Cellulose in . Origin of Cellulose Synthase?

David R. Nobles, Dwight K. Romanovicz, and R. Malcolm Brown, Jr.* Section of Molecular Genetics and Microbiology, The University of Texas, Austin, Texas 78712

Although cellulose biosynthesis among the cyanobacteria has been suggested previously, we present the first conclusive evidence, to our knowledge, of the presence of cellulose in these organisms. Based on the results of x-ray diffraction, electron microscopy of micTofibrils, and cellobiohydrolase I-gold labeling, we report the occurrence of cellulose biosynthesis in nine species representing three of the five sections of cyanobacteria. Sequence analysis of the genomes of four cyanobacteria revealed the presence of multiple amino acid sequences bearing the DDD35QXXRW motif conserved in all cellulose synthases. Pairwise alignments demonstrated that CesAs from plants were more similar to putative cellulose synthases from Anabaena sp. Pasteur Culture Collection 7120 and punctiforme American Type Culture Collection 29133 than any other cellulose synthases in the database. Multiple alignments of putative cellulose synthases from Al1abal'l1i1 sp. Pasteur Culture Collection 7120 and N. pUl1ctlforme American Type Culture Collection 29133 with the cellulose synthases of other prokaryotes, Arabidopsis, Gossypium hirsutllm, Populus alba x Populus tremula, corn (Zea I"Iwys) , and Dictyostelium discoideurn showed that cyanobacteria share an insertion between conserved regions Ul and U2 found previously only in eukaryotic sequences. Furthermore, phylogenetic anatysis indiciltes that the cyanobacterial cellulose synthases share a common branch with CesAs of vascular plants in a manner similar to the relationship observed with cyanobilcterial and chloroplast 16s rRNAs, implying endosymbiotic transfer of CesA from cyanobacteria to plants and an ancient origin for cell nlose synthase in eukaryotes.

Ceilulose is the most abundant biopolymer on Decho, 2000; Nicolaus et al., 1999), a tightly bound earth with some lOll tons produced annually (Hess et sheath that is often highly fibrillar and sometimes al., 1928). To date, clear examples of this process have crystalline (Frey-Wyssling and Stecher, 1954; Singh, been found in prokaryotes (Acetobacter xylinus, 1954; Tuffery, 1969; Hoiczyk, 1998), mucilaginous Agrobacterium tume/aclms, Rhizobium spp, [Ross et al., slime loosely associated with cells (often partially 1991]; Escherichia coli, Klebsiella pneumoniac, Salmonella water soluble; Drews and Weckesser, 1982), or a tran­ typhimurium, [Zogaj et al., 2001); Sarcina ventriculi siently attached slime tube, present in motile fila­ [Roberts, 1991]) and eukaryotes, including animals ments (Castenholz, 1982; Drews and Weckesser, (tw1icates), algae, fungi, vascular plants such as 1982; Hoiczyk, 1998). Cyanobacterial EPS are in­ mosses and ferns, gymnosperms and angiosperms volved in a wide range of functions including desic­ (Brown, 1985), and the cellular slime mold Dictyo­ cation tolerance, protection from UV light, and adhe­ stelium discoideum (Blanton et al., 2000). Thus far, sion to substrates, as well as motility (Ehling-Schulz evidence is lacking for cellulose biosynthesis among et al., 1997; Phillipis and Vincenzini, 1998; Kodani et the Euryarchaeota, although \,ve have found puta­ al., 1999; Nicolaus et al., 1999). Although several tive cellulose synthases in the genome databases of reports in the literature have suggested the presence Thennoplasma (http://www.ncbLnlm.nih.gov/ cgi-in/ of cellulose in cyanobacterial EPS (Frey-Wyssling Entrez/framik?db=Genome&gi = 168) and Fcrroplas­ and Stecher, 1954; Singh, 1954; Tuffery, 1969; Winder ma (http://spider.jgi-psf.org/JGLmicrobial/html/). 1990), none has conclusively demonstrated ceilulose One of the most ancient extant groups of living biosynthesis among this group of organisms. There­ organisms is the cyanobacteria, having been in exis­ fore, we sought to examine representatives from di­ tence for more than 2.8 billion years (KnoU, 1992, verse genera of the cyanobacteria for the presence of 1999). Fossil records of cyanobacteria-like forms date cellulose, employing stringent methods for positive as far back as 3.5 billion years (Schopf and Walter, identification. Using cellobiohydrolase I (CBHI)-gold 1982). Cyanobacteria produce a wide array of extra­ labeling and x-ray diffraction, we demonstrate the cellular polysaccharides (EPS), which can take the presence of celluLose in six strains of five genera. form of released polysaccharides (Kawaguchi and Four additional strains appear to have cellulose as evidenced by CBHI-gold labeling Three of the five sections of cyanobacteria are represented among cel­ J This work was supported in Pilrt by grJnts from the Division lulose producing strains. of Energy Biosciences, the Department of Energy (grant no. DE­ FG03-94ER20145), and the Welch Foundation (grant no. F-1217). Recent genome sequencing projects allowed us to , Corresponding author; e-mail [email protected]; fax mine databases of cyanobacteria and other pro­ 512-471-3573. karyotes for protein sequences with similarity to cel­ Article, publication date, and citation information can be found lulose synthases. In all, 17 prokaryotic (five of which at www.plantphysiol.org/cgildoi/l0.1104/pp.010557. were cyanobacterial) and eight eukaryotic cellulose

Planl Physiology, October 2001, Vol. 127, pp. 529-542, www.plantphysiol.org

Electron Microscopy X-Ray Diffraction Microfibrils of varied morphology were observed X-ray analysis of acetic nitric-treated EPS fractions in the EPS isolates of eight cyanobacteria and in the revealed patterns typical for cellulose I in Oscillatoria slime tube isola tes of Oscillatoria princeps (Table I). sp. UTEX 2435 (data not shown), Nostoc sp. UTEX These microfibrils were strongly labeled with CBHI­ 2209 (data not shown), Gloeocapsa sp. UTEX L795, S. gold, indicating that they are composed of 13-1,4­ JlOjmanni UTEX 2349, Anabaena sp. UTEX 2576, and P. glucans (Okuda et a1., 1993; Tomme et a1., 1995; autumnale UTEX 1580. Figure 4 represents the diffrac­ Fig. 1). The narrow and wide axes of the microfibrils were measured from representative samples of all tion patterns of four genera. Line broadening in these cyanobacterial cellulose microfibrils. The mean micro­ diffractograms suggests a low crystallinity, which is fibril thickness is rather constant: 1.7 nm (±0.4 nm) consistent with the small microfibril size observed. based on 65 measurements, ranging from 1.1 nm to 2.8 Electron diffraction patterns were also obtained (data nm. The mean microfibril width was more variable, not shown), but were very diffuse, supporting the with a mean 10.3 nm (±4.1 nm) based on 10 measure­ size measurements and the x-ray line broadening. ments, ranging from 5 nm to more than 17 nm. The presence of contaminating crystalline materials Cellulose microfibrils exist as only a fraction of the is evidenced by the existence of peaks not related to EPS investments of cyanobacteria, as demonstrated cellulose. Note that these peaks do not produce uni­ by thin sections of S. hofmanni UTEX 2349 (Fig. 2) and form overlaps with all tour genera. untreated slime tubes from O. princeps (Fig 3A). In The acid-insoluble EPS fraction from Synechocystis the case of S. hojmanni, a single layer of CBHI-gold sp. American Type Culture Collection (ATCC) 27184 labeled material is seen near the outer boundary of the thick sheath The labeled region is approximately yielded an x-ray diffraction pattern with d-spacings 300 nm wide, slightly less than one-third of the thick­ of 4.6 and 4.0 A, indicating a crystalline product. ness of the sheath. Microfibrils, which seem to be These reflections, however, do not correspond with enveloped in an extracellular matriX, are somewhat reflections of cellulose I or II. It is interesting to note difficult to resolve in the untreated slime tubes of O. that the contaminating peak preceding the 002 re­ princeps. The chemical composition of the matrix is flection in the S. hojmanni x-ray diffraction pattern

Table I. Summation of cyanobacteria investigated and results of experiments ACId Microfibrils Organism Motility Location CBHI-Gold X-Ray D,ffr

Section I Synechocystis sp. ATCC 27184 Yes Yes No NAb No 0.46 and 0.40 nm No Synechococcus sp. UTCC 100 No Yes No NA No Not tested No Cloeocaps

e TEM, Transmission electron microscopy. b NA, Not applicable.

530 Plant Physiol. VoL 127, 2001 Cellulose in Cyanobacteria

"., ...*' D

• • ~'t;~ • 'WI .,...

E F

Figure 1. A through F, Various cellulose microfibnls Isolated from cyanobacteria (all negatively stained with 1% lv/vi aqueous uranyl acetate and labeled with CBHI-gold; the gold complex is 10 nm in diameter). A, Oriented bundles of microfibrils from Cloeocapsa sp. L795, many of which are stained with the CBHI-gold complex, which specifically binds to the surface of cryst

Plant Physio!. Vol. 127, 2001 531 Nobles et al.

Figure 2. NegatIvely stained thm section of S. hofmannl UTEX 2349 with Epon removed and labeled with CBHI-gold. Note the layering of the shedth dnd the position of the cellulose region nc-,H the outer boundary of the sheath. Sheath thickness is designated by the white line. Bar 60 nm.

~ ' ..

(http://www.kazusa.or.jpI cyanol cyano.html; http:// atic errors such as long branch attraction when con­ spidcr.jgi-psf.org/JGLmicrobial/htmi/) revealed cel­ structing phylogenetic trees (Xiong et aI, 2000). lulose-synthase-like (CSL) sequences in 5ynechocystis Criteria for inclusion of ORFs were simply the pres­ sp. Pasteur Culture Collection (PCC) 6803, Anabaena ence of the conserved motif and an arbitrary expec­ sp. PCC 7120, and N. punct!forme. Searches of the tation value <10-9 when compared with any known Anabaena sp. PCC 7120, N. pU11ctifonne, and Synecho­ cellulose synthase. The search yielded sequences coccus sp. WH8102 databases revealed the presence of from both gram-negative and gram-positive Eubac­ ORFs with significant sequence similarity to known teria as weB as members of Euryarchaeota. Multiple prokaryotic and eukaryotic cellulose synthases. alignments using all prokaryotic amino acid se­ We were interested in determining the relationship quences (17 sequences total) as well as sequences of the putative cellulose synthases from cyanobacte­ from Arabidopsis, Gossypium hirsutum, corn (Zea ria to known cellulose synthases. BLAST searches mays), Populus alba X Populus tremulus, and D. discoi­ with the amino acid sequences from known cellulose deum were constructed with ClustalX (Thompson et synthases of prokaryotes and D. discoideum against aI., 1994). The presence of an insertion region corre­ the Arabidopsis genome database demonstrated that sponding to the plant-specific and conserved region the CesA of D. discoidewn had the lowest expectation (CR-P; Delmer, 1999) between regions U1 and U2 value when compared with CesAs of Arabidopsis. was observed in eukaryotic sequences as previously Subsequent pairwise alignments showed that the pu­ described by Blanton et al (2000). The amino acids of tative cyanobacterial cellulose synthases represented by Anabaena sp. PCC 7120 (contig 326) and N. punc­ this insertion block were highly conserved aLUong tiforme ATCC 29133 (contig 499) had expectation val­ plants but showed very little sequence similarity to ues significantly lower (2 X 10-20 and 2 X 1O-J8 the insert of D. discoideum. Two putative ceBulose times, respectively) them that of the cellulose syn­ synthases from Anabaena sp. PCC 7120 and N. PU!1C­ thase of D. discoideum when compared with IRX3 of tiforme showed the presence of the insertion region Arabidopsis. Although BLAST searches are some­ found in O. discoideum and vascular plants but lack­ times used to infer phylogenetic relationships, pair­ ing in other prokaryotes (Delmer, 1999; Blanton et al., wise alignments are inherently limited in this capac­ 2000). Additionally, the cyanobacterial insertions ity. Thus, we chose to perform multiple alignments showed greater similarity to those of plants than D. for the construction of phylograms. Toward this end, discoideum (Fig. 5). The sequences of cyanobacteria we sought to identify as many prokaryotic cellulose and O. discoideum lacked a second insertion region synthase sequences as possible from the above data­ between U2 and U3 present only in plants (Delmer, bases in order to minimize the possibility of system­ 1999; Blanton et aI., 2000).

532 Phmt Physiol. Vol. 127, 2001 Cellulose in Cyill10bacteriil

groups in the tree. The topologies were analyzed A using the test (5% significance) of Kishino and Ha­ segawa (1989) and compared to the original topol­ • ogy. All other topologies showed log likelihood val­ ues smaller (even when SES were taken into account) and were rated by TreePuzzle as significantly worse trees than the maximum likelihood topology (Figs. 7 and 8). • The trees generated place cyanobacterial sequences in three distinct branches (Figs. 6 and 7). Synechocys­ tis sp. PCC 6803 (D90912), N. punctiforme ATCC 29133 •• (contig 583), Anabaena sp_ PCC 7120 (contig 294), and B. subtihlS (CAB12237) display a close grouping in illl - trees. Based on their small size and low similarity to • •• known cellulose synthase sequences, these sequences most likely represent CSLs rather than cellulose syn­ • thases. Two cyanobacterial sequences (N. punctiforme B ATCC 29133 [contig 640] and Synechococcus sp. WH8102 [contig 251]), present in a second main branch of the tree, group distantly with cellulose .. "", •.,. .. synthases of other gram-negative . Anabaena • ..... sp. PCC 7120 (contig 326) ilnd N. punctiforme (contig .! .. .. 499) share a sister branch with vascular plants, indi­ •" • • • • • ... •• • cating a distinct phylogenetic relationship between the cellulose synthase homologs of cyanobilcteria and .. ..• " .­ ·" CesAs of vascular plants. •••••• t DISCUSSION ". Cellulose Structure Although several previous reports suggested the presence of cellulose among the cyanobacteriil, it was Figure 3. A, Untreated slime tube material collected from the liquid actually a surprise to find that the product is so culture 0. pnlJceps. The lllicrofibriis are somewhat masked In the EPS matrix. B,.lr = 20 nm. B, Trlfluoroacetic-dlgested slime tube widespread among this group of organisms. Thus matertal collected from liquid media. Microfibrils are more apparent far, we have found three of the five sections of cya­ here as is a decrease in matriX material. Bar = 40 nm. nobacteria to have genera exhibiting cellulose bio­ synthesis. It is intriguing to note cellulose produc­ tion, but it is also necessary to understand where it is Trees generated using both neighbor joining (NJ) assembled. We have found cellulose in slime tubes, and maximum parsimony (MP) showed similar over­ sheaths, and extracellular slime, the three major all results (Fig. 6). As expected, the vascular plant classes of extracellular polysaccharides in the cya­ sequences formed a monophyletic group. As de­ nobacteria. The presence of cellulose in the sheath of scribed previously (Holland et aI., 2000), CesA or­ S. hofmanni represents the first report of prokaryotic thologs were seen to be more similar than paralogs. secretion of cellulose as an attached, capsular-type NJ and MP methods demonstrated a close phyloge­ EPS. This is also the first report of cellulose in slime netic relationship between the amino acid sequences tubes of motile cyanobacterial trichomes. of cyanobacteria and plants supported by high boot­ What is known about the molecular machinery strap values (Fig. 6). The branch connecting the cya­ involved in cyanobacterial gliding motility? Gliding nobacteria to plants is very deep, thus raising the motility takes place with the concomitant secretion of possibility that the placement of the cyanobacterial polysaccharides. Circumstantial evidence implicates branch was a long branch attraction artifact. To test junctional pore complexes (JPCs) as points of poly­ the validity of the cyanobacterial/plant grouping, a saccharide secretion associated with gliding motility quartet puzzling (QP) maximum likelihood tree was (Hoiczyk and Baumeister, 1998). Recently, outer constructed from the multiple alignment. The maxi­ membrane pores necessary for the secretion of a Type mum likelihood phylogeny reinforced the cyanobac­ 1 capsular polysaccharide in E. coli were described teria/plant relationship demonstrated with NJ and (Drummelsmith and Whitfield, 2000). These pores MP methods (Fig. 7). Of 954 trees generated, only are exposed on the outer surface and, based on trans­ 1.3% failed to show this relationship. Additionally, mission electron microscopy examinations, have alternative topologies were generated by manually structural similarity with cyanophycean IPCs Cellu­ grouping the cyanobilcterial clade with other lose present in the slime tubes of actively motile

Plant Physio!. Vol. 127, 2001 533 Nobles et al.

3

B c

D

10 12 14 16 16 20 22 24 26 28 30 32 34 36 2 theta Imm) Figure 4. X-rdY analysi> of cellulose from four different cyanobacterial strains. A, Cloeocapsa; B, Scytonema; C, Phor­ midium; D, Anabdena. All s~nera exhibit a typical cellulose I pattern specified by the overlapping peaks. Peak 1 is the 101 d-spacing; peak 2 is the -101 d-spaCing; and peak 3 is the 002 d-spacing. For the 002 spacings, all genera with the exception of S. hofmanni (4.0 A) have reflections of 3.9 A. For the -101 spacings, all genera with the exception of CloeocapsiJ (5.4 A) have 5.3-A reflections. For the 101 spacings, all genera with the exception of P. autumnale (5.9 A) have 6.0-A reflections. The presence of contaminating crystalline materials is evidenced by the existence of peaks not related to cellulose. Note that these peaks do not always produce uniform overlaps with all four genera. cyanobacterial filaments suggests a possible tie to alga Erythrocladia also have small dimensions of 1 to gliding mOtility. The maximum rate of gliding in 1.5 X 10 to 33 nm (Okuda et al., 1994). Distinct linear cyanobacteria is more than 100 times the rate of terminal complexes assemble these thin yet wide mi­ cellulose ribbon synthesis in A. xylinus, suggesting crofibrils. Evidence implicating the IPCs as sites of that the cyanobacterial movement is based on a polysaccharide secretion in the cyanobacteria is cir­ more complex set of rna terials and processes than A. cumstantial; however, it is interesting to note that the xylinus. In A. xylinus, the rate of movement is ap­ IPCs are located in a ring around the filament, thus in proximately 2 to 3 um min-1, whereas cyanobacte­ a favorable position to secrete polysaccharides in rial filaments have demonstrated the ability to glide general. It appears that the cellulose from cyanobac­ at up to 10 um S-1 (Castenholz, 1982). We can only teria is only a small percentage of the totClI extracel­ speculate on the role of cellulose assembly in cya­ lular polysaccharide, as compared with A. xylinwrl nobacterial gliding motility at present. Perhaps the and cell walls of vascular plants (Brown, 1985). It hClS slime tube is a frictional bearing for anchoring of other motility elements responsible for the glid­ been demonstrated that when A. xylinus synthesizes ing; however, relatively little is known about this cellulose in the presence of carboxymethyl cellulose process. and other polysaccharides, the crystallinity of the Interestingly, the microfibrils of cyanobacteria cellulose is reduced (Hirai et aI., 1998). Perhaps, in a have smaller dimensions than those of other organ­ similar fashion, the copious production of extracellu­ isms, including A. xylinus (Brown et aI., 1976), certain lar materials other than cellulose by cyanobacteriCl algae such as Valonia (Itoh and Brown, 1984), and prevents lateral aggregation of microfibrils into Boergesenia (Itoh et al., 1984), Clnd vascular plants larger units, thus accounting for the small observed (Brown, 1985). The microfibrils of the Rhodophycean crystallite size.

534 Plant Physiol. Vol. 127, 2001 Cellulose in Cyanob

134 214 ~;;;;;~~~!~ ~~:j:jjj:j::~:j~jjjj~:~jj~~~jj~~~:~~~:~::~~::~j~:~:~jjjj:j:jj:j:j~ 190 ._,..... 172 184 A_aeolicus__D70422_ TALAAINMDYPS~I~K------­189 ~7l20__c294_ LAKNLCNLBYPNGQYEVWIIDqNST------­ 148 SYD_6803__D90912_ IVOOLCSLDYPGDRBEVWIvpqNS~------­150 N-pun__c583_ L~YEVWII~RSS------148 Ath__BA809693.l_ FVLSILAvDrPvpKvSCYVS~FSS~S~ARKWVPFCXKYSI.PRA------­412 ZIDa__AAF89961.1_ TVLSILSVDYPvpKVSCYVSDDGSAMLTFSSL~A!FARKWVPFCKKBNI.PRA------­424 Gbi T10797 TVLSILALDYPvpKVSCYISpDGAAMLTFBS~ADFARKWVPFCKKFSI.PRA------­326 Pt/Pa__AAC78476.1:: TVLSILSVDYPvpKVSCYVS~BMLLFDS ABrARKWVPFCKKBNI.PRA------375 Ath__AAC29067.1_ TVLSILAVDYPvpKVACYVSPQGAAMLTFZALS ~FARKWVPFCKKYCI.PRA------­ 426 Ath__T51579_ TVLSI~VBKISCYVS~SMLT~LS~~ARKWVPFCKKFSI~PRA------­391 Ghi__AAD39534.2_ FVLSILAVDYP~SCYVSDDGAAMLT~~S!FARKWVPFCKXYNI.PRA------­413 N_P~c499_ TAKAALVKDYPPTKLR~SA-~RLCIEDLOSPOLOQEANRIDA------­176 A_7120__c326_ TAQAALAIDypFT!L~SP-AMRAMTIKLCIADLQSPLLOQEANRIDA------­203 212 w:~~~--~~~~- ~~~~;~~======123 D_discoidaW;;_AAF0020SYD 0 . 1:: TFRTAISMDYPS~WIGLLDIlSVNYRESRal'IAHLQSVEKNFLYVLr.gKAVYSVIDIIRPPV'1'SQ- ·PBGILNitTSSiU.SillS.m~lClAIl 438 91 131 169 T_aCido::::::~~1~1~~:A_tume£acienS__I39714_ TLAAAKNM;YP~~~~~======. FTVWLL ------303 ruler 450 460 470 480 490 ....•••500 510 520 530 .

E_coli_AAC76558.1_ ------GREEFRQFAQNVG------147 B_cepacia ------RRPE~BBvG------­227 A_XY1inus__2019362A_ ------RREEFARF CG------­ 203 R_1egum1nosarum~28574.1_------QKRNSGKLLEAQAAAARBIELBQLC LO------­ 201 R.sphseroides ------QRCMSPDPELAQKAQERRRELQQLCRlLG------213 220 A_ae01*c~~20D1~~;~ ======~~~~~~:~~~~======~======:==163 SYD_6803__D90912:: ------D--R-TPAILDQLROOyp------.------165 ~_Pun__c583_ ------D--S-TPBLLAELEQKYD------163 Ath_BAB09693 . 1_ ---- -P&wYFAAJ::r~LKl:lKVQTB1"VKDRRl\IlIOlEYDFKI- --lUKALVSltALJ'(CPeQ%lGTPWPG------1OI-TRllBPGMIg 486 ZIDa_AAF89961 . 1_ ---- -P/IFYl"AQlUDYL~KIQPSFV1lERRAIIJtRE~FKV- - -RDJALVAKAQItVP TAWPG------1IIl'l-PRDBPGMIQ 498 Ghi__TI0797_ -----P~FYF~DYL~QPS~Yl8YKI---aIRALVAKAQKTPD TSWPG------NN-PRDBPGMIQ 400 Pt/Ps__AAC78476.l_ -----P~lUDYLKIKVBP~RRAMKREr.lBFKV---aIRALVSKAQKKP QDGTPWPG------NI-TRDBPGMIQ 449 Ath_AAC29067.l -----PIIWYP'CIl1aIDYLKNKVBPAP'VRERRAMItRDYDFKV---1tDlALVAT~B'ilowTMQl\GTPWPG------IilS-VRDHPGMIQ500 Ath T51579 -----P~XVIlYLQ~P'rFVItERlll\ImRE~FKV---RI1I1AQVAKASKVP~MQBaTPWPd------NN-TlqlBPGMIQ465 Ghi_AAD39534.2 -----PpnrAQItI~KVOTSrvJO)RRAIIKREYBBn:V---RDlGLV~~D1QDCn'1'WPG------IOI-TRDBPGMIQ487 ~_PUD....c499. -----EHSSLLQRLKELENL'1'PNTQGAEQWLQTSSS PVVN---QLATEFVQSLRQF-ILWLPPTBQSISD8PVI'1'~L~RKALB 257 A_7120__c326_ -----ERSBLVl:RLQKLESFTSNIQVAEEWLSItS8SVPNQD--~SLQQP-IRWLSPKHBSIA------ERLK~QVLAQ277 ~_PUD__c640_ ------RRPBI~------­ _ 225 SYD_WB8102__c251_ ------REEVRQLARRLG------135 D_discoideum__AAF00200.1_ VQWFIEYPLLNSWFGVGQCIPRDADDAERALIAKLRoDNFSFYaTFTKSESEKI8NFTIDSLQ8LWHGSA------FFRPLIRSILLK 520 106 B~subt~l~~~~;:::~; ======~~~~~======144 T_acidopbi1um-=CACl1626.1_ ------KIVSOLAEVcwILG------183 A_tume£aciens_I39714. ------VO~IVEAO~QRRBEELltKLClPLD------­333 ruler 540 550 560 570 580 590 600 610 620 .

: eo: • E_coli_AAC76558.1_ ------VXYIARTTB---E-~ 168 B_ce.pacia 248 A.xylinus__2019362A_ ======~~===~=~ 224 R.1eguminosarum__AAD28514.1_ ------DVSYLTRDRN---E-BAKAGNLNNQM 223 R_sphaeroidea 235 A_seolicu8__D70422_ 245 A.7120_c294. 186 Syn_6803__D90912_ ======~~===~~ 188 N_PUD__c583. ======~~===~~~ 186 Ath__BAB09693.1. --VF------LG------.------QNQG~L . PRLVYVSREKRPOPQB8KKAaAMNALV 529 ZIDa_AAF89961.1 ... --VF------LG------B8OQL DG~LPRLVYVSREKRPOFQ~I 541 Ghi T10797 --V7------LG------Y~IPRLVYV8REKRPGYQBBKKAQAZNALV 443 Pt/Ps__AAC78476.1= --VY------LG------LPBLVYV8REKRP~I 492 Ath_AAC29067.1 --V7------LG------B PRLVYVSREKRPOPD~SLI 543 Atb_TS1579 508 Ghi__AAD39534.2_ ==~~:::~======~~~:&s~~ 530 N..pUD__c499. --TIRK--KBI.------L LSRrRYIARPItT~AI 296 A 7120 c326 --AIRQ--KBL------LTRFRYIARPKPAOVPBBAKAONLRYAI 316 N....PUD-c640= ------CEYITRPDN----~ 246 SYD.WB8102_c251. ------CRrQBRPBR----RRAKAaNLRAQL 156 D_discoideum__AAF00200.1 ~SJLN1'lQIl!U_R!'L~Q1rQVLMMGRQIL IgllONVRI~DOPIVBPltCTlrLRRRKPP--:IPIINIWIN:IIOIAL 608 8_subtilis__D69769_ ------Y:zKMVITltPPN---AQKQltS8ALNSGF 130 F _AC idarmanU8 ------AvYIKRTDR----SoYKAQALNNAL 165 T.scidopbi1um.-CAC11626.1. ------IKFVBRNNR----RQ~IRDAL 204 A.tumefsciens__I39114 ------.------VRYL~----~L 354 ruler ... 630 640 • •••••• 650 ...... • 660 670 680 , .690 700 710 . Figure 5. Multiple alignment of amino acid sequences from 17 prokaryotic cellulose synthase homologs with CesA sequences from Arabidopsis, Cossypium hirsutum, corn, Populus tremu/us x Populus alba, and D. discaideum. The alignment demonstrates the presence of a CR-P region between the U1 and U2 domains present only in eukaryotic and cyanobacterial sequences.

Phylogenetic Analysis tool for determi..ning phylogenetic relationships. Us­ With the exciting progress in genome database ing these advances, we were able to gain insight into mapping, researchers have an extremely valuable possible relationships of cellulose synthase among

Plant Physiol. Vol. 127, 2001 535 Nobles et a1.

Ath 1 NJ, MP, and QP trees based on the above multiple ~,.'''' Ghi 2 alignment showed that putative cyanobacterial cellu­ 62(67) 2ma lose synthases branch closely with plant cellulose 61 synthases. Even though the bootstrap and confidence (100) Ath 3 T1 scores are high, the branch connecting the cyanobac­ Ghl 1 91 terial sequences of vascular plant sequences is quite PtlPa 100 deep, leaving open the possibility of a long branch 1 Alh 2 attraction artifact. However, maximum likelihood ,-- N puncurorme 1 evaluation shows that tree topologies separating the '00 T3 '------1 (100) Anabaena 1 183) '------'-----'------­ cyanobacterial and plant clades are not viable. An '----- Ddiscoideum alternative explanation for the long branch lengths may lie in the ancient origin of chloroplast endosym­ _____ .~ __ R leguminosarum biosis which, according to the fossil record, must ,-----1 '00 (1001 I100 ------A tumetaclens have occurred more than 2.1 billion years ago (Han (941 69 '------R sphao(oides and Runnegar, 1992). Such a vast time for divergent IT2) fJ'------r------, ASQoHcus evolution could easily account for the formation of ,­ 99 ,/ (86)~'/ ,------Ecoll deep branches. ' 10°(971 There is substantial evidence that chloroplasts of _ 100 '------=:....------B cepacja - 84 (100) vascular plants have a cyanobacterial origin and that ~'----- Axyllnu$ (57) genes can be transferred from the chloroplast to the Ir------N punc:tlfonne 2 nuclear genome (Weeden, 1981; Martin and Schnar­ '----- Synechoc:occ:us renberger, 1997; Martin and Herrmann, 1998; Martin ,-- Anabaena 2 et al., 1998; Krepinsky et al., 2001; Rujan and Martin, ,------1 100(100} 2001). Chloroplasts, originatll1.g from the endosymbi­ ,------11 °0 '----- N plJnc:tlfonno 3 IL_11_00_'' _ otic capture of cyanobacteria, may have shed their 99 Syneehoc:ystts (98) cellulose synthases and donated them to the nuclear 1141 '------Ssubtliis genome. It is known that a subgroup of cellulose r------F acidannanus 85 synthases in Arabidopsis exists that bears minor se­ '----- Tacldophllum quence similarity to prokaryotic cellulose synthases 0.1 (Taylor et al., 1999). Our findings offer a possible Figure 6. Comparison of Nj and MP trees. The tree shown is an NJ tree subjected to 5,000 bootstrap trials. Bootstrap values are shown o drs<:olc1oum as percentages with MP bootstrap values shown in parentheses. Differences in the MP tree are denoted by bold IIlles (multifurcations) and dashed arrow (variable position), and an asterisk indicates root­ T llcidopllilum ing at the base of the tree. Note the distribution of cyanobacterial A rumofaclons sequences in the tree: two sequences from Anabaena sp. pee 7120 SYI'ICehOUU$ R lr.gunllnosarum 16 R sp~aoroldD1l and N. punctiforme branch with vascular plants; two sequences from N. pUIJctiforme and Synechococcus branch distantly with Thermo­ togales and Proteobacteria; and three sequences from Synechocystis, F acid.annanus 67 N. punctiforme, and Anabaena, which are most likely eSL proteins, group with Bacillus sublilis. The high bootstrap values support the N pum:bfo(l'J1e 2 validity of the tree.

N puneijfonne 1 8 ecpacla """-----< " ,. Anilbaena 1 the cyanobacteria and other organisms. Pairwise alignments against the Arabidopsis genome database Axyllnus demonstrate that cyanobacterial cellulose synthase homologs have significantly lower expectation val­ Onl1 ues than any other cellulose synthases not originat­ .. 'th' 95 Ath1 N punetlfor'l1'olll 3 T' ing from plants. Further evidence for a close relation­ 'T '" 61 Ghl'}. Aroabaen... 2 ship of cyanobacterial and plant cellulose synthases 6~: PtlPa was gained by performing a multiple alignment of 5ynechoc)'stls 'th2 the 17 prokaryotic sequences with cellulose syn­ E!subtlll, thases of D. discoideum and vascular plants. Unlike other prokaryotic sequences, two cyanobacterial se­ In L =-37924.32 qLlences share a CR-P region with D. discoideum and Figure 7. The maximum likelihood phylogeny for the cellulose syn­ vascular plants. In addition, plant CR-P regions thase sequences showing confidence values and the log likelihood. show greater sequence similarity to the correspond­ With the exception of O. discoideum (which groups with the Eur­ ing cyanobacterial regions than to the insert of D. yarchea in this tree), the relatlon;hips in this tree are nearly identic~1 discoideu111. to those shown in Figure 6.

536 Plant Phy~iol. Vol. 127, 2001 Cellulose in Cyanoba.cteria

F acldarmanus

8subtili

N pUl1ctJforme 1 Anabaena 1 Ath1 Zma N punctlforme 2 Synech cystls Zm. Ath1 Ghi 1 Anabaena 2 h ZAth 3 PtiP A tumefacie Ghl2 N punctlforme 3 R leguminosal'\Jl'riA. aeollc s E coli A Kyllnus Synechococcus Atl12 $Phaeroid~ R cepacia Ath 3 Ghl1 o disc ideum

In lr -37998.15 S.E.= 15.84

F acldarmanus E co~ xyllnu5 Synechococc N punctlfonne 2 B cepacla Ddlscol eum

T .cldophllum Anabaena 2 N punctlforme 3 Synechocystls In L= -37998.67 F acldarmanus S.E.~ 15.74

B subtills

N DUl1c:tiforme 1

Anabaena 1

Ath1 Zma ~~__ A xylinus discoidellm o F acidarrnanU5 P1IP. Ghi 1 E call Ghllth Z Ath 3 B cepacia A tumetaclens N punctlfoan_ t R legum inoS3r'\Jm Anabaena 1 R sphaero\des N punctiforme 2 R leguminosarum T acloophllum R sphaeroides Synechococcus

In L= -37996.50 S.E.= 16.23 N punctifonne 2 B cepacia

Athl

N punctlforme 3 B sublihs Anabaena 2 Synechocystls In lr -37976.88 S.E.= 13.55 Figure 8. Four distinct phylogenies alternative to the maximum likelihood phylogeny for N. punctiforme 1 and Anah)(>na I sequences (underlined). The log likelihood and the SE are shown for each tree. explanation for this similarity. The relationship be­ cellulose synthases originated in the cyanobacteria. If tween cyanobacterial cellulose synthase homologs invalid trees and convergent evolution are eliminated and vascular plant cellulose synthases shown in our as possibilities, the relationship of cyanobacterial se­ phylogram closely resembles the relationship of 16sr­ quences to vascular plant sequences seems inexplica­ RNAs from cyanobacteria and chloroplasts (Olsen et ble except in terms of synology. Unfortunately, the a!., 1994). This evidence suggests that vascular plant current lack of CesA sequences from the algae and

Plant Physiol. Vol. 127, 2001 537 Nobles et at. from more primitive cyanobacteria leaves a significant hormogonia, (Cohen et a1., 1994). These different gap in the tree and makes it impossible to determine, genes may be conditionally expressed in specific with any certainty, the precise nature of the relation­ differentiated cell types serving different functions ship between cyanophycean and plant cellulose syn­ in each. For example, cellulose could serve as a thases. Recently, however, the McCesA3 gene from means of attachment to the host plant in the forma­ Mesotaeniul11 caldariorum UTEX 41 was sequenced by tion of symbiotic relationships in a manner similar Roberts et a1. (2000). Multiple alignments and phylo­ to A. tumefaciens' attachment to its host plant (Mat­ gri:1ms (data not shown) demonstrate not only the thysse, 1983). In addition, cellulose could have roles expected close relationship of the algal sequence to in gliding motility of hormogonia, desiccation tol­ vascular plants CesAs, but also a branch point closer erance, nitrogen fixing efficiency of , en­ to the cyanobacterial/vascular plant divergence hancing Viability of akinetes! or protection from UV point. light. Tn conclusion, the proof of cellulose biosynthesis in No sequences with similarity to cellulose synthase the strains of AnabaCila UTEX 2576 and Nostoc punc­ exist in genomes of chloroplasts or cycmelles se­ tiforrnc ATCC 29133 correlates directly with the dis­ quenced to date. Given the lack of any such ho­ covery of the putative cellulose synthase homologs mologs in the cyanelle genome of the primitive alga from these same organisms. This adds validity to the Cyanophora paradoxa, it is reasonable to assume that if identity of these sequences as cellulose synthases. the cellulose synthases of plants were obtained via Given that the cyanobacteria were probably among cyanobacterial endosymbioses, translocation of the the earliest forms of life on earth, contributing to the gene to the nucleus must have occurred relatively oxygen of the planet's atmosphere over eons, it is early in the evolution of algae. The divergence of interesting to speculate why these organisms, of all cyanobacterial and plant sequences from the branch potential photosynthetic life forms! gained promi­ point reinforces the idea of an early transfer of cellu­ nence in the primitive earth environment. In the re­ lose synthase from cyanobacteria to a eukaryotic cell. ducing atmosphere, severe UV radiation could have Proteins associated with vascular plant cellulose posed a serious hazard; however, a cellulose sheath, synthases have been elusive, and cyanobacteria slime tube, or extracellular matrix could have could offer a simplified model system for identifica­ shielded the cell from UV radiation If the cellulose tion of these proteins. Homologs to the known cellu­ contributed to motility, this ability could have been lose synthase-associated proteins found in A. xylinus important to allow these early cells to move into and A. turnefaciens are lacking in all cyanobacteria shade and avoid the intense radiation. It is ironic that analyzed in this study. In fact, many of the ORFs in the earliest producers of cellulose may be the last the vicinity of the putative cyanobacterial cellulose major group of organisms to have cellulose discov­ synthases have no sequence similarity to any pro­ ered among them! teins of known function. Therefore, it is likely that cyanobacteria possess a heretofore undescribed reg­ ulatory system for cellulose biosynthesis in pro­ MATERIALS AND METHODS karyotes. Curatti et a1. (2000) recently isolated a pro­ Strains and Growth Conditions karyotic Sue synthase gene from Anabaena sp. PCC 7119 This gene, which is very similar to the Sue All cultures were grown photoautotrophically with a 12-h light/dark cycle under fluorescent light. Cultures of synthase of vascular plants, also has substantial se­ Oscillatoria sp. UTEX L2345, Oscillatoria princeps CE-95-0P quence similarity with ORFs found in databases of Cll CC 5122, Phormidium autunmale UTEX 1580, Syneeho­ the cellulose producing N. punctifor71lc ATCC 29133 cyetis sp. ATCC 27184 (also known as PCC 6803), Cl'inalium and Anabaena sp. PCC 7120. Since it is thought that epipsammul11 ATCC 49662, and Synechoeoeeus sp. UTCC 100 Susy from vascular plants is tied to cellulose synthe­ (syn. PCC 7942) were maintained in BCll medium. sis (Amor et a1., 1995), it would not be surprising to Anabaena sp. UTEX 2576 (syn. PCC 7120), N. museorum find that the cyanobacterial homolog to SusY is as­ UTEX 2209, Nostoc punctlforme ATCC 29133, N. muscorum sociated with cellulose synthesis in the cyanobacte­ UTEX 1037, Scytonema hofmanni UTEX 2349, Fischerella am­ bigua UTEX1908, and Gloeocapsa sp. UTEX L795 were main­ ria. If this is the case, perhaps other cellulose tained in BG11 0 medium. Cultures were routinely main­ synthase-associated proteins are shared between vas­ tained on agar plates prepared as previously described cular plants and the cyanobacteria. (Golden et aI., 1987). When possible, contaminated cultures Two putative cellulose synthase proteins and one were purified by the method of Rippka (1988). With the likely CSL protein exist in N. punctiforme ATCC exceptions of Scytonema !lofmanni, Fischerella ambigua, N. 29133, a nitrogen fixing, facultatively heterotrophic, museorum UTEX 1037, and O. princeps, all cultures were symbiotically competent cyanobacterium capable of axenic and were periodically monitored for contamination differentiation into heterocysts, akinetes, and motile by microscopic examination and by growth, in the dark, on

538 Plant Physiol. Vol. 127, 2001 Cellulose in Cyanobacteria

BGll or BGllo plates supplemented with 0.5% (w Iv) Glc Monsanto, St. Louis)-coated grids was labeled without ill1d .05% (wIv) Vitamin-Free Casamino Acids (Difco Lab­ further treatment. CBHI-gold labeling was performed oratories, Beckton Dickinson & Co., Franklin Lakes, NJ). es-sentially as described previoLlsly (Okuda et aI., 1993) Batch cultures were grown either as 500-mL cultures in with the follOWing exceptions: (1) 10 nm gold was used for 2-Lflasks on an orbital shaker or as 2.5-L cultures in growth the CBHI-gold complex, (2) rather than floating grids, chambers with agitation by magnetic stirrers and aeration. 8-J.LL drops of enzyme complex were added to Formvar All cultures were maintained at 28°C with the exception of grids with the EPS isolates, and (3) enzyme complex and 0. princeps, which was maintained at 35 Q C. Cultures were product were incubated for 3 min at room temperature harvested by centrifugation and Llsed directly for experi­ (RT). Grids were negatively stained with 2% (w Iv) uranyl

ments, washed, and resuspended in 0.1 IV! K2HP03 buffer, acetate. pH 7.0 (PB), dnd frozen at -SO°C, or washed in deionized

H20 and lyophilized. Specimen Fixation

Cells were initially fixed in 0.1 IV! K2HP03 buffer, pH 7.0 Isolation of EPS with 2% (v Iv) glutaraldehyde for 4 h at RT. Fixed cells were given three 15-min washes and post-fixed with 2'1u EPSs were isolated through a variation of the method (v Iv) osmium tetroxide in 0.1 Mbuffer, in the dark, for 4 h previously described (Barclay and Lewin, 1985). Lyophi­ at 4°e. Cells were washed three times (10 min each) with lized cells resuspended in 30 mL of PB and frozen cells deionized water and stained with 1% (v Iv) aqueous ura­ thawed at 37°C were broken by three to five passages nyl acetate for 1 h, in the dark, at 4°e. Cells were wilshed through a pre-chilled French Press at 20,000 psi. A micro­ with deionized water for 10 min at 4°C and taken through scope was used to monitor cell breakage. The lysate was a graded eth,mol series: 30%, 50%, 70%, and 90% for 15 spun at 9,180g for 40 min. The supernatant was decanted min each at RT. This was followed by two lS-min dehy­ and the pellet resuspended in 30 mL of PB, and the sus­ drations each with 100% ethanol and 100% acetone. Cells pension was incubated with an excess of lysozyme (1:5 were infiltrated with 33% and 66% EMbed 812 (Electron ratio dry protein weight to wet weight of pellel) at 37°C Microscopy Sciences, Fort Washington, PA) in absolute with gentle shaking for 24 to 48 h. The insoluble fraction acetone for 12 h each at RT and polymerized for 24 h was collected by centrifugation at 14,300g and the pellet at 60°C. washed with PB. The pellet was resuspended in 30 mL of 10 mM Tris, pH 7.5, 5 mM EDTA, 0.4°;;.> (wIv) SDS, and 66.67 J.LM proteinase K and incubated overnight at 37°C. CBHI-Gold Staining of Sectioned Material The insol uble fraction was collected as above and the pellet washed twice with PB. The pellet was extracted twice each Ultrathin sections of fixed specimens were placed on with 30 mL of chloroformlmethanol 1:1 and acetone, carbon-coated Formvar grids. Grids were incubated with washed twice in PB, and incubated with 165 units of Epoxy Resin Remover (Polysciences, Inc., Warrington, PA) a-amylase (A-6380, Sigma, St. Louis) and 360 units of amy­ for 10 min and washed with glass-distilled water. Grids loglucosidase (Sigma A-3042B) in 30 mL of PB, at 28°C, were labeled and negatively stained in the same manner as with gentle shaking, for 24 to 48 h. The insoluble fraction EPS and slime tube material. was collected as above, washed twice with deionized wa­ ter, and lyophilized. Alternatively, sheaths were isolated X-Ray Diffraction essentially as described previously (Hoiczyk, 1998). In this case, cells were broken by passage through a French Press X-ray diffraction ~nalysis was performed with Ni­ at 20,000 psi rather than by sonication. filtered CuKa (1.542 A) at 35 kV dnd 25 mA on a PW 729 x-ray generator (Philips, Eindhoven, The Netherlands) and a Debye/Sherer camera system (Philips). Acetic nitric-digested, lyophilized EPS isolates were Collection of Slime Tubes packed into 0.3- or 0.7-mm capillary tubes and diffraction Slime tube material was pipetted from liquid cultures of patterns collected under the above conditions for 1 h. Re­ actively motile O. princeps 24 to 72 h postinoculation. flections were measured manually with measurements cor­ rected for film shrinkage. Digital images of film negatives were generated using the Interactive Bild Analysung Sys­ tem image processing system. The images were converted CBHI-Gold Labeling to digital diffractograms using NIH Image, Image J, and Microsoft Excel. Slime tube material harvested from liquid cultures of O. princeps and EPS isolates from all other cultures were either observed without further treatment or digested with 1 N Phylogenetic Analysis TFA for 1 hat lOO°C andlor acetic nitric reagent (Upde­ graff, 1969) for 30 min at lOO°e. Slime tube material col­ Sequences with similarity to cellulose synthases were lected from motile filaments on Formvar (polyVinyl formyl; identified with BLAST (Altschul et al., 1990) searches

Plant Physiol. Vol. 127, 2001 539 Nobles et al. against the following databases: (a) Department of Energy Amor Y, Haigler CH, Johnson 5, Wainscott M, Delmer Joint Genome Institute, http://spider.jgi-psf.org/JGL DP (1995) A membr<:lne-associated form of sucrose syn­ microbial/htmll; (b) National Center for Biotechnology thaseand its potential role in synthesis of cellulose and Information, http://www.ncbi.nlm.nih.gov/cgi-bin/Entrez/ callose in plants. Proc Natl Acad Sci USA 92: 9353-9357 framik?db=Genome&gi = 168; and (c) Kazuza, http://www. Barclay W, Lewin R (1985) Microalgal polysaccharide pro­ kaZllsa.or.jpI cyanolcyano.htm!. duction for the conditioning of agricultural soils. Plant Putative cellulose synthase amino acid sequences and Soil 88: 159-169 Blanton RL, Fuller D, Iranfar N, Crimson MJ, Loomis WF were then compared with sequences at '\Jational Cen­ (2000) The cellulose synthase gene of Dictyostelium. Proc ter for Biotechnology Information Arabidopsis Genome Nat! Acad Sci USA 97: 2391-2396 database (http://www.ncbi.nlm.nih.gov:80/cgi-in/Entrez/ Brown RM Jr (1985) Cellulose microfibril assembly and map search?chr=arabid.inf) with BLAST (Altschul ct aI., orientation: recent developments. J Cell Sci Suppl 2: 1990). The contig listings and amino acid accession numbers 13-32 of the sequence legend in the phylognetic trees are as Brown RM Jr, Willison JH, Richardson CL (1976) Cellu­ follows: E. coli, AAC76558.1; Burkholderia cepaeia, contig lose biosynthesis in Ael.'tobacter xylinum: visualization of 411; A. xylinus, 2019362A; Rhizobium leguminosarum, the site of synthesis and direct measurement of the in AAD28574.1; Rhodobaeter sphaeroides, contig 168; Aqulfex vivo process. Proc Nat! Acad Sci USA 73: 4565-4569 aeolicus, D70422; AnabaenaI, contig 326; Anabaena2, contig Castenholz RW (1982) Motility and taxes. 1/7 NG Carr, BA 294; Syneehocystis, D90912; N. pUl1eti(ormel, contig 499; N. Whitton, eds, The Biology of Cyanobacteria. Blackwell puneti(orme2, contig 640; N. puncti[onue3, contig 583; Athl, Scientific Publications, Boston, pp 413--440 BAB096931; Zm(l, AAF8996l.1; Ghil, Tl0797; Pt/Pa, Cohen MF, Wallis JG, Campbell EL, Meeks JC (1994) AAC78476.1; Ath2, AAC29067.1; Ath3, T51579; Ghi2, Transposon mutagenesis of Nostoc sp. strain ATCC AAD39534.2; Oictyostelium discoideum, AAF00200.1; Bacillus 29133, a filamentous cyanobacterium with multiple cei­ subtilis, 069769; Therrnoplasma acidophilum, CAC11626.1; lutar differentiation alternatives. Microbiology 140: Agrobacterium tumefaeiens, 139714; and Ferroplasma aeidarma­ 3233-3240 nus, contig 137. Curatti L, Porchia AC, Herrera-Estrella L, Salerno GL Multiple alignments for the creation of unrooted phylo­ (2000) A prok<:1ryotic sucrose synthase gene (susA) iso­ genetic trees were constructed with ClustalX (default set­ lated from a filamentous nitrogen-fixing cycl1lobacterium tings; Thompson, et aI., 1994). Sequences in which the UI, encodes a protein similar to those of plants. Planta 211: U2, U3, or U4 conserved regions were misaligned were 729-735 aligned manually. All trees were constructed with original, Delmer D (1999) Cellulose biosyntheis: exciting times for a gapped alignments. For NJ trees, 5,000 bootstrap tri(lls difficult field of study. Annu Rev Plant Physiol Mol BioI were performed and trees were constructed by the method 50: 245-276 of Saitou and Nei (1987). MP trees were created with PAUP Drews G,Weckesser J (1982) Function, structure and com­ 4.08 beta version (Sinaur Associates, Sunderland, MA) us­ position of cell walls and external layers. In NG Carr, BA ing a heuristic search. algorithm with 100 replicates. Boot­ Whitton, eds, The Biology of Cyanobacteria. Blackwell strap analysis was performed with maximum parsimony Scientific Publications, Boston, pp 333-357 as optimization criterion with resampling (100 replicates). QP Drummelsmith J, Whitfield C (2000) Translocation of group 1 capsular polysaccharide to the surface of Escl1e­ maximum likelihood topologies were constructed with Tree­ ridlia coli requires a mu1timeric complex in the outer Puzzle (bttp:llwww.tree-puzzle.de) using the variable lime membrane. EMBO J 19: 57-66 model of substitution (Muller and Villgron, 2000) and assum­ Ehling-Schulz M, Bilger W, Scherer S (1997) UV-E­ ing uniform rates of heterogeneity with 1,000 puzzling steps. induced synthesis of photoprotective pigments and ex­ Tree evaluations were made with the test of Kishino and tr(lcellular polysaccharides in the terrestrial cy(lnobacte­ Hasegawa (1989). Trees were drawn and edited with Tree­ rium Nostoc comnnme. J Bacteriol 179: 1940-1945 view (http://taxonomy.zoology.gla.ac.uk/rodl treeview. Frey-Wyssling A, Stecher H (1954) 471'55ben Fenball Des html). Nostoc-Schleimes. Z Zellforsch Mikrosk Anat Abt Histo­ chem 39: 515 Received June 25,2001; returned for revision July 2, 2001; Golden SS, Brusslan J, Haselkorn R (1987) Genetic engi­ accepted July 13, 2001. neering of the cyanobacterial chromosome. Methods En­ zymol153: 215-231 Han T, Runnegar B (1992) Megascopic eukaryotic algae LITERATURE CITED frol11 the 2.l-billion-year-old Negaunee iron-formation, Michigan. Science 257: 232-235 Altschul SF, Cish W, Miller W, Myers EW, Lipman OJ Hess K, Haller R, Katz JR (1928) Die Chemie der Zellulose (1990) Basic local alignment search tool. J Mol Bioi 215: und ihrer Begleiter. Ak<:1demische Verlagsgesellschafl, 403--410 Leipzig

540 Plant Physiol. Vol. 127, 2001 Cellulose in Cyanobacteria

Hirai A, Tsuji M, Yamamoto H, Horii F (1998) In situ and production of exopolysaccharides from representa­ crystalliza tion of bacterial cellulose: Ill. Influences of tive members of heterocystous and non-heterocystous different polymeric additives on the formation of micro­ cyano-bacteria. Phytochemistry 52: 639-647 fibrils as revealed by transmission electron microscopy. Okuda K, Li L, Kudlicka K, Kuga 5, Brown RM Jr (1993) Cellulose 5: 201-213 I3-Glucan synthesis in the cotton fiber: 1. Identification of Hoiczyk E (1998) Structural and biochemical analysis of the 13-1,4- and 13-1,3-g1ucans synthesized in vitro. Plan l sheath of Phormidiu m uncinatum. J Bacterial 180: Physioll01: 1131-1142 3923-3932 Okuda K, Tsekos I, Brown RM Jr (1994) Cellulose micro­ Hoiczyk E, Baumeister W (1998) The junctional pore com­ fibril assembly in Erythrocladia subintegra Rosenv.: an plex, a prokClryotic secretion orgClnelle, is the molecular ideal system for understanding the relationship between motor underlying gliding motility in cyanubacteria. Curr syntheSizing complexes (TCs) and microfibril crystalliza­ BioI 8: 1161-1168 tion. Protoplasma 180: 49~58 Holland N, Holland D, Helentjaris T, Dhugga K, Olsen GJ, Woese CR, Overbeek R (1994) The winds of Xoconostle-Cazares B, Delmer D (2000) A comparative (evolutionary) change: breathing new life into microbi­ analysis of the plant cellulose synthase (CesA) gene fam­ ily. PIClnt Physiol 123: 1313-1323 ology. J B,lcteriol 176: 1-6 Hoh T, Brown RM Jr (1984) The assembly of cellulose Phillipis R, Vincenzini M (1998) Exocellular Polysaccha­ microfibrils in Valonia macrophysa. Planta 160: 372-381 rides from cyanobacteria and their possible applications. Hoh T, O'Neil RM, Brown RM Jr (1984) Interference of cell FEMS Microbiol Rev 22: 151-175 wall regeneration of Boergesenia forbesii protuplasts by Rippka R (1988) Isolation and purification of cyanobacte­ Tinopal LPW, a fluorescent brightening Clgent. Proto­ ria. In L Packer, A Glazer, eds, Methods in Enzymology, plasma 123: 174-183 Vol 167. Academic Press, San Dil'go, pp 3-27 Kawaguchi T, Decho A (2000) Biochemical characteriza­ Roberts A, Roberts E, Delmer D (2000) The cellulose syn­ tiun of cyanobacterial extracellular polymers (EPS) from thase (CesA) genes of ferns, mosses, and algae. Poster, modern marine stromatolites (Bahamas). 30: 321-330 Year 2000 Annual Meetings, American Association of Krepinsky K, Plaumann M, Martin W, Schnarrenberger C Plant Biologists, San Diego (2001) Purification and cloning of chloroplast 6-phospho­ Roberts E (1991) Biosynthesis of cellulose II and related gluconate dehydrogenase from spinach. Eur f Biochem carbohydrates. PhD thesis. Univeristy of Tex,ls, Austin 268: 2678-2686 Ross P, Mayer R, Benziman M (1991) Cellulose biosynthe­ Kishino H, Hasegawa M (1989) Evaluation of the maxi­ sis and function in bacteria. Microbiol Rev 55: 35-58 mum likelihood estimate of the evolu tionary tree topol­ Rujan T, Martin W (2001) How many genes in Arabidopsis ogies from DNA sequence data, and the branching order come from cyanobacteria?: an estimate from 386 protein in Hoinoidea. J Mol Evol 29: 170-179 phylogenies. Trends Genet 17: 113-20 Knoll A (J 992) The early evolution of eukaryotes: a geo­ Saitou N, Nei M (1987) The neighbor joining method: a logical perspective. Science 256: 622-627 new method for reconstructing phylogenetiC trees. Mol Knoll A (1999) A new molecular window on early life. Science 285: 1025-1030 BioI Evol 4: 405-425 Kodani S, Ishida K, Murakami M (1999) Occurrence and Saxena 1M, Brown RM Jr, Fevre M, Geremia RA, Henris­ identification of UDP-N-acetylmuramyl-pentapeptide sat B (1995) Multidomain architecture of l3-glycosyl from the cyanobacteri urn Anabaena cylindrica. FEMS Lett transferases: implications for mechanism of action. J Bac­ 176: 321-325 teriol 177: 1419-1424 Martin W, Herrmann R (1998) Gene transfer from or­ Schopf JW, Walter MR (1982) Origin and eClrly evolution ganelles to the nucleus: how much, what happens, and of cyanobacteria: the geological evidence. In N Carr, B why? Plant Physiol118: 9-17 Whitton, eds, The Biology of Cyanobacteria, Vol 19. Martin W, Schnarrenberger C (J997) The evolution of Blackwell Scientific Publications, Boston the Calvin cycle from prokaryotic to eukaryotic chro­ Singh R (1954) Electron Micrographs of the Mucilage of the mosom,'s: a case study of functional redundancy in Blue-Green Algae 1. ScytonemG pseudogyanense. Rapport ancient pathways through endosymbiosis. Curt Genet Europees Congres Toegeposte Electronen-microscopie, 32: 1-18 Ghent, Belgium, pp 63-67 Martin W, Stoebe B, Goremykin V, Gansmann S, Hase­ Taylor NG, Scheible WR, Cutler S, Somerville CR, gawa M, Kowallik K (1998) Gene transfer to the nu­ Turner SR (1999) The irregular xylem3 locus of Arabi­ cleus and the evolution of chloroplasts. Nature 393: dopsis encudes a cellulose synthase required for second­ 162-165 Matthysse A (1983) Role of bacterial cellulose fibrils in ary cell wall synthesis. Plant Cell 11: 769-780 Agrobacterium tumefaciens infection. J Bacteriol 154: Thompson JD, Higgins DG, Gibson TJ (1994) CLUSTAL 906-915 W: improving the sensitivity of progressive multiple se­ Muller T, Vingron M (2000) Modeling amino acid replClce­ quen.ce alignment through sequence weighting, position­ ment. J Comput Bioi 7: 761776 specific gap penalties and weight matrix choice. Nucleic Nicolaus B, Panico A, Lama L, Romano I, Cristina M, De Acids Res 22: 4673-4680 Giulio A, Gambacorta A (1999) Chemical composition Tomme P, Driver DP, Amandoron EA, Miller RC Jr,

Plant PhY';iol. Vol. 127, 2001 541 Nobles et a!.

Antony R, Warren J, Kilburn DC (1995) Comparison of taus cyanobacterium with trichomes composed of ellip­ fungal (family I) and bacterial (family II) cellulose­ tical cells and containing poly-,B-IA-glucan cellulose. binding domain. J Bacteriol177: 4356-4363 J Gen Microbial 136: 1645-1653 Tuffery AA (1969) Light and electron microscopy of the Xiong J, Fischer WM, Inoue K, Nakahara M, Bauer CE sheath of a blue-green alga. J Gen Microbial 57: 41-50 (2000) Molecular evidence for the early evolution of pho­ Updegraff DM (1969) Semimicro determination of cellu­ tosynthesis. Science 289: 1724-1730 lose in biological materials. Anal Biochem 32: 420-424 Zogaj X, Nimtzv M, Rohde M, Bokram; W, Romling U Weeden NF (1981) Genetic and biochemical implications of (2001) The multicellular morphotypes of Salmonella typhi­ the endosymbiotic origin of the chloroplast. J Mol Evol murium and Escherichia coli produce cellulose as the sec­ 17: 133-139 ond component of the extracellular matrix. Mol Micro­ Winder B (1990) Crinalium epipsammum sp. nov.: a filamen­ bial 39: 1452-1463

542 Plant Physio!. Vo!. 127, 2001