Spirirestis rafaelensis gen. et sp. nov. (Cyanophyceae), A New Cyanobacterial Genus from Arid Soils

Valerie R. Flechtner, Sarah L. Boyer, Jeffrey R. Johansen

and Marisa L. DeNoble

Department of Biology, John Carroll University, University

Heights, OH 44118, U.S.A.

With 11 figures and 4 tables

Author for reprint requests (E-mail: Flechtner@jcu.edu)

Now in print: Nova Hedwigia 74:1-24. (February 2002)

ABSTRACT

A new cyanobacterial genus and species, Spirirestis rafaelensis, is described from soils of a semi-arid Utah juniper community in the San Rafael Swell, Utah, U.S.A. Multiple isolates of the organism have only been recovered from well-crusted, protected, and totally undisturbed soils at this site; it has not been recovered from any of the other 40 sites we have examined in the Sonoran, Mojave, Chihuahuan, Colorado Plateau, or Great Basin deserts during the last eight years. Spirirestis shares morphological characters with members of both the Scytonemataceae and Microchaetaceae, principally heterocyte formation, false branching, and presence of sheath. However, unlike the trichomes of all previously described genera in these families, most trichomes of Spirirestis are tightly spiraled. 16S rRNA sequence data suggest that Spirirestis is more closely related to members of the Microchaetaceae than to members of the Scytonemataceae or Rivulariaceae. The data also support the maintenance of Microchaetaceae and Scytonemataceae as separate families.

Key index words: 16S rRNA, cyanobacteria, phylogeny, microbiotic crusts, Scytonemataceae, Microchaetaceae, soil algae, Spirirestis, Utah.

INTRODUCTION

The classification of cyanobacteria is currently problematic. The traditional botanical classification scheme is most completely given in Geitler (1932), and serves as the basis for the later treatments of Desikachary (1959) and Starmach (1966). The botanical scheme is based primarily on morphological characters of cyanobacteria as observed in field samples. Bacteriologists have developed a taxonomic scheme that uses a polyphasic approach based upon physiological, molecular, ultrastructural, and morphological characteristics of cyanobacteria in clonal (often axenic) culture (Castenholz 1989, Castenholz & Waterbury 1989, Waterbury 1989, Waterbury & Rippka 1989). A modern revision of the botanical system has been completed for the genera (Anagnostidis & Komárek 1988, 1990, Komárek & Anagnostidis 1986, 1989), but many species still require further study to determine correct placement in the modern genera.

During the last decade, scientists have begun to employ molecular techniques to answer questions about cyanobacterial taxonomy. The rRNA operon, consisting of three rRNA molecules (16S, 23S, 5S) separated by internal transcribed spacer (ITS) regions (Fig. 8), has been a popular target for sequence analysis. Sequence determination of the entire 16S rRNA gene has provided insight into the phylogenetic relationships of genera within the different orders proposed by Komárek and Anagnostidis (e.g. Wilmotte et al. 1992, Nelissen et al. 1994, Turner 1997, Garcia-Pichel et al. 1998). Other investigators (e.g. Rudi et al. 1997) have focused on the three particularly variable regions (V6, V7 and V8) identified by Gray et al. (1984) as lying within bp 400-900 of Escherichia coli 16S rRNA to characterize and classify isolates of Nostoc, Anabaena, Aphanizomenon, Microcystis, and Planktothrix. Still other investigators (reviewed by Boyer et al. 2001) have focused their attention on the sequence of the 16S-23S ITS region to answer questions about taxonomy, phylogeny and population biology. The ITS region is particularly interesting because of the presence of sequences encoding structural genes for tRNA (tRNA^ala and tRNA^ile) and intervening sequence (IS) regions that do not encode structural products. It is within these latter sequences that the greatest freedom for sequence divergence exists.

We have recovered a highly unusual member of the Nostocales from adjacent plots in a single site in the San Rafael Swell region of eastern Utah. The soil of this region is well protected from disturbance due to its relative inaccessibility and has an extensively developed microbiotic crust. We have used a polyphasic approach that combines extensive morphological characterization with DNA sequence data from the 16S rRNA gene and the 16S-23S ITS region to determine the taxonomic position of our isolate in the Nostocales. This paper reports our morphological and molecular studies of this taxon, and describes it as Spirirestis rafaelensis gen. et sp. nov.

SITE DESCRIPTIONS

This semi-arid site is located in a nearly inaccessible perched shelf located below the Wedge Overlook area of the San Rafael Swell, Emery County, Utah (39º 06.066’ N latitude, 110º 44.713’ W longitude). The soil at the site is a reddish loam with very low organic matter (0.17% OM), slightly alkaline pH (7.6), electrical conductivity of 2.4m S^.cm^-1, sodium absorption ratio of 0.11, 2.9 ppm nitrate-N, 5.7 ppm phosphate-P, 134 ppm exchangeable K, 613 ppm soluble Ca, 39 ppm soluble Mg, and 55 ppm soluble Na (based on analyses conducted by the Soil Analysis Laboratory, Brigham Young University). Soils are very shallow (0.2-1.5 m to slickrock base) and support a sparse vascular plant community dominated by widely spaced Juniperus osteosperma. The soil is covered with a well-developed pedicled microbiotic crust. This site is located in the Colorado Plateau, a large geological province that supports microbiotic crusts in much of its range. Additional information on the site is available at http://www.jcu.edu/mcp. It was sampled on 19 May 1995 and again on 22 May 1996.

This arid hot desert site is located near the Montana Mine in an upland bajada site in the Paradise Range, San Bernardino County, California (35º 13.730’ N latitude, 116º 49.509’ W longitude). The soil at the site is a gravelly silt loam with a surface covering of gravels and coarse sands. It is low in organic matter (0.7% OM), neutral (pH 7.12), and low in electrical conductivity (0.5 m S^.cm^-1). It has been characterized thoroughly in another paper (Johansen et al. 2001) under the site name PR1. This typical Mojave Desert site is dominated by the following shrubs: Ambrosia dumosa, Atriplex canescens, Larrea tridentata, Krameria erecta, Lycium andersonii, Mirabilis bigelowii, Haplopappus cooperi, Xylorhiza tortifolia, Encelia actonii, and Yucca brevifolia. The soil has a very patchy microbiotic crust cover, and the samples used for isolation of strains were essentially uncrusted. The site was sampled on 23 May 1998.

This arid hot desert site is located in the Sevilleta Long-Term Ecological Research Project area in Soccoro County, New Mexico (34^o 19.462' N latitude, 106^o 0.283' W longitude). The soil is a sandy loam with low organic matter (0.51%), slightly alkaline pH (7.45), and low electrical conductivity (0.5 m S^.cm^-1). This site is located in the Chihuahuan Desert near the intersection of several biomes. Additional information on the specific site (SEV2) is available at http://www.jcu.edu/mcp, with general information on the Sevilleta available at http://sevilleta.unm.edu.

This site is in Meigs Creek, a tributary of the Little River in the Great Smoky Mountains National Park, Tennessee. This stream is bordered by a mixed mesic hardwood forest and flows through the Thunderhead Sandstone Formation. It was sampled 9 April 1999. The specific site was located on the county line between Blount and Sevier Counties, and was also just at the point where the stream flows from the Thunderhead Sandstone into the Metcalf Phylite Formation. The site was at 530 m above sea level, 35º 39.739' North latitude, 83º 39.773' West longitude.

MATERIALS AND METHODS

Methods

Isolation and Characterization of strains. Soil samples from the three desert sites were composites of ten 5 gram samples from systematically placed quadrats within each site. Soil chemistry samples were also composites of ten subsites, but were collected separately. Soil chemistry was determined by the Soil Analysis Laboratory at Brigham Young University, Provo, Utah. The aquatic sample from Meigs Creek was a composite scraping of attached algae in rapidly flowing water in the stream.

Cyanobacterial isolates used in this study were all collected, isolated, and identified by Valerie Flechtner and Jeff Johansen. For those cyanobacteria isolated from soils, dry soil samples were crushed, subsampled, diluted, and plated on Bold's basal (Bold & Wynne 1978) and Z-8 media (Carmichael 1986) as described in Flechtner et al. (1998). Cyanobacteria were isolated into unialgal culture from the plates and kept on agar slants of Z-8 medium. Coleodesmium wrangelii was isolated directly from a stream sample onto agar slants of Z-8 medium. All isolates were examined on Olympus photomicroscopes with Nomarski DIC optics. Strains were kept in dim light (<50 µE·cm^-2·s^-1 illuminance) at 7ºC on a 12:12 hr light:dark cycle.

All isolates were examined on Olympus photomicroscopes with Nomarski DIC optics. All strains were photographed, characterized, and measured on at least two occasions to document morphology and identity of the species.

Spirirestis received additional characterization. Cultures of two isolates, SRS6 and SRS70 (isolated by Valerie Flechtner from the 1995 samples) were examined at 3 weeks, 9 weeks, 3 months, and 6 months after transfer to fresh BBM, Z-8, and nitrogen-free Z-8 agar-solidified media. Although many clonal isolates of Spirirestis were identified from initial platings in 17 different soil samples, only two (SRS 6 and SRS 70) were characterized thoroughly. In addition to examining cultures at time intervals indicated above, statistical data on each of the two clonal isolates were collected from 4-week-old cultures on BBM and nitrogen free Z-8 medium. Filament width, trichome width, cell length, and cell length to width ratios were determined for 50 individual cells (Table 1). The species was also observed in soils under epifluorescence microscopy immediately following moistening of the soil. The filaments were identical in size and coiling in these uncultured samples, but were not extensively characterized due to their rarity.

DNA was extracted from 20 mg of fresh unialgal cultures using the Cullings (1992) modification of the Doyle and Doyle technique (Doyle & Doyle 1987). The resultant DNA was suspended in 50 µL TE and stored at –20ºC.

Primers were designed after Wilmotte et al. (1993) and Nübel et al. (1997). They were designated:

Primer 1 5’ CTC TGT GTG CCT AGG TAT CC 3’ (after Wilmotte et al 1993)

Primer 2 5’ GGG GGA TTT TCC GCA ATG GG 3’ (after Nübel et al. 1997)

Primer 5 5’ TGT ACA CAC CGG CCC GTC 3’ (after Wilmotte et al. 1993)

Primer 6 5’ GAC GGG CCG GTG TGT ACA 3’ (after Wilmotte et al. 1993)

Primer 7 5’ AAT GGG ATT AGA TAC CCC AGT AGT C 3’ (after Nübel et al.

1997)

Primer 8 5' AAG GAG GTG ATC CAG CCA CA 3' (after Wilmotte et al. 1993)

The position of these primers with regard to the 16S RNA gene, the 23S RNA gene, and the transfer RNA genes which had previously been found between them in cyanobacteria is shown in Fig. 8. Primer 6 is the reverse complement of the original Wilmotte (1993) primer (here designated Primer 5). Primer 8 was modified by one base pair from Wilmotte’s WAW1486R in accordance with initial data. Primers were ordered from the Midland Certified Reagent Company in concentrations that we brought to 100 µM. For use in PCR, a mix of 1.2 µL each of two primers and 7.6 µL sterile water was made.

Each DNA sample was amplified using primers 1 and 2. This resulted in a chain approximately 1600 bp long (long PCR) which was then used as a template for a reamplification using primer pairs 1 and 5, 2 and 6, and 7 and 8. Primer pairs 1 and 5 and 7 and 8 amplified sequences in the 16S rRNA gene; primer pair 2 and 5 reamplified 16S-23S ITS sequences. Reamplification products from the long PCR product were subsequently cloned prior to sequence determination (see below).

Each 100 µL reaction contained 86 µL sterile water, 10µL 10x buffer (Promega), 0.5 µL of each dNTP (G, A, T, C) at 10 mM, 0.5 µL of the primer mixture described above, 0.5 µL Taq polymerase (Promega), and, typically, 1.0 µL template DNA (either genomic DNA or PCR product).

The most commonly used profile for the initial 16S + ITS PCR reaction using primers 1 and 2 was 94ºC for 1 min, 57ºC for 1 min, 72ºC for 4 min (35 cycles), followed by a 10 minute extension at 72ºC. For PCR reamplifications, the most commonly used profile was 94ºC for 1 min, 56ºC for 45 s, 72ºC for 2 min (20 cycles). Reactions were carried out using Thermolyne’s Amplitron and Temptronic thermocyclers. Results were checked using a 1% agarose gel.

PCR product was cloned into plasmids containing the sites for the universal primers M13 forward and reverse on either side of the cloning site using Invitrogen’s TOPO^® TA Cloning Kit for Sequencing, Version A. Plasmid DNA was obtained from, generally, three of the resultant clones using Qiagen’s QiaPrep Spin Kit.

Automatic sequencing with the universal primers M13 forward and reverse was performed by Cleveland Genomics.

Forward and reverse primer sequences were checked against each other by generating the reverse complement of the "reverse" sequence with Oxford Molecular Group’s Omiga™ and aligning it with the "forward" sequence with CLUSTAL W Multiple Sequence Alignment Program, version 1.7 (Thompson et al. 1994) via the Baylor College of Medicine’s Search Launcher (Smith et al. 1996) at http://dot.imgen.bcm.tmc.edu:9331/. This resulted in the longest possible read of the sequence, in addition to acting as a check on the sequencing. Where forward and reverse sequences did not agree with each other or with published sequences, the sequence data were carefully checked against the sequence chromatograms. Sequences from amplifications using the three different primer pairs were joined using alignments of overlaps between the sequences. All sequences were submitted to GenBank. The accession numbers are: Spirirestis rafaelensis - ITS: AF236659, 16S: AF334690, AF334691, AF334692; Coleodesmium wrangelii -- ITS: AF236652, 16S: AFAF334701, AF334702, AF334703; Tolypothrix distorta - ITS AY007689, 16S: AF334693, AF334694, AF334695; Calothrix parietina - ITS: AF236642, AF236643, 16S: AF334695, AF334696, AF334697; Scytonema hyalinum - ITS: AF236650, AY007688, AY007689, 16S: AF334698, AF334699, AF334700.

Sequences from different clones and isolates were aligned using CLUSTAL W. These alignments were checked by eye, and again, when sequences did not agree with each other or published sequences, the sequence data were carefully checked against the sequence chromatograms. The variability among sequences from different clones was preserved in the analysis through the inclusion of multiple non-identical sequences.

In all analyses, we compared our 16S sequences to Nostoc sp. strain 152 (GenBank AJ133161), Nostoc ATCC53789 (GenBank AF062638), Nostoc sp. TDI#AR94 (GenBankAF027653), Anabaena sp. strain PCC 7108 (GenBank AJ133162), Anabaena variabilis Kützing ex Bornet et Flahault (GenBank AB016520), Scytonema hofmanni Agardh ex Bornet et Flahault (GenBank AF132781), Calothrix sp. PCC7714 (GenBank AJ133164), and Calothrix D253 (GenBank X99213).

Phylogenetic trees based on 16S sequence data were constructed by a variety of methods using PAUP 4.2 (Swofford 1998). Methods used include: neighbor-joining using logdet, HKY85, and Jukes-Cantor distance metrics and assuming both equal and gamma substitution rates, maximum parsimony using the logdet distance metric and assuming equal substitution rates, and maximum likelihood using the HKY85 distance metric and assuming equal substitution rates. Bootstraps were performed with 1000 replicates on all trees.

We describe Spirirestis rafaelensis following the requirements of the International Code of Botanical Nomenclature (Greuter et al. 2000). However, we have also followed recommendations to meet the minimal requirements for the Bacteriological Code by depositing a clonal isolate in a recognized culture collection (in this case, the UTEX Culture Collection), and upon publication submitting the name and relevant reprint to the International Journal of Systematic Bacteriology (Castenholz & Waterbury 1989, Lapage 1992, Murray 1996). Construction of Latin diagnoses and descriptions follows the format suggested in Stearn (1992).

Type materials were prepared in several ways. The holotype was prepared by filtering a young, healthy culture of Spirirestis onto Whatman filter paper. The paper was allowed to air dry and then attached to lichen herbarium cardstock, and placed in herbarium envelopes. Two isotypes were additionally prepared in this way. Three vials containing Spirirestis preserved in 2% glutaraldehyde were also prepared, and these are considered isotypes as well. Finally, a feral soil sample from the site was curated with the above materials at the Herbarium of Non-Vascular Plants at the Monte L. Bean Museum, Brigham Young University, Provo, Utah. In addition to these materials, cultures of Spirirestis rafaelensis were deposited in the UTEX culture collection at the University of Texas, Austin, Texas.

RESULTS

Spirirestis Flechtner et Johansen gen. nov.

A Scytonema et Tolypothrix filis regulatim arcte torsivis in helices dextrales differt.

Fila libra, regulatim arcte torsiva. Rami falsi duplices vel simplices. Vaginae firmae, tenues vel crassae, interdum lamellatae. Trichomata solitaria intra vaginam. Cellulae breviores quam latae. Heterocyteae intercalares vel basilares, sphaericae ad ovales vel adpressae. Akineta non observata.

Differs from Scytonema and Tolypothrix in that the filaments are regularly, tightly coiled in a right-handed helix.

Filaments free, forming tight, regular spirals. False branches double or single. Sheaths firm, thin to thick, sometimes lamellated. Trichomes solitary within the sheath. Cells shorter than broad. Heterocytes both intercalary and basal, spherical to oval or appressed. Akinetes not observed.

Spirirestis rafaelensis Flechtner et Johansen sp. nov. (Figs 1-4, Figs 5-7)

Colonia molliter acervata in agaro, marginibus leviter irregularibus, non caespitosa, atroveneta, remantes atroveneta ad menses quatuor; olivacea vel ochracea ad mortem cellulosum in culturis vetustioribus siccis. Colonia compacta, filis dense dispositis, non mucilagina. Fila non horizontaliter vel verticaliter extensa e colonia etsi visibiles intra coloniam. Morphologia colonialis similaris in ambobus agaro basali Boldii et agaro Z-8.

Fila spiras arctas regulares formantia, 6-20 m m diametro. Spirae leviter decrescentes in culturis novis, 16-28 m m diametro ad partem latissimam, 12-16 m m diametro ad extrema angustata spirarum, distantia inter spiras regulares 8-12 (-16) m m. Fila interdum in spiras laxiores contorta vel etiam extensa, haec in culturis vetustioribus frequentiora. Rami falsi duplices vel simplices, in spiras laxas vel in fila extensa praesentes, non observati in spiras arctas. Rami falsi plerumque ab heterocyteis distantes. Vagina firma plerumque incolorata, interdum succinea vel umbrina, in culturis novis tenuis, in culturis vetustioribus crassiorescens, 1-4 m m lata. Trichomata ad septa haud constrictae ad leviter constrictas, in culturis vetustioribus et in filis laxiter spiralibus vel extensis frequenticus constrictae, 5.5-14.0 m m diametro. Cellulae plerumque breviores quam longae, infrequenter quadratae, 2.0-7.2 m m longae, sine aerotopis, contentis interdum granularibus

praecipue in culturis vetustioribus. Heterocyteae intra filas intercalares vel baseless, pro parte maxima simplices, in positione intercalari infrequenter duplices, raro triplices, sphaericae ad ovales, interdum in positione basali adpressae in latere uno, brunnescentes, 8-9 m m diametro, 4-9 m m longae.

Typus SRS70 die 19. Maji 1995 a solo deserti, lat. bor. 39^o 06.066', long. occ. 110^o 44.713', Prospectus Cunei, Sanctus Raphael Unda, Utah, USA. Holotypus: BRY C 35162, Herbarium Cryptogamorum Nonvascularium, Brigham Young University, Provo, Utah. Isotypi: BRY C 35162, Herbarium Cryptogamorum Nonvascularium, Brigham Young University, Provo, Utah. Isotypus in statu vivo: ZZ-24, UTEX Congeries Culturarum, University of Texas, Austin, Texas.

Colony softly mounded on agar, with slightly irregular margins, not tufted, dark blue-green, maintaining its colour at four months; the only evidence of colour change to olive or ochre occurring with cell death in drying, older cultures. Colony compact, with densely arranged filaments, but no evidence of external mucilage production. Although filaments can be seen within the colony at 30 X magnification, filaments do not extend horizontally or vertically from the colony. Colony morphology is similar on both Bold's Basal and Z-8 media.

Filaments forming tight, regular spirals which can be seen to taper slightly in young cultures. Filaments 6-20 m m in diameter; spirals 16-28 m m in diameter at their widest point, tapered end of spirals about 75-80% of the maximum diameter; distance between regular spirals 8-12 (-16) m m. Some filaments more loosely coiled, or even extended; these filaments increasing

in frequency in older cultures. Both single and double false branching occasionally observed in more loosely coiled or extended filaments, but not in tightly spiraled filaments. False branching usually distant from the heterocytes. Sheath firm, usually colourless, occasionally golden to brown, thin in young cultures becoming thicker and lamellated in older cultures, 1-4 m m thick. Trichomes not to slightly constricted at the crosswalls, more commonly constricted in older cultures and in loosely coiled or extended filaments, 5.5-14.0 m m in diameter. Cells mostly shorter than wide, occasionally quadratic, 2.0-7.2 m m long, without aerotopes, cell contents sometimes granular, especially in older cultures. Heterocytes both intercalary and basal within filaments, mostly single but occasionally double in intercalary position, rarely triple, spherical to oval, sometimes appressed on one side in basal position, light brown, 8-9 m m in diameter, 4-9 m m long.

Type SRS70 collected on 19 May 1995 from desert soil surface, 39^o 06.066' N latitude, 110^o 44.713' W longitude, hanging plateau below Wedge Overlook, San Rafael Swell, Emery County, Utah, USA. Holotype: BRY C 35162, Herbarium of Nonvascular Cryptogams, Brigham Young University, Provo, Utah, USA. Isotypes: BRY C 35162, Herbarium of Nonvascular Cryptogams, Brigham Young University, Provo, Utah, USA. Living isotype: ZZ-24, UTEX Culture Collection, University of Texas, Austin, Texas, USA.

Spirirestis means "spiral rope", chosen to indicate the key diagnostic feature of the genus, and indicate its thicker nature than Spirulina (spiral thread).

Morphological characters of cultured Spirirestis demonstrated some variability, being affected by both medium composition and culture age. Although tightly coiled filaments is a consistent character of Spirirestis on all media (Figs 1 & 2), filaments are typically more tightly coiled in cultures grown on Z-8 medium than in cultures grown on Bold's Basal Medium (BBM) and in younger cultures compared to older cultures (Figs 3 & 4). Colonies grown on BBM are more olive in colour than the typically blue-green colonies observed on Z-8 medium due to yellowing of the sheath materials and colonies become more olive coloured on both media with the passage of time. Filaments of cultures grown on BBM are wider than those grown on Z-8 both because the trichomes were wider (Table 1) and because more sheath material is produced. It is interesting to note that the differences between clones were very minor, and in most cases statistically insignificant. We feel that differences seen between media were primarily due to the fact that cultures are healthier on Z-8 (even when nitrogen free) than on Bold's Basal Medium.

Although heterocytes are present in tightly coiled filaments (Fig. 2), they become much more evident in relaxed filaments (Fig. 4). In cultures growing on nitrogen-free Z-8 plates, heterocyte formation is abundant and conspicuous even in tightly coiled specimens (Fig. 7). Hormogonia formation is also more prevalent in N-free Z-8, likely due to frequent fragmentation at the heterocyte. False branching is not often observed in tightly coiled filaments. The slightly constricted walls seen towards the ends of the trichomes are much more evident in loosely coiled filaments (cf. Fig. 5). The diameter of the coil was greater at the basal end than at the apical end (Fig. 1 & Fig. 6).

Phylogenetic Analysis of Spirirestis

Sequence data for a partial sequence of the 16S rRNA gene and the associated 16S-23S Internal Transcribed Spacer (ITS) region were collected for each of the following: Spirirestis rafaelensis (SRS70), Calothrix parietina Thuret ex Bornet et Flahault, Coleodesmium wrangelii Borzi ex Geitler, Scytonema hyalinum Gardner, and Tolypothrix distorta Kützing ex Bornet et Falahault. In addition, nine 16S rRNA sequences of other members of the Nostocales published on GenBank were also aligned and incorporated in cladistic analyses. Since this paper gives the first report of molecular sequence data for the strains which we isolated, we here document the morphology of these strains. Following the morphological characterization of the strains, the phylogenetic analyses based on the full molecular data set for the 14 nostocalean strains is given.

Calothrix parietina Thuret ex Bornet et Flahault.

Strain data: SRS-BG14, isolated by Valerie Flechtner from a May 1995 sample from the Wedge Overlook of the San Rafael Swell. Thallus flat and spreading, vivid blue-green, becoming orange in senescence. Sheath thin and colourless in recently transferred cultures, becoming lamellated and yellow with age. Trichomes heteropolar, with basal heterocytes and tapering apices, not constricted at the crosswalls at the base and mid-filament, constricted at the crosswalls near the apices, with single false branching, 7.5-10.0 µm wide at the base, to 3.2-5.0 µm wide at the apex. Cells 3-5(7) µm long. Heterocytes, shorter than wide, 5-10 µm wide.

Coleodesmium wrangelii Borzi ex Geitler

Strain data: MC-JRJ1, isolated by Jeff Johansen from a sample from Meigs Creek, Great Smoky Mountains National Park, Tennessee, collected 9 April 1999. Thallus a branching filament up to 20 µm wide with numerous trichomes in a common sheath. Trichomes slightly constricted at the crosswalls, 11-13 µm at the base and middle, tapering down to 8 µm at the ends. Vegetative cells 3.3-5.0 µm long. Heterocytes basal, hemispherical to elongated and rounded, 12-14 µm wide, 10-22 µm long.

Scytonema hyalinum Gardner

Strain data: FI-8A, isolated by Valerie Flechtner from a soil sample from the Paradise Range, California. Colony adherent to agar. Filaments uniseriate, with both single and double false branching, 11-14 µm wide. Sheath clear, thin to roughened. Trichomes twisted, unconstricted at the crosswalls, often with a meristematic region near the end, 8-12 µm wide. Cells shorter than broad to quadratic. Some end cells with a cap in older cultures. Heterocytes 9.0 µm wide, 6.4 µm long.

Tolypothrix distorta Kützing ex Bornet et Flahault

Strain data: SEV2-52G, isolated by Valerie Flechtner from a soil sample from the Sevilleta Long-Term Ecology Research Project Area, New Mexico. Colony slightly olive and cushion-shaped with grassy uprights. Sheath clear, somewhat lamellated. Trichomes straight, unconstricted to slightly constricted at crosswalls. Single branching at heterocyte; branches long. Necridia and hormogonia evident. Filament 9 to 13 µm wide, usually thin. Trichomes 7.5-12.0 µm wide; cells shorter than wide (2.4-6.4 µm wide). Heterocytes oblong to spherical, 8 to 10 µm long, 4 to 10 µm wide; singles or in pairs.

16S rRNA sequence data for all strains was obtained by amplifying a 1400 bp portion of the rRNA operon beginning about 400 bp in downstream from the 5' end of the 16S rRNA gene and extending approximately 40 bp into the 23S gene. It encompasses the V₆, V₇ and V₈ regions of 16S rRNA gene as well as the entire 16S-23S ITS region (long PCR product). When we examined the 16S-23S ITS regions from each of the strains obtained by using primers 1 and 5 with the long PCR product (Fig. 8) we found that while three taxa (S. rafaelensis, T. distorta, and C. wrangelii) exhibit a single band on 1% agarose gels, the remaining two taxa (S .hyalinum and C. parietina) have multiple bands (Fig. 9). Given these results, we decided to clone all PCR products prior to sequence analysis.

Sequence Data from the 16S rRNA Operon.

For Calothrix parietina, Coleodesmium wrangelii, Scytonema hyalinum and Spirirestis rafaelensis, 16S rRNA sequence data were obtained from three individual clones of PCR-amplified 16S rRNA sequences. Sequences were obtained for two clones of PCR-amplified T. distorta 16S rRNA genes. We detected microheterogenities among 16S sequence determined from individual clones prepared from a single PCR amplification in all taxa examined. However, the between-genera differences (Table 2) were always greater than the within-genus variability (Table 3).

Phylogenetic analyses of the 16S gene from our five strains and the nine GenBank strains gave trees with consistent topologies across methods when Nostoc and Anabaena were used as outgroups. Three major clades were revealed (Fig. 10). Clade 1 includes three taxa (S. rafaelensis, T. distorta, and C. wrangelii) representing two established genera and one new genus of the family Microchaetaceae. Clade 2 includes three taxa of the genus Calothrix (family Rivulariaceae). Clade 3 includes two taxa of the genus Scytonema (family Scytonemataceae). Differences among the trees generated involved the placement of taxa within the Calothrix clade and the placement of the Nostoc and Anabaena taxa.

Because of the generally conservative nature of the 16S rRNA sequences, several researchers have proposed that sequence comparisons of the ITS region separating the 16S and 23S rRNA genes may be useful in refining our understanding of the phylogenetic relationships among closely related taxa. Our ITS sequence data from the five taxa considered in this study provide some additional insights into the similarity of the rRNA operons of these taxa. Two distinct sequence organizations were identified in the ITS regions of these organisms (Fig. 11). Some ITS regions contained the structural genes for tRNA^ile and tRNA^ala (pattern 1); other ITS regions lacked these genes (pattern 2). Pattern 1 ITS sequences were found in all five taxa while pattern two was found only in C. parietina and S. hyalinum.

Comparisons of the sequence composition of the pattern 1 ITS regions from the five taxa reveal that even when the genes for two tRNA's are present, significant differences in sequence arrangement can still exist (Fig. 11). The ITS regions from S. rafaelensis, C. wrangelii, and T. distorta are similar in arrangement. The ITS region of S. hyalinum is generally similar, but has a larger IS3 compared to the Clade 1 organisms. The ITS regions of C. parietina is most dissimilar; the IS2 is extremely small (9 bp).

The two pattern 2 ITS regions recovered from clones of S. hyalinum were identical in size (400 bp) but not in sequence (Boyer et al. 2001). A single pattern 2 ITS region of 329 bp was identified in a C. parietina clone. The presence of different kinds of ITS regions in a single organism has been observed in other cyanobacteria. Two groups (Iteman et al. 2000; Li 2000) have reported pattern 1 and pattern 2 ITS regions in Nostoc. We have made similar observations in Microcoleus (unpublished data).

Specific areas of the ITS region have been identified as critical to the proper folding of the primary transcript for release of structural ribosomal and transfer RNA's during processing. One of these areas, D1, occurs at the very 5' end of the 16S rRNA gene, a region sequenced in this study. Alignment of D1 sequences of the five field species examined in this study reveals that members of the Microchaetaceae all share a common sequence that differs in one base pair from the sequence seen in members of the Scytonemataceae and Rivulariaceae (Table 4).

The possession of morphological characteristics intermediate between the two genera Scytonema and Tolypothrix, and thus intermediate between the two families recognized by Komárek & Anagnostidis (1989), suggests that aside from the coiling, Spirirestis is distinct from both Scytonema and Tolypothrix if the whole suite of other morphological characters is considered. Our work suggests that the two families represent different monophyletic assemblages. Based on the frequent heteropolar nature of Spirirestis and on our molecular data, we are placing Spirirestis in the Microchaetaceae sensu Komárek & Anagnostidis (1989).

DISCUSSION

Justification for taxonomic recognition

Spirirestis shows clear phenotypic divergence from other heterocyteous genera, but only small genotypic separation from closely related genera in the Microchaetaceae based upon the 16S rRNA gene. We are aware that some workers would argue that without genetic separation in the 16S rRNA, Spirirestis rafaelensis should not even be recognized at the species level, let alone as a new genus. It seems necessary, therefore, to discuss at this point species concepts within cyanobacteria and other prokaryotes, systematic practice in broader taxonomic groups, and our reasons for choosing to recognize this possibly endemic species in its own genus.

Castenholz (1992) recognized a pragmatic species concept for cyanobacteria; i.e. a species is a cluster of similar strains that have recognizable discontinuities with other known clusters. As stated, this concept appears to be a phenetic construct, since individuals are assigned to species based on similarity to one another, and separated out based upon distance from other clusters. Castenholz & Waterbury (1989) recommended using a polyphasic approach, i.e. the use of multiple character sets (morphology, ultrastructure, DNA-DNA hybridization, 16S rRNA sequence data, %GC, pigmentation and other biochemical characters, physiological characters, etc.). This polyphasic but essentially phenetic approach is very similar to the phylogenetic species definitions proposed by Eldredge & Cracraft (1980), Nelson & Platnick (1981), and Nixon & Wheeler (1990). These concepts were unified by Wheeler & Platnick (2000) into their definition of a phylogenetic species, which for asexual taxa is "the smallest aggregation of.…lineages diagnosable by a unique combination of character states." They further state that apomorphic character states do not exist within species, and that the concept of monophyly is inapplicable at this taxonomic level.

Spirirestis is diagnosed from all other Microchaetaceae, Scytonemataceae, and Rivulariaceae by the autapomorphic trait of regular tight spiraling of the trichome that is due to cellular arrangement rather than constriction within the sheath. It is defined within the cyanobacteria by the unique combination of character states that includes regularly spiraling trichomes, heterocyte production, presence of false branching, and absence of true branching. Thus, it fits the criteria of Wheeler & Platnick (2000) to be recognized at least on the species level. It also fits the criteria for the evolutionary species concept (Wiley & Mayden 2000) because it maintains its identity through time and over space, and the phylogenetic species concept of Mishler & Theriot (2000a) because of the evidence for both apomorphy and monophyly. The taxon satisfies typological and morphological species constructs as well (Cracraft 2000). The only species concept it may not comply with strictly is Castenholz’s cyanobacterial species concept. A single population likely may not comprise a "cluster of similar forms". Castenholz’s concept could be used to refute the recognition of true endemic species within the cyanobacteria, and would certainly argue against endemic genera. While we are unsure Spirirestis is endemic, one of us has studied North American desert soil algae for over 25 years, and recently we have among us examined the soils of over 70 sites scattered throughout the western United States and Mexico. If not endemic, it is certainly very rare. We consider the recognition of Spirirestis rafaelensis as a new species to be compliant with the majority of theoretical systematic constructs directly applicable to asexual species.

However, the molecular data collected in this study do not provide much evidence for the separation of Spirirestis rafaelensis from Tolypothrix distorta, let alone for the separation of Spirirestis from the other members of the Microchaetaceae for which sequence data exist. The Ad Hoc Committee on Reconciliation of Approaches to Bacterial Systematics (Wayne et al. 1987) recommended the use of DNA-DNA hybridization studies to define prokaryotic species. They stated "the phylogenetic definition of a species generally would include strains with approximately 70% or greater DNA-DNA relatedness and with 5° C or less D T_m. Both values must be considered." They further stated that phenotypic characteristics should agree with this definition, but would be allowed to override this phylogenetic species concept in exceptional cases. They recommended that genospecies defined by their genetic criteria should not be named until they can be differentiated by some phenotypic property. They also remarked that the depth in an RNA dendrogram at which a given hierarchical line separating species is to be drawn may vary along the different major branches due to different ages in the branches.

DNA-DNA hybridization between species is only possible with axenic strains, and is much less commonly performed than sequence analysis of the 16S rRNA gene. Stackebrandt & Goebel (1994) correlated DNA-DNA reassociation values and 16S rRNA sequence similarity for 54 pairwise comparisons of bacteria strains. They found that the relationship was not linear, but that species with 70% or greater DNA similarity usually have more than 97% 16S rRNA sequence similarity. Thus if sequence similarity is below 97%, one can be fairly confident that DNA-DNA reassociation is below 70%, and thus the two strains being compared should be recognized as separate species if any phenotypic separation exists. However, from their data presentation, it is clearly not possible to use 16S rRNA sequence similarity above 97% as a rationale for placing two strains in the same species. Thirty-five of the 54 pairwise comparisons showed taxa with 16S rRNA sequence similarity above 97%, but DNA-DNA reassociation values below 70%. Indeed, a number of strains were over 99% similar in their 16S rRNA sequences, but had less than 50% DNA-DNA reassociation. In cases where 16S rRNA sequence data are more similar than 97%, the sequence data simply are uninformative, and other character sets must be examined to separate or combine the two strains being compared. Spirirestis rafaelensis and Tolypothrix distorta are two such forms, and we have separated them based on their morphology.

Separating genera is more problematic than separation of species, as fewer criteria have been put forth. Even though genera form the essential basis of bacterial systematics, Wayne et al. (1987) indicated that there was no satisfactory phylogenetic definition of a genus at the time of their writing. The phylogenetic species concept of Wheeler & Platnick (2000) is not applicable to genera and families (Mishler & Theriot 2000b). The only species concept which can be extended in principal to higher taxa is the phylogenetic taxonomic concept of Mishler & Theriot (2000a, b). By this standard, groups of species demonstrating evidence of monophyly through the existence of shared apomorphic characters would be considered separate genera. With a clear morphological autapomorphy (regular tight spiraling of trichomes) indicating monophyly in the population of forms present in the soils of the Wedge Overlook in the San Rafael Swell, this concept supports the recognition of Spirirestis as a distinct genus. Further support for recognition of Spirirestis as a genus separate from Tolypothrix can be found in the recommendations with regards to generic delineation by Anagnostidis & Komárek (1985). They recommend recognition of small, unambiguously defined genera, and indicate "no particular purpose is served by maintaining large genera and subdividing them into subgenera, sections, subsections, etc." Given the genetic and morphological similarity of Spirirestis, Tolypothrix and Coleodesmium, it seems our choices would be to expand Tolypothrix to include both other genera (and possibly Microchaete as well) or recognize all as small, unambiguously defined genera in the Microchaetaceae. We prefer the latter choice.

Biogeography of Spirirestis

It is interesting to note that Spirirestis, although fairly common at the collection site in the San Raphael Swell, is absent from all other soils we have examined in the arid and semiarid western United States. These localities include semi-arid shrub-steppe in the Great Basin Desert (Johansen & St. Clair 1986, Johansen et al. 1984, unpublished studies), Tintic Mountains (Johansen & Rushforth 1985), Colorado Plateau (Johansen et al.1981), and San Rafael Swell (unpublished studies), as well as the hotter desert sites in the Sonoran Desert (Cameron 1960, unpublished studies), Mojave Desert (unpublished studies) and Chihuahuan Desert (unpublished studies). What is especially remarkable is that we have been unable to detect Spirirestis in soils of very similar texture, organic matter, and chemistry immediately above the study site. The major difference between the two adjacent locales is that the upper site is periodically grazed by cattle and disturbed by hikers and vehicles, while the lower site is the most pristine, undisturbed crust community we have seen. Thus, Spirirestis may be an indicator of absence of trampling disturbance in the reddish, silty soils of semi-arid regions of the Intermountain West.

There are two schools of thought concerning the taxonomic placement of the genera Calothrix, Scytonema, and Tolypothrix. Using the polyphasic approach preferred by bacteriologists, Castenholz (1989) recognizes three families in the order Nostocales (Nostocaceae, Scytonemataceae and Rivulariaceae) and subsume Tolypothrix into the genus Scytonema, the sole genus listed in the family Scytonemataceae. They argue that because culture medium can influence the pattern of false branching seen, this character cannot be used reliably to separate the genera and that "…there may not be a clear genetic boundary between Scytonema and Tolypothrix." Under this scheme, the taxa we identify as S. hyalinum, T. distorta, C. wrangelii, and our new isolate S. rafaelensis would all be included in the genus Scytonema of the family Scytonemataceae and Calothrix parietina would fall in a separate family Rivulariaceae. Komárek & Anagnostidis (1989), relying on the botanical scheme that emphasizes cell and colony morphology, recognized four families (Microchaetaceae, Nostocaceae, Scytonemataceae and Rivulariaceae). In this scheme, the genera Tolypothrix, Coleodesmium and Spirirestis would all be members of the Microchaetaceae rather than the family Scytonemataceae.

This paper reports the first ribosomal rRNA sequence data from members of the family Microchaetaceae. Our molecular data support the taxonomic scheme of Komárek and Anagnostidis. We offer the following arguments supporting the placement of Tolypothrix, Coleodesmium and Spirirestis in a family distinct from the Scytonemataceae. (1) The phylogenetic tree constructed from sequence data representing 1000 bp of the 16S rRNA gene shows three distinct clades (Fig. 10). Clade 1 (Microchaetaceae) groups together T. distorta, C. wrangelii and S. rafaelensis, taxa we would argue belong to the family Microchaetaceae. Clade 2 (Rivulariaceae) groups together sequences from our isolate identified as C. parietina with two published Calothrix sequences. Clade 3 (Scytonemataceae) groups together our sequence from our isolate identified as S. hyalinum with published sequences of S. hofmanni. The similarity matrix of 16S (Table 2) reveals as much difference between the Microchaetaceae clade and the Scytonemataceae clade as there is between the Rivulariaceae clade and the Scytonemataceae clade. (2) The ITS data are consistent with conclusions drawn from the 16S data. Gel electrophoresis of PCR amplification products (long PCR) of whole cell DNA using cyanobacterial specific rRNA primers reveal a single band for members of the Microchaetaceae clade and bands of two different sizes from the other clades. Upon cloning ITS regions reamplified from the short PCR products, we recovered a single type of ITS region containing two tRNA's in the Microchaetaceae clade and two different types of ITS regions, one with two tRNA's and one with no tRNA's, from the other two clades. Furthermore, sequence analysis revealed that the D1 region of the ITS, a highly conserved region known to be important in forming the secondary structure essential for rRNA processing, is identical in the Scytonemataceae and Rivulariaceae clades but shows a single base pair deletion in all genera of the Microchaetaceae clade. These findings underline the usefulness of the ITS region, in conjunction with 16S rRNA sequence data, in resolving taxonomic questions in the cyanobacteria.

ACKNOWLEDGEMENTS

Bruce Webb provided soil chemistry analyses of the study site. Larry St. Clair provided transportation and other logistical support while we were in Utah. Funds for travel were provided by a grant to JRJ (Contract DACA88-95-C-0015) from the U.S. Army Construction Engineering Research Laboratory in Champaign, IL, for whom other related studies are being conducted. The Zeiss Axioskop was purchased with funds from the Instrumentation and Instrument Development Program of the National Science Foundation (Award 9319239). John Carroll University provided funds for laboratory supplies and research release time for both authors. VRF was the recipient of a John Carroll University Summer Research Faculty Fellowship.  This research is also part of a funded study from the National Science Foundation through the special program Biotic Surveys and Inventories.

LITERATURE CITED

Table 1. Comparison of filament, trichome, and cell dimensions of two different clones grown in nitrogen free Z-8 and Bold's Basal media. Dimensions given in microns, length to width ratio is cell length divided by trichome width.

Parameter		Z-8 Medium		Bold's Basal Medium
			Clone 1  Clone 2	Clone 1     Clone 2
______________________________________________________________________________
Filament width (range)	6-10     6-12        	10.0-17.5   13.5-20
Filament width (x±SE)	8.8±0.2  8.1±0.2 	14.5±0.3    16.7±0.3
Trichome width (range)	5.5-10.0 6-11     	8-14        8-13
Trichome width (x±SE)	7.8±0.1  7.1±0.1 	10.9±0.2    10.6±0.2
Cell length (range)	2.0-6.5  2-6.4   	2.0-7.2     2.0-4.5
Cell length (x±SE)	3.8±0.2  3.7±0.2     	4.4±0.1	    3.3±0.1
Length/width (range)	.25-.84  .28-.89     	.18-0.85    .18-0.50
Length/width (x±SE)     .50±.02  .52±.02   	.34±.02	    .32±.01

______________________________________________________________________________

Table 2. 16S sequence similarities among taxa using the V6, V7 and V8 regions for comparison.

Table 3. 16S sequence similarity among clones of PCR product from single isolates.

_______________________________________________________________________________

_______________________________________________________________________________

Calothrix parietina              95.6-99.1%

Coleodesmium wrangelli    99.1-99.5%

Scytonema hyalinum          97.9-99.1%

Spirirestis rafaelensis         99.5-99.9%

Tolypothrix distorta            99.7%

Table 4. Comparison of the D1 ITS sequences among five members of the Nostocales isolated from desert soils.

______________________________________________________________________

Taxon and ITS configuration                  D1 Sequence

_____________________________________________________________________

Spirirestis rafaelensis                         AGG-AGACC

Tolypothrix distorta                           AGG-AGACC

Coleodesmium wrangelii                    AGG-AGACC

Calothrix parietina (2 tRNAs)           AGGGAGACC

Calothrix parietina (no tRNAs)        AGGGAGACC

Scytonema hyalinum (no tRNAs)      AGGGAGACC

Scytonema hyalinum (no tRNAs)      AGGGAGACC

Scytonema hyalinum (2 tRNAs)        AGGGAGACC