on culture conditions and their changes.Cellulose-producing cells dominate in stationary cultures and produce on the surface of culture medium a thick cellulose mat, called a pellicle, in which the embedded bacterial cells have contact with the oxygen-rich liquid/air interface.Sufficient and uniform aeration of liquid culture media under agitated culture conditions favors spontaneous appearance of Cel -cells, which become dominating in the population.Thus aeration of culture media is believed to be a factor against discriminating Cel + A. xylinum cells (Leisinger et al., 1966).
Studies on the effect of culture medium composition on the efficiency of cellulose synthesis under agitated culture conditions revealed that the highest yields of this biopolymer were achieved in media supplemented with ethanol (2%, v/v).According to Son et al. (2001) this phenomenon resulted not from a change in metabolism but from the lack of conversion of Cel + cells to Cel -ones.Hence, the spontaneous change of A. xylinum Cel + cells to the cellulosenegative phenotype in agitated cultures can be decreased by addition of ethanol to culture media (Son et al., 2001;Krystynowicz et al., 2002).
Glucose-1-phosphate uridylyltransferase (also known as UDPG pyrophosphorylase) is thought to play an important role in cellulose synthesis because the Cel -forms of A. xylinum lack this enzyme (Valla et al., 1989).However, the genetics and regulation of transcription of this enzyme in bacteria producing cellulose or related polysaccharides have not been explained.
According to Valla et al. (1987), also plasmids can be involved in cellulose biosynthesis.Based on electrophoretic separations of plasmid DNA isolated from A. xylinum it was found that nine of thirteen isolated mutants displayed changes in plasmid DNA profile as compared to the wild-type strain.
It was also reported that insertion elements could be involved in polysaccharide synthesis in Pseudomonas atlantica (Bartle� & Silverman, 1989) and Xanthomonas campestris (Ho�e et al., 1990).An unstable polysaccharide production by Zoogloea ramigera resulted from rearrangement of its DNA (Easson et al., 1987).The presence of insertion element IS1031 (950 bp) was detected in A. xylinum ATCC 23769.Cel -mutants of this strain possessed two or more IS1031 elements, and furthermore, their DNA was rearranged within the IS elements, as compared to cellulose-producing wild-type strain (Coucheron, 1991).Therefore, the loss of capability of some A. xylinum strains to produce cellulose is supposed to result from dislocation of insertion elements and inactivation of gene(s) responsible for cellulose synthesis.
The presented studies have been undertaken to compare wild A. xylinum E 25 strain and Cel -forms possessing cellulose-negative phenotype at the molecular level.Genome DNA was isolated from Cel + and Cel -A.xylinum E 25 cells, and compared by PCR-MP (PCR-melting profiles) analysis.Independently, protein profiles of both types of cells were compared by means of two-dimensional electrophoresis.Spots corresponding to the proteins inherent for only one type of the cells were excised from gels, digested and sequenced.The peptide sequences determined in this way were subsequently aligned with protein sequences deposited in databases.Two enzymes involved in the synthesis of various glucose-containing polysaccharides, namely phosphoglucomutase and glucose-1-phosphate uridylyltransferase, were identified as lacking in the Cel -phenotype.Possible reasons of this phenomenon are discussed.

Microorganism.
A. xylinum E 25 strain from the pure culture collection of the Institute of Technical Biochemistry of the Technical University of Lodz was used for the studies.It was maintained on agar slants at 4 o C.
Culture conditions.Inoculum was prepared by transferring a single A. xylinum Cel + or Cel -colony from SH agar medium into a 50-ml Erlenmeyer flask containing 10 ml of liquid SH medium.This cell suspension (10 ml) was added into a 500 ml Erlenmeyer flask containing 100 ml of a fresh SH medium.Agitated cultures were incubated at 30 o C on a rotary shaker at 90 r.p.m.
Selection of non-reverting mutant.A non-reverting form of A. xylinum E 25 was obtained by reiterated passages under agitated culture conditions.The cultures were carried out for 48 h at 30 o C on a rotary shaker at 90 r.p.m.This process was repeated three times (3 passages).Cell suspension from each passage was spread on SH agar medium and incubated for 6 days at 30 o C. Colonies with morphology characteristic for Cel -forms were suspended in 5 ml of liquid SH medium and incubated for 3 days at 30 o C to test if the revertants were able to form the cellulose pellicle.Non-reverting forms were isolated from cellulose-free culture broths and used in further studies.
Sample preparation for 2D electrophoresis.A single colony of Cel + or Cel -A.xylinum E 25 was suspended in 5 ml of SH medium and incubated for 48 h.From this culture, a 0.5 ml inoculum was added do 10 ml of SH medium (in a 50 ml Erlenmayer flask) and incubated for 48 h at 30 o C. Bacterial cells were harvested by centrifugation (9000 r.p.m., 20 min, 4 o C), washed with 40 mM Tris/HCl and disrupted by sonication (3 × 2 min).Proteins were separated using 2-D Clean-Up Kit (Amersham Biosciences) and suspended in a solution containing: 7 M urea, 2 M thiourea, 4% Chaps, 1% DTT and 0.8% ampholites pH 3-10, according to Westemeier et al. (2002).
Two-dimensional gel electrophoresis.Twodimensional electrophoresis was carried out in an Ettan apparatus (Amersham Biosciences).In the first dimension the separation of proteins was based on their isoelectric points (Immobiline DryStrip gel pH 3-10), and in the second dimension the proteins were separated by SDS/PAGE gel (12.5%, 200 × 260 mm) according to the protocol of an Amersham Pharmacia Biotech Technical Manual (1999).The analyzed samples contained 75 µg of proteins which were visualized by Silver Staining Kit (Sigma).
DNA isolation.DNA was isolated from Cel + and Cel -cells of A. xylinum E 25 according to Maniatis et al. (1989).Bacterial cells were pelleted by centrifugation (1300 × g, 15 min), washed with 10 ml TGE (25 mM Tris/HCl, 50 mM glucose, 10 mM EDTA, pH 8) and again centrifuged (1300 × g, 15 min).The pellet was suspended in 10 ml TGE supplemented with lysozyme (10 mg/ml) and incubated at room temperature for 10 min.Next, proteinase K (100 µg/ml) and 1% SDS were added and the mixture was kept at 37 o C for 30 min.Then, 2 ml of 5 M aqueous NaCl solution and 1.5 ml of 10% CTAB in 0.7 M NaCl were added to the mixture, which was incubated at 65 o C for 20 min.A�er mixing with equal volume of phenol/chloroform, the mixture was centrifuged (3600 × g, 10 min), extracted with a double volume of chloroform/isoamyl alcohol (24:1, v/v), carefully mixed with 2.5 volumes of cold (-20 o C) ethanol (96%), and le� for 15 min.DNA was collected by centrifugation, the pellet washed 2 times with 70% ethanol, centrifuged at 3600 × g for 10 min and dried.
Then, PCR reaction was carried out in an MJ Research PTC200 thermocycler at denaturation temperatures (T d ) 80, 81 and 82 o C. The reaction mixture contained 50 pmol of POWBAGCT primer 5'-CT-CACTCTCACCAACGTCGAAAGCTT-3', 100 µmol each dNTPs and 1 µl of ligation mixture in 50 µl of PCR buffer.Samples (8 µl out of 50 µl) were loaded on 6% polyacrylamide gel (bisacrylamide/acrylamide = 1:60) with TAE buffer.The gel was stained with ethidine bromide and visualized in UV at 302 nm, according to the protocol of Masny and Płucienniczak (2003).DNA fragments of different thermal stability were isolated from the gel (Dybczyński & Plucienniczak, 1988), digested with HindIII restriction endonuclease, ligated with the pBluScriptSK -plasmid (Stratagene) and cloned in Escherichia coli, strain NM522.The nucleotide sequences of the inserts were A. Krystynowicz and others determined in the Institute of Biochemistry and Biophysics (PAS, Warsaw, Poland).
Protein identification.Proteins were identified by peptide mass fingerprinting using MALDI-TOF mass spectrometry and by partial peptide sequencing with ion trap electrospray mass spectrometry.Trypsin autodigestion peaks were exluded from the database searching.Molecular mass data were obtained from EXPASY (h�p://www.matrixscience.com/).Monoisotopic peptide masses were used to search protein databases to match and subsequently identify individual protein spots.

PCR-MP
Figure 1 presents a profile of DNA fragments liberated by HindIII restrictase from the genomes under investigation and amplified at low T d 's according to the procedure described by Masny and Płucienniczak (2003).It appeared that in the case of the cellulose producer strain (Cel + form) one of the DNA fragments migrating just above 242 bp was less stable than the corresponding fragment from Cel -.The DNA fragment from Cel + cells was amplified at the denaturation temperature of 81 o C, while its counterpart from Cel -became visible when the T d was 1 o C higher.The less stable fragment (from Cel + cells) has a stretch of four T-residues (positions 76-79, Fig. 2), while its counterpart from Cel -contains at the same positions three T residues.One can suggest that the deletion of a single T residue in the Cel -fragment is responsible for its increased thermal stability as compared to that of the corresponding Cel + DNA fragment.

Two-dimensional electrophoresis
Figures 3A and 3B present intracellular proteins (75 µg per gel) isolated from the wild A. xylinum E 25 (Cel + ) strain and its Cel -counterpart.Electrophoretic separation was carried out in the pH range from 4 to 7 in 12% polyacrylamide gel.Proteins produced only by Cel + cells and not synthesized by the Cel -mutant, as well as proteins found only in Cel -mutant and lacking in wild-type strain are marked on the gels.

Protein identification
Proteins unique to Cel + or Cel -cells were excised from the gel and sequenced.The results of the analysis of peptides derived from intracellular A. xylinum E 25 (Cel + ) proteins supposedly involved in cellulose biosynthesis and missing in cells of the cellulose-nonproducing mutant are collected in Table 1.

DISCUSSION
Cellulose non-producing forms of A. xylinum were for the first time described in 1954 (Schramm & Hestrin, 1954).An influence of culture conditions on the frequency of Cel -phenotypes (Leisinger et al., 1966;Son et al., 2001;Krystynowicz et al., 2002) and activities of enzymes involved in the synthesis of this polysaccharide (Valla et al., 1989) have been reported by several groups.An influence of plasmid DNA and insertion elements in the A. xylinum genome on the efficiency of cellulose synthesis was also described (Valla et al., 1987;Coucheron, 1991).
However, the reasons of the appearance of cellulose non-producing A. xylinum forms in agitated cultures remain obscure.One can suppose that this effect results from different expression of genes encoding enzymes responsible for glucose metabolism and/or cellulose synthesis.Wong et al. (1990) identified in A. xylinum the bcs operon encoding four proteins essential for bacterial cellulose synthesis.The first gene in this operon, bcsA, encodes the regulatory subunit of cellulose synthase (CS), which binds c-di-GMP acting as bacterial second messanger and activator in the cellulose synthesis process (Wong et al., 1990).The second gene, bcsB, encodes the catalytic subunit of CS, which binds the substrate, UDP- glucose, and presumably catalyzes the polymerization of 1,4-β-�-glucan.Two other genes belonging to this operon, dgc and pdeA, encode diguanylate cyclase and phosphodiesterase A, which are respon-sible for the synthesis and degradation of c-di-GMP, respectively (Ross et al., 1991).Coucheron (1991) reported insertion of the IS1031 element upstream from the transcription start of this operon, which resulted in the deficiency of cellulose synthesis.
Our studies carried out with the use of the PCR-MP approach (Masny & Płucienniczak, 2003) indicate that genomic libraries from Cel + and Cel - strains contain DNA fragments of different thermal stability.In the case of the cellulose producer (Cel + form) a DNA fragment migrating just above 242 bp is less stable than the corresponding fragment from Cel -cells.A comparison of their sequences has revealed that the thermally less stable fragment (Cel + ) has array of four T-residues (positions 76-79, Fig. 2), while its counterpart from Cel -cells contains at the same positions only three T residues.
Alignment of the 218 nt sequence identified in our studies with those deposited in the GenBank database (National Center for Biotechnology Information) reveals that several genes of Acetobacter xylinum and Gluconacetobacter oxydans contain in their 3'-terminal parts (downstream from their ORFs) nucleotide sequences similar to the fragment sequenced in our studies.The 2, 702, 173 nt-long Gluconacetobacter oxydans genome contains five regions showing significant homology to this fragment (Table 4).However, it should be noticed that all these sequences are shorter than 218 nt (see Fig. 2).Because the 218 nt fragment contains inverted repeats (see Fig. 2), one can speculate that the corresponding RNA arising during transcription can form stem-loop or hairpin secondary structures (see Fig. 4).It is known that these types of RNA structures can regulate gene expression by their involvement in the transcriptional terminator/antiterminator mechanism (Turner et al., 1994;Yanofsky, 2000).It cannot be excluded that the loss of the single T residue results in a change of the RNA secondary structure which may influence transcription of the genes involved in cellulose synthesis.

2005
A. Krystynowicz and others Table 1.Results of analysis of peptide fragments derived from gel slices numbered 3, 4 and 9 (according to Fig. 3).

Gel slice number 3
Gel

Molecular aspects of bacterial cellulose synthesis
A comparison of the protein profiles of wildtype (Cel + ) A. xylinum strains and their Cel -mutants revealed some differences.Cel -cells do not synthesize two key enzymes involved in cellulose biosynthesis, namely phosphoglucomutase and glucose-1-phosphate uridylyltransferase.The lack of glucose-1-phosphate uridylyltransferase in cellulose non-producing cells was reported earlier (Valla et al., 1989).Our studies proved that A. xylinum cells of Cel -phenotype are also deprived of phosphoglucomutase.Although it is known that phosphoglucomutase and glucose-1-phosphate uridylyltransferase (or UDP-glucose pyrophosphorylase) play a critical role in cellulose synthesis, their genetics and regulation have not been studied in detail.
Taking into account the results of the PCR-MP analysis and 2D electrophoresis as well as the alignment of the nucleotide and peptide sequences, one can suggest that the cellulose deficiency in A. xylinum Cel -cells can result from the lack of expression of the genes encoding phosphoglucomutase and glucose-1-phosphate uridylyltransferase.Our finding that in the Cel -form the ORF encoding phosphoglucomutase may be flanked by a DNA fragment of a sequence different than that in Cel + cells can indicate a possible transcriptional repression of the gene expression.

Figure 1 .
Figure 1.Electrophoretic pa�ern a�er PCR-MP of DNA preparations from Cel -and Cel + bacteria (lanes -and +, respectively).Experiments for each temperature were repeated two times.T d 's of the PCR are marked above electrophoretic lines.DNA bands taken for cloning and nucleotide sequencing are marked with arrows.Lanes marked M, DNA markers: 110+111, 147, 190, 242, 331, 404, 476 and 489 base pairs.

Figure 2 .
Figure 2. The nucleotide sequence of the 218 nt element from Cel + A. xylinum E 25 strain isolated a�er PCR-MP (shown in capital le�ers) and its comparison with the corresponding Gluconacetobacter xylinus and Acetobacter europaeus nucleotide sequences.The array of T residues where deletion of one T occurs in Cel -is shown in bold.Inverted repeat sequences are shown by gray background.The GenBank accession numbers of the sequences under comparison are shown in square brackets.

Figure 3
Figure 3. 2-D separation of a lysate of Acetobacter xylinum E 25 cells.Proteins from (Cel -) (A) and wild-type (Cel + ) (B) were separated by 2-D electrophoresis and visualized by silver staining.The pH gradient (horizontal direction) runs from pH 3-10 and the SDS/PAGE separation (perpendicular direction) was performed in 12.5% gel.The arrows indicate the proteins identified as phosphoglucomutase (spots 3 and 4) and UDP-glucose pyrophosphorylase or glucose-1phosphate uridylyltransferase (spot 9).