Phylogenetic Diversity of Gram-Positive Bacteria Cultured from Marine Sediments

Abstract

Major advances in our understanding of marine bacterial diversity have been gained through studies of bacterioplankton, the vast majority of which appear to be gram negative. Less effort has been devoted to studies of bacteria inhabiting marine sediments, yet there is evidence to suggest that gram-positive bacteria comprise a relatively large proportion of these communities. To further expand our understanding of the aerobic gram-positive bacteria present in tropical marine sediments, a culture-dependent approach was applied to sediments collected in the Republic of Palau from the intertidal zone to depths of 500 m. This investigation resulted in the isolation of 1,624 diverse gram-positive bacteria spanning 22 families, including many that appear to represent new taxa. Phylogenetic analysis of 189 representative isolates, based on 16S rRNA gene sequence data, indicated that 124 (65.6%) belonged to the class Actinobacteria while the remaining 65 (34.4%) were members of the class Bacilli. Using a sequence identity value of ≥98%, the 189 isolates grouped into 78 operational taxonomic units, of which 29 (37.2%) are likely to represent new taxa. The high degree of phylogenetic novelty observed during this study highlights the fact that a great deal remains to be learned about the diversity of gram-positive bacteria in marine sediments.

Gram-positive bacteria can be divided into two major subdivisions: the phylum Actinobacteria, also described as the high-G+C gram-positives, and the phylum Firmicutes, also known as the low-G+C gram-positives, a group that includes such well-known genera as Bacillus and Clostridium (21). Gram-positive bacteria typically have a cell wall consisting of a thick layer of peptidoglycan (19), while a few rather unusual genera lack a cell wall entirely (42). Many in this large group of primarily chemoorganotrophic bacteria are also known to produce spores in response to starvation or harsh chemical or physical conditions (17, 40, 50). Aerobic gram-positive bacteria, specifically actinomycetes (defined here as bacteria within the order Actinomycetales) and members of the order Bacillales, are generally saprophytic and include well-known producers of important secondary metabolites (23, 53).

While the most thoroughly studied gram-positive bacteria include human pathogens (e.g., Mycobacterium tuberculosis, Bacillus anthracis) and soil-derived, antibiotic-producing actinomycetes (2), relatively little is known about the diversity and distribution of gram-positive bacteria in the marine environment. This lack of information persists despite the fact that gram-positive bacteria have been cultured from the ocean for decades (5, 26, 32, 43, 68) and consistently appear in culture-independent studies (e.g., references 62 and 66), including the report of a new and as-yet-uncultured order within the class Actinobacteria (54). Gram-positive bacteria are likely to play important microbiological roles in the marine environment, yet without a fundamental understanding of their diversity and ecophysiology, it is difficult to assess the ecological significance of this relatively overlooked component of the marine bacterial community.

Although gram-positive bacteria have been cultivated from seawater, marine invertebrates, and other sample types (25, 27, 29, 47, 69), marine sediments (32, 34, 45, 48, 64), including deep-sea sediments (39, 56, 68), are the primary oceanic habitat from which they have been recovered (1). While it is probable that some marine-derived gram-positive bacteria are terrigenous microorganisms, washed or blown into the marine environment, species occurring exclusively in the sea have been described (25, 26, 69). The recovery of gram-positive bacteria that require seawater for growth, including several Bacillus species (24, 28, 56, 71) and the recently described actinomycete genus Salinispora (44), suggests that additional, obligate marine taxa reside in marine sediments.

Encouraged by recent work that clearly demonstrated how improved, selective cultivation methods are an effective means of isolating significant new examples of bacterial diversity (36, 55, 57, 70), we performed a series of culture-dependent experiments designed to assess the diversity of gram-positive bacteria in marine sediments. The results revealed a diverse assemblage of bacteria spanning 22 gram-positive families, including many that appear to represent new taxa.

MATERIALS AND METHODS

Sediment collection and bacterial isolation.

A total of 225 sediment samples were collected from the intertidal zone to depths of 500 m during a research expedition to the Republic of Palau (7°30′N, 134°30′E), from 6 to 17 March 2004. Sediment samples were collected either by scuba divers or by using a modified, surface-deployed sediment sampler (model no. 214WA110; Kahlisco, El Cajon, CA). Following collection, samples were placed in sterile 50-ml plastic Whirl-Pak bags (NASCO, Modesto, CA) and kept cool until processed (within 4 h) by one or more of the following four selective methods.

The first processing method involved drying 10 ml of wet sediment overnight in a laminar-flow hood before stamping onto agar media. The method was performed as described previously (34) with the exception that a polyester fiber-tipped sterile swab (Fisher Scientific, Hampton, NH) was used to press the dried sediment onto the agar surface 35 to 40 times, creating a serial-dilution effect. The second processing method involved adding 0.5 g of sediment (dried overnight) to 4 ml of autoclaved seawater passed through a 0.2-μm-pore-size filter (AFSW) either with (final concentration, 5 μg/ml) or without kanamycin. After vigorous shaking for 30 s, the sediment was allowed to settle for 5 min before 50 μl was inoculated onto agar media and spread with an alcohol-sterilized glass rod. For the third processing technique, wet sediment was diluted (1:4) in AFSW and then heated for 6 min at 55°C. The diluted sample was then vigorously shaken for 30 s and further diluted (1:4), and 50 μl of each dilution was plated onto agar media. Finally, pour plates were prepared by adding 0.5 g of wet sediment to 25 ml of autoclaved, molten (∼42°C) 100% seawater agar amended with cycloheximide (100 μg/ml) and rifampin (5 μg/ml).

Processed samples were inoculated onto one or more of 11 different isolation media (A1 to A11). All agar media were prepared with filtered (0.2-μm pore size), deionized (DI) water and/or natural seawater and were amended with filtered (0.2-μm pore size) cycloheximide (100 μg/ml) and a second antibiotic (if noted), after autoclaving. The isolation media consisted of the following: A1, 18 g agar, 10 g starch, 4 g yeast extract, 2 g peptone, 1 liter natural seawater, rifampin (5 μg/ml); A2 (10% A1), 18 g agar, 1 g starch, 0.4 g yeast extract, 0.2 g peptone, 1 liter natural seawater; A3, 18 g agar, 2.5 g starch, 1 g yeast extract, 0.5 g peptone, 0.2 g glycerophosphate (disodium pentahydrate), 750 ml natural seawater, 250 ml DI water; A4 (100% seawater agar), 18 g agar, 1 liter natural seawater; A5 (75% seawater agar), 18 g agar, 750 ml natural seawater, 250 ml DI water; A6 to A9, 18 g agar, 1 liter natural seawater, one antibiotic (5 μg/ml polymixin B sulfate, 5 μg/ml kanamycin, 25 μg/ml novobiocin, or 5 μg/ml rifampin, respectively); A10, 8 g noble agar, 0.5 g mannitol, 0.1 g peptone, 1 liter natural seawater, 5 μg/ml rifampin; A11 (Munz medium [49]), 18 g agar, 1 g KNO₃, 0.1 g MgSO₄-7H₂O, 2 g Na₂HPO₄-7H₂O, 0.14 g KH₂PO₄, 1 g NaCl, 1 liter DI water, 5 ml light liquid paraffin (added after autoclaving).

Inoculated plates were incubated at 25 to 28°C for up to 12 weeks, and all well-separated bacterial colonies, observed by eye or using a stereomicroscope at a magnification of up to ×64 (Leica Microscopy Systems Ltd., Heerbrugg, Switzerland), were removed from the original isolation plates and subcultured on A1. The Gram reaction of all pure cultures was determined via the nonstaining (KOH) method (6). The majority of the gram-positive strains possessed morphological features characteristic of the recently described actinomycete genus Salinispora (44). Multiple strains from each Salinispora-like morphotype were cryopreserved at −80°C along with all of the remaining gram-positive strains. All strains were grouped according to colony color, morphology, and pigment production, and representatives from each phenotype were subjected to phylogenetic analysis.

Nucleic acid extraction, 16S rRNA gene amplification, and sequencing.

Genomic DNA was extracted according to the DNeasy protocol (QIAGEN Inc., Valencia, CA) with the following modifications. After RNase A (2 mg/ml) was added to the enzymatic lysis buffer, the resuspended bacterial pellet was incubated for 2 h at 37°C. Following the addition of proteinase K, the sample was held for 1 h at 70°C. Genomic DNA was eluted from the spin column with 100 μl of elution buffer for immediate use or storage at −20°C.

The 16S rRNA genes were amplified from genomic DNA by PCR using the primers FC27 (5′-AGAGTTTGATCCTGGCTCAG-3′) and RC1492 (5′-TACGGCTACCTTGTTACGACTT-3′). The 50-μl PCR mixture contained 20 to 50 ng of DNA, 250 pmol of each primer, ThermoPol Buffer (New England BioLabs Inc., Beverly, MA), 2.5 U of Taq DNA polymerase (New England BioLabs Inc., Beverly, MA), and 100 μM deoxynucleoside triphosphate mixture. The PCR program consisted of 30 cycles of 94°C for 1 min, 60°C for 1 min, and 72°C for 1 min followed by a final extension step at 72°C for 7 min. Amplification products were examined by agarose gel electrophoresis and purified using the QIAquick PCR cleanup kit (QIAGEN Inc., Valencia, CA) according to the manufacturer's suggested protocol. A partial consensus sequence (Escherichia coli nucleotide numbering 20 to 531) for each isolate was obtained using the primers FC27 and R530 (5′-CCGCGGCTGCTGGCACGTA-3′). Nearly complete sequences were obtained for select 16S rRNA gene amplicons (E. coli nucleotide numbering 20 to 1392) using four additional primers: RC1492, R936 (5′-GTGCGGGCCCCCGTCAATT-3′), F514 (5′-GTGCCAGCAGCCGCGGTAA-3′), and F1114 (5′-GCAACGAGCGCAACCC-3′). Sequencing reactions were carried out with an ABI 3100 DNA sequencer at the DNA Sequencing Shared Resource, UCSD Cancer Center (funded in part by NCI Cancer Center support grant 2 P30CA23100-18).

Phylogenetic analyses and diversity estimates.

All nucleotide sequences were assembled, analyzed, and manually edited using the Sequencher software package (version 4.5; Gene Codes Co., Ann Arbor, MI) and compared to sequences within the NCBI database (http://www.ncbi.nlm.nih.gov/) using the Basic Local Alignment Search Tool (BLAST). All partial 16S rRNA gene sequences sharing a phylogenetic affiliation with either Actinobacteria or Firmicutes were imported into ARB (41) and aligned. Aligned partial 16S rRNA gene sequences (E. coli numbering 20 to 531) were analyzed using the Clusterer program (http://www.bugaco.com/bioinf), and the number of operational taxonomic units (OTUs) was calculated using sequence identity values ranging from ≥90% to 100%. For at least one representative of each OTU generated using the ≥98% sequence identity value, a nearly complete 16S rRNA gene sequence was obtained. Phylogenetic analyses were performed using PAUP (63), and trees were drawn using distance neighbor-joining methods, the unweighted-pair group method using average linkages (UPGMA), and maximum parsimony.

In order to estimate the taxonomic novelty of the bacteria cultured, strains within OTUs sharing a sequence identity value of ≥98% were subjected to further analysis. An OTU was considered a new phylotype if all strains within the OTU shared <98% sequence identity with any previously cultured bacterium for which sequence data were available (as determined by a BLAST search); otherwise, the OTU was designated a known (previously cultured) phylotype. In addition to determining whether the members of each OTU had been previously cultured, an OTU's taxonomic novelty was assessed using the OTU's nearest type strain (http://www.bacterio.cict.fr/). If all isolates within an OTU shared <98% sequence identity with the nearest type strain, as calculated using the ARB distance matrix, the OTU was considered to have a high probability of representing a new taxon. OTUs calculated using a sequence identity value of ≥ 98% were further used to estimate gram-positive bacterial diversity using the abundance-based coverage estimator (9) and Chao's richness estimator (8) implemented in EstimateS (version 7; R. K. Colwell; available at http://viceroy.eeb.uconn.edu/estimates).

Effects of seawater on growth.

Select isolates were screened to determine whether they required seawater for growth. Using a sterile loop, cells from a single colony were streaked onto A1 prepared with natural seawater and A1 prepared with DI water. Plates were incubated at 25 to 28°C for 6 to 8 weeks, and growth was monitored at a magnification of up to ×64. Strains that grew on the medium prepared with seawater but not on the medium prepared with DI water were scored as requiring seawater for growth.

Nucleotide sequence accession numbers.

16S rRNA gene sequences have been deposited in the GenBank database (http://www.ncbi.nlm.nih.gov/GenBank/index.html) under the accession numbers DQ092624, DQ224159, and DQ448693 to DQ448806.

RESULTS

From a total of 225 sediment samples, 1,624 gram-positive bacteria were isolated. Interestingly, 1,353 (83.3%) of these strains possessed morphological features characteristic of the genus Salinispora (44). Four hundred seven of the Salinispora-like strains, along with the remaining 271 gram-positive strains, were cryopreserved at −80°C. Of these 678 strains, 199 were chosen for phylogenetic analysis based on colony color and morphology. These 199 isolates included 25 Salinispora-like strains and 64.2% (174) of the remaining gram-positive strains cultured. NCBI nucleotide BLAST searches using the partial 16S rRNA gene sequences of these 199 strains revealed that 189 (95.0%) of the isolates were gram positive and shared a phylogenetic affiliation with members of the Actinomycetales or Bacillales. These results further validate the KOH method (6) as a rapid and effective means to determine the cell wall type of an isolate. (For additional information on the 189 gram-positive isolates, including collection depth, isolation method and medium, seawater requirement, and nearest type strain, see the supplemental material.) Even though sediment-processing methods were not applied equally to all samples, stamping dried sediments onto low-nutrient agar proved to be a highly successful method to cultivate gram-positive bacteria. In fact, over 70% of the gram-positive strains were cultured on low-nutrient media, particularly A4 to A6.

A phylogenetic analysis of 25 of the 1,353 strains that shared morphological similarities with the genus Salinispora (44) revealed that 23 shared >99% 16S rRNA gene sequence identity with members of this taxon. The other two strains belonged to the closely related genus Micromonospora. Of the 23 Salinispora strains, 16 (69.9%) shared 100% sequence identity with Salinispora arenicola, further supporting the pantropic distribution and lack of intraspecies 16S rRNA gene diversity within this taxon (35). None of the 23 strains clustered with Salinispora tropica, which to date has only been reported from the Bahamas. The remaining seven (30.4%) strains belonged to a new phylotype for which the name “Salinispora pacifica” has been proposed (35).

The diversity of gram-positive bacteria cultured in this study was estimated by performing cluster analyses using the 189 partial 16S rRNA gene sequences. The numbers of OTUs calculated using various levels of sequence identity were as follows: ≥90%, 8 OTUs; ≥91%, 9 OTUs; ≥92%, 15 OTUs; ≥93%, 18 OTUs; ≥94%, 35 OTUs; ≥95%, 43 OTUs; ≥96%, 49 OTUs; ≥97%, 63 OTUs; ≥98%, 78 OTUs; ≥99%, 95 OTUs; 100%, 116 OTUs. Of the 116 distinct gram-positive sequences identified, 70 (60.3%) were phylogenetically affiliated with the order Actinomycetales (Fig. 1). These actinomycetes are most closely related to 25 different genera that fall within 18 separate family level groupings and span 8 of the 10 suborders within the order Actinomycetales.

FIG. 1.

Neighbor-joining distance tree constructed in PAUP (63) using the aligned, partial 16S rRNA gene sequences (512 nucleotide positions) of strains representing each of the 70 Actinomycetales OTUs (generated using a sequence identity value of 100%) and the type strains of the most closely related genera. Sequences from this study are shown in boldface, and GenBank accession numbers are given in parentheses. Bootstrap values (in percent) calculated from 1,000 resamplings using the neighbor-joining method are shown at the nodes for values of ≥60%. Sphaerobacter thermophilus was used to position the root. For multiple strains that shared an identical partial 16S rRNA gene sequence, the number of additional isolates is presented in brackets. The suborders to which the strains belong are presented on the right. Family- and genus-level affiliations were maintained when distance UPGMA and maximum-parsimony treeing methods were applied, although some within family branching patterns changed.

The remaining 46 (39.7%) OTUs calculated using 100% sequence identity shared a phylogenetic affiliation with the order Bacillales (Fig. 2). The majority of these OTUs (67.4%) formed a highly diverse clade, all of whose members are most closely related to the genus Bacillus. The remaining 15 OTUs were most closely related to the genera Exiguobacterium, Halobacillus, Laceyella, Paenibacillus, Pontibacillus, and Staphylococcus. Contrary to formal taxonomic assignment, the single Staphylococcus strain (CNJ-924) and the Exiguobacterium strains (CNJ-771 and CNJ-781) did not appear to form a clade with their respective families when partial 16S rRNA gene sequences were used (Fig. 2). However, when using nearly complete 16S rRNA gene sequences, these relationships were rectified. CNJ-924 grouped with its appropriate family, the Staphylococcaceae (data not shown), and the Exiguobacterium strains, while deeply rooted, grouped with the Bacillaceae (Fig. 3).

FIG. 2.

Neighbor-joining distance tree constructed in PAUP (63) using the aligned, partial 16S rRNA gene sequences (512 nucleotide positions) of strains representing each of the 46 Bacillales OTUs (generated using a sequence identity value of 100%) and the type strains of the most closely related genera (with the exception of Exiguobacterium aurantiacum and Halobacillus halophilus, for which alternative sequences were used). Sequences from this study are shown in boldface, and GenBank accession numbers are given in parentheses. Bootstrap values (in percent) calculated from 1,000 resamplings using the neighbor-joining method are shown at the nodes for values of ≥60%. Coprothermobacter proteolyticus was used to position the root. For multiple strains that shared an identical, partial 16S rRNA gene sequence, the number of additional isolates is presented in brackets. The families to which the strains belong are presented on the right. Family- and genus-level affiliations were maintained when distance UPGMA and maximum-parsimony treeing methods were applied, although some within family branching patterns changed.

FIG. 3.

Neighbor-joining distance tree based on the nearly complete and aligned 16S rRNA gene sequences of 41 Actinomycetales and Bacillales OTUs (calculated using a sequence identity value of ≥98%) observed in this study and their nearest type strains. The strains used to construct this tree represent the 29 OTUs that have not yet been formally described and the 12 OTUs whose nearest type strain was isolated from a marine source. A total of 1,367 nucleotide positions were included in the analysis, and Deinococcus radiophilus was used to position the root. GenBank accession numbers are given in parentheses following the strain identification (in boldface). Bootstrap values (in percent) calculated from 1,000 resamplings using the neighbor-joining method are shown at the nodes for values of ≥60%. The number of additional isolates within each OTU is presented in brackets. Double asterisks indicate OTUs in which all of the tested isolates required seawater for growth. The topology of the distance neighbor-joining tree is supported by distance UPGMA and maximum-parsimony treeing methods.

While 100% 16S rRNA gene sequence identity was used to highlight the 16S rRNA diversity of the cultured isolates, the 52 Actinomycetales OTUs (Table 1) and 26 Bacillales OTUs (Table 2) generated using the more conservative identity value of ≥98% were used to estimate the phylogenetic novelty of the culture collection. When all strains within these OTUs shared <98% sequence identity with all previously cultured bacteria for which comparable sequence data were available, the OTU was considered a new phylotype. According to this criterion, 12 of the 52 Actinomycetales OTUs (23.1%) and 9 of the 26 Bacillales OTUs (34.6%) represent new phylotypes (Table 3). Thus, 21 of the 78 observed phylotypes (26.9%) have not been previously cultured and reported using 16S sequence-based methods. Of the 40 known Actinomycetales OTUs and 17 known Bacillales OTUs, 35 and 10, respectively, had not been previously reported from marine sources. Given that the samples were collected close to shore, (within 10 km), these 45 strains represent bacteria that appear to be adapted to both marine and nonmarine environments. Using the abundance-based coverage estimator and Chao's richness estimator, it can be predicted that the numbers of cultivable OTUs (≥98% sequence identity) in these sediments are 131 and 130, respectively. Relative to the 78 OTUs detected, these estimators suggest that further processing of the sediment samples would result in the cultivation of additional gram-positive bacterial diversity.

TABLE 1.

List of isolates representing the 52 Actinomycetales OTUs generated using a 16S rRNA percent identity value of ≥98%^a

Phylogenetic group (family)	Representative isolate (accession no.)	Sequence length (bp)	No. of strains in OTU	Nearest type strain (accession no.)	Sequence identity (%)b	Source of nearest type strain
Brevibacteriaceae	CNJ-737 (DQ448693)	1,480	1	Brevibacterium epidermidis (X76565)	99.8	Skin
Corynebacterium	CNJ-954 (DQ448694)	1,480	1	Corynebacterium variabile (AJ222815)	98.6	Animal fodder
Dermacoccaceae	CNJ-855 (DQ448695)	1,476	3	Kytococcus sedentarius (X87755)	99.6	Skin
Dietziaceae	CNJ-898 (DQ448696)	1,469	8	Dietzia maris (X79290)	99.9	Marine sediment
Geodermatophilaceae	CNJ-868 (DQ448697)	1,471	1	Blastococcus saxobsidens (AJ316570)	98.8	Stone surfaces
	CNJ-793 (DQ448698)	1,472	2	Modestobacter multiseptatus (Y18646)	98.1	Soil
Gordoniaceae	CNJ-756 (DQ448699)	1,473	5	Gordonia bronchialis (X79287)	98.4	Human sputum
	CNJ-863 (DQ448700)	1,475	1	Gordonia nitida (AF148947)	100	Industrial wastewater
	CNJ-754 (DQ448701)	1,473	3	Gordonia polyisoprenivorans (Y18310)	98.4	Deteriorated automobile tire
	CNJ-752 (DQ448702)	1,475	4	Gordonia terrae (X81922)	100	Soil
Intrasporangiaceae	CNJ-824 (DQ448703)	1,477	3	Ornithinimicrobium humiphilum (AJ277650)	97.2	Soil
	CNJ-927 (DQ448704)	1,495	1	Serinicoccus marinus (AY382898)	100	Seawater
Microbacteriaceae	CNJ-745 (DQ448705)	1,479	2	Agromyces aurantiacus (AF389342)	98.2	Soil
	CNJ-930 (DQ448706)	1,477	2	Microbacterium flavescens (Y17232)	97.3	Soil
	CNJ-743 (DQ448707)	1,475	1	Microbacterium imperiale (X77442)	97.9	Moth's alimentary tract
	CNJ-797 (DQ448708)	1,474	1	Microbacterium schleiferi (Y17237)	99	Activated sludge
Micrococcaceae	CNJ-723 (DQ448709)	1,377	2	Kocuria marina (AY211385)	96.6	Marine sediment
	CNJ-900 (DQ448710)	1,479	1	Kocuria palustris (Y16263)	100	Cattail rhizosphere
	CNJ-770 (DQ448711)	1,479	2	Kocuria rosea (X87756)	98.2	Soil
	CNJ-719 (DQ448712)	1,475	13	Micrococcus luteus (AF542073)	99.3	Wall painting
Micromonosporaceae	CNS-326 (DQ448713)	1,468	1	Micromonospora endolithica (AJ560635)	97.8	Antarctic sandstone
	CNJ-878 (DQ448714)	1,469	1	Micromonospora endolithica (AJ560635)	98.6	Antarctic sandstone
	CNS-051 (DQ448715)	1,468	16	Salinispora arenicola (AY040619)	100	Marine sediment
	CNS-143 (DQ092624)	1,468	7	Salinispora tropica (AY040617)	99.6	Marine sediment
Mycobacteriaceae	CNJ-859 (DQ448716)	1,472	3	Mycobacterium brisbanense (AY012577)	98.5	Human
	CNJ-823 (DQ448717)	1,472	7	Mycobacterium porifera (AF480589)	99.9	Finland sponge
Nocardiaceae	CNS-044 (DQ448718)	1,471	2	Nocardia arthritidis (AB108781)	99.3	Human
Nocardioidaceae	CNJ-889 (DQ448719)	1,475	1	Aeromicrobium erythreum (AF005021)	94.8	Soil
	CNJ-780 (DQ448720)	1,471	1	Marmoricola aurantiacus (Y18629)	94.8	Marble statue
	CNJ-872 (DQ448721)	1,472	1	Marmoricola aurantiacus (Y18629)	97	Marble statue
	CNJ-892 (DQ448722)	1,468	1	Nocardioides ganghwensis (AY423718)	97.9	Tidal flat sediment
Nocardiopsaceae	CNR-923 (DQ448723)	1,485	1	Nocardiopsis lucentensis (X97888)	99	Salt marsh soil
Promicromonosporaceae	CNJ-734 (DQ448724)	1,474	1	Promicromonospora sukumoe (AJ272024)	97.3	Soil
Pseudonocardiaceae	CNJ-888 (DQ448725)	1,483	1	Pseudonocardia antarctica (AJ576010)	94.4	Soil
	CNS-139 (DQ448726)	1,472	1	Pseudonocardia yunnanensis (AJ252822)	97.7	Soil
	CNS-004 (DQ448727)	1,474	1	Pseudonocardia zijingensis (AF325725)	98.7	Soil
Streptomycetaceae	CNR-884 (DQ448728)	1,478	1	Streptomyces arenae (AJ399485)	99.5	Soil
	CNR-926 (DQ448729)	1,470	2	Streptomyces aureofaciens (AY289116)	97.7	Soil
	CNR-881 (DQ448730)	1,476	1	Streptomyces bikiniensis (X79851)	98.7	Soil
	CNR-918 (DQ448731)	1,480	1	Streptomyces caviscabies (AF112160)	99.5	Potato lesion
	CNR-924 (DQ448732)	1,478	1	Streptomyces chartreusis (AJ399468)	99.4	Soil
	CNR-875 (DQ448733)	1,478	1	Streptomyces galilaeus (AB045878)	98.7	Soil
	CNR-872 (DQ448734)	1,497	1	Streptomyces hebeiensis (AY277529)	95.5	Soil
	CNR-880 (DQ448735)	1,475	1	Streptomyces koyangensis (AY079156)	99.2	Soil
	CNS-177 (DQ448736)	1,481	1	Streptomyces lydicus (Y15507)	99.2	Potato scab
	CNJ-962 (DQ448737)	1,495	1	Streptomyces sampsonii (D63871)	95.5	Potato scab
	CNR-887 (DQ448738)	1,477	1	Streptomyces sampsonii (D63871)	99	Potato scab
	CNR-885 (DQ448739)	1,481	3	Streptomyces tendae (D63873)	98.9	Potato scab
	CNR-877 (DQ448740)	1,423	1	Streptomyces thermocarboxydovoran (U94489)	95.6	Soil
	CNR-940 (DQ448741)	1,482	2	Streptomyces thermocoprophilus (AJ007402)	96.8	Poultry feces
	CNR-925 (DQ448742)	1,480	2	Streptomyces thermocoprophilus (AJ007402)	97.3	Poultry feces
Thermomonosporaceae	CNU-125 (DQ448743)	1,472	1	Actinomadura cremea (AF134067)	99.1	Soil

^aFor each OTU, the nearest type strain and its source are identified. When an isolate and its nearest type strain share <98% sequence identity, the percent sequence identity is shown in boldface.

^bSequence identity shared between the representative isolate and its nearest type strain.

TABLE 2.

List of isolates representing the 26 Bacillales OTUs generated using a 16S rRNA percent identity value of ≥98%^a

Phylogenetic group (family)	Representative isolate (accession no.)	Sequence length (bp)	No. of strains in OTU	Nearest type strain (accession no.)	Sequence identity (%)b	Source of nearest type strain
Bacillaceae	CNJ-803 (DQ448744)	1,502	7	Bacillus algicola (AY228462)	99.8	Fucus evanescens thallus
	CNJ-796 (DQ448745)	1,505	3	Bacillus aquimaris (AF483625)	96.1	Seawater
	CNJ-733 (DQ448746)	1,504	3	Bacillus aquimaris (AF483625)	99.5	Seawater
	CNJ-815 (DQ448747)	1,502	3	Bacillus barbaricus (AJ422145)	99.4	Exptl wall painting
	CNJ-826 (DQ448748)	1,501	1	Bacillus bataviensis (AJ542508)	96.9	Soil
	CNJ-732 (DQ448749)	1,504	2	Bacillus cereus (AE017013)	100	Air
	CNJ-816 (DQ448750)	1,504	1	Bacillus cohnii (X76437)	98.2	Horse meadow soil
	CNJ-828 (DQ448751)	1,503	5	Bacillus decoloationis (AJ315075)	97.9	Mural painting
	CNJ-958 (DQ448752)	1,444	1	Bacillus endophyticus (AF295302)	95	Cotton plant inner tissue
	CNJ-905 (DQ448753)	1,502	1	Bacillus firmus (AJ717384)	97.8	Nonsaline alkaline groundwater
	CNJ-933 (DQ448754)	1,503	2	Bacillus firmus (AJ717384)	99.7	Nonsaline alkaline groundwater
	CNJ-759 (DQ448755)	1,504	1	Bacillus horikoshii (X76443)	99.5	Soil
	CNJ-775 (DQ448756)	1,462	3	Bacillus humi (AJ627210)	96.4	Soil
	CNJ-782 (DQ448757)	1,490	2	Bacillus indicus (AJ583158)	99.9	Arsenic-polluted sand
	CNJ-778 (DQ448758)	1,504	6	Bacillus megaterium (X60629)	99.9	Soil
	CNJ-748 (DQ448759)	1,503	1	Bacillus methanolicus (AB112727)	96.3	Sugar beet wastewater facility
	CNJ-742 (DQ448760)	1,500	3	Bacillus pumilus (AY876289)	99.4	Soil
	CNJ-771 (DQ448761)	1,514	4	Exiguobacterium aestuarii (AY594264)	99.7	Seawater, Korea
	CNJ-915 (DQ448762)	1,517	3	Halobacillus litoralis (X94558)	99.2	Salt marsh soil
	CNJ-895 (DQ448763)	1,518	2	Halobacillus salinus (AF500003)	99.9	East Sea coast salt lake
	CNJ-812 (DQ448764)	1,475	3	Halobacillus trueperi (AJ310149)	98	Saline sediment, salt lake
	CNJ-912 (DQ448765)	1,515	1	Pontibacillus chungwhensis (AY553296)	97.1	Korean solar saltern
Paenibacillaceae	CNJ-934 (DQ448766)	1,508	1	Paenibacillus turicensis (AF378697)	91.9	Cerebrospinal fluid shunt
Staphylococcaceae	CNJ-924 (DQ448767)	1,503	1	Staphylococcus capitis (AY688040)	99.8	Human skin
Thermoactinomycetaceae	CNR-949 (DQ448768)	1,497	2	Laceyella sacchari (AF138737)	91.9	Soil
	CNJ-795 (DQ448769)	1,469	3	Laceyella sacchari (AF138737)	92.3	Soil

^aFor each OTU, the nearest type strain and its source are identified. When an isolate and its nearest type strain share <98% sequence identity, the percent sequence identity is shown in boldface.

^bSequence identity shared between the representative isolate and its nearest type strain.

TABLE 3.

Number of OTUs, generated using a 16S rRNA gene sequence identity of ≥98%, and strains belonging to new and known phylotypes^a

Order	New		Known (marine)		Total
	OTUs	Strains	OTUs	Strains	OTUs	Strains
Actinomycetales	12	14	40 (5)	110 (39)	52	124
Bacillales	9	16	17 (7)	49 (22)	26	65

^aAll new phylotypes share <98% 16S rRNA gene sequence identity with cultured and sequenced bacteria. Known phylotypes that are most closely related to a type strain isolated from a marine source are shown in parentheses.

The 21 OTUs identified as new (not previously cultured) phylotypes have a high probability of representing new taxa. Additionally, six previously observed Actinomycetales and two previously observed Bacillales OTUs may also represent new taxa, as all strains within these OTUs shared <98% 16S rRNA gene sequence identity with their nearest type strains. Thus, in total, 29 of the 78 OTUs (37.2%) cultured as part of this study have the potential to be described as new taxa. The nearly complete 16S rRNA gene sequences of representative isolates from each of these OTUs were used to construct a phylogenetic tree (Fig. 3). Also included in this tree are representative isolates from the 12 OTUs that were not considered new but were most closely related to a type strain isolated from a marine source.

Of the 144 strains tested, 57 required seawater for growth, while the remainder grew either poorly (24 strains) or equally well (63 strains) when seawater was replaced with DI water in the growth medium. Forty-five of the 57 seawater-requiring strains were divided among 14 OTUs (≥98% sequence identity) that were comprised solely of seawater-requiring strains. These strains either belonged to a new OTU or an OTU most closely related to a type strain isolated from a marine source (Fig. 3). Ten additional seawater-requiring strains fell into seven previously observed OTUs that contained from one to five strains that did not require seawater for growth. The final two seawater-requiring strains, each the sole member of a separate OTU, belonged to known OTUs that had not been previously described as requiring seawater.

Thirty-three of the 57 seawater-requiring strains belonged to the order Actinomycetales. In addition to strains related to the known seawater-requiring genus Salinispora, seawater-requiring actinomycetes were also most closely related to the genera Dietzia, Kocuria, Kytococcus, Marmoricola, Microbacterium, Mycobacterium, and Pseudonocardia. Outside of the genus Salinispora, these strains are among the first seawater-requiring actinomycetes to be reported. Also requiring seawater were 24 strains within the class Bacilli. While the majority of these seawater-requiring strains were most closely related to Bacillus species, seawater-requiring strains related to Halobacillus, Laceyella, and Paenibacillus species were also cultivated.

DISCUSSION

Marine bacterioplankton represent one of the most thoroughly studied environmental communities on the planet (22), yet bacteria inhabiting marine sediments remain largely uncharacterized. This lack of information hinders an effective assessment of marine bacterial diversity and limits our understanding of the fundamental differences between the bacterial populations inhabiting two major ocean ecosystems. One apparent yet relatively unexplored difference between seawater and sediment bacterial communities is the relative abundance of gram-positive bacteria. While early research estimated that only 5% of the bacteria in the ocean are gram positive (72), more recent studies suggest that the abundance and diversity of gram-positive strains in sediments may be considerably greater (33, 52, 59). The present study employed cultivation-dependent methods to assess the diversity of gram-positive bacteria in marine sediments collected around the islands of Palau. In total, 78 gram-positive OTUs were cultured, of which 21 are considered to be new phylotypes based on the sharing of <98% 16S rRNA gene sequence identity with any previously cultured isolate for which sequence data are available. Eight other OTUs, previously observed but not yet formally described, bring the total number of potentially new taxa cultured as a part of this study to 29. These results indicate that considerably diverse gram-positive microbial populations can be cultured from marine sediments and reinforces the concept that relatively simple cultivation techniques can be used successfully to isolate many as-yet-undescribed taxa (13, 30, 43).

The frequent use of high-nutrient media in previous studies of bacterial diversity may explain why some gram-positive bacteria have gone uncultured. During the present study, the majority of isolates were obtained using low-nutrient media (e.g., seawater agar [for additional information, see the supplemental material]). In fact, 24 of the 29 OTUs for which formal taxonomic descriptions are not yet available were isolated exclusively from low-nutrient media. While all of the cultured strains were ultimately capable of growth on a high-nutrient medium (i.e., A1), our observations support the results from previous studies (13, 14, 60), which suggest that lower nutrient concentrations improve the initial isolation and recovery of diverse microorganisms as they help avoid contamination and overgrowth by fast-growing strains.

The identification of 21 new gram-positive phylotypes, despite extensive culture-independent investigations of seawater, might suggest that seawater and sediment communities are significantly different. The fact that the number of new phylotypes falls by only three to 18 when the results of culture-independent analyses are included in the comparison (data not shown) supports this possibility. Alternatively, biases associated with culture-independent methods (18, 61, 67) may have contributed to the underestimation of specific groups of gram-positive bacteria that occur in both seawater and sediments. This may be particularly applicable in the case of spore-forming gram-positive bacteria, as it is known that even when specific steps are taken to lyse spores, these bacteria are underrepresented in environmental clone libraries when spore counts are ≤10³/ml of sediment (15, 46, 48). Although culture-dependent approaches also have well-known biases (38, 51, 65), these methods may prove to be the most effective way to detect certain groups of marine bacteria. In addition, cultured strains can be subjected to taxonomic characterization, and their physiology, ecology, and biotechnological potential can be explored.

While the number of OTUs was reported using multiple 16S rRNA gene sequence identity values, only those clusters generated using values of ≥98% were subjected to additional diversity analyses. This value was chosen based on the relationship between percent DNA-DNA reassociation and 16S rRNA gene similarity, where 70% DNA relatedness is expected to correspond to >98% 16S rRNA gene sequence identity (16). Although Stach et al. (58) suggested that a 16S rRNA gene sequence identity value of ≥99% could be used to define an OTU, that study was focused solely on delineating actinobacterial OTUs. The use of a sequence identity value of ≥98% may not provide the most conservative estimate of OTU numbers; however, even at this value it is probable that diversity will be underestimated.

Members of the actinomycete families Micromonosporaceae, Nocardiaceae, and Streptomycetaceae have dominated previous studies of terrestrial and marine-derived Actinobacteria (11, 12, 26, 43), and isolates most closely related to members of each of these three families were cultured during the present study. Based on morphological characterization, the majority of the isolates recovered were identified as Micromonosporaceae, supporting previous observations that these bacteria are among the dominant actinomycetes cultivable from marine sediments (31, 68). Also readily cultured from marine sediments were actinomycetes of the families Nocardiaceae and Streptomycetaceae. While we were surprised not to recover Rhodococcus isolates, which are among the most common members of the Nocardiaceae recovered from marine samples (11, 12, 26), our processing methods clearly did not select against other mycolate actinomycetes, including strains most closely related to Corynebacterium, Dietzia, Gordonia, Mycobacterium, and Nocardia. Within the Streptomycetaceae, a diverse assemblage of filamentous, spore-forming actinomycetes grouped into 15 OTUs. Five of those Streptomycetaceae OTUs shared <98% 16S rRNA gene sequence identity with the most closely related type strain, and thus considerable new examples of taxonomic diversity appear to have been cultured within this well-studied family.

The phylogenetic identification of what appear to be new taxa within the Actinomycetales and Bacillales confirmed previous observations that marine sediments harbor new diversity within these groups (11, 26, 43, 44). These two orders are responsible for almost 50% of the known bioactive microbial metabolites discovered to date, including many well-known antibiotics (2). Although marine microorganisms have only recently become a target for natural product drug discovery, it has become increasingly clear that gram-positive strains are a rich source of new structures that possess promising antimicrobial and anticancer activities (3, 4, 37) and that a better understanding of microbial diversity will provide important insight into how to devise intelligent strategies for natural product discovery (7). The present study helps to establish a fundamental understanding of the diversity of gram-positive bacteria in the marine environment and provides a diverse, marine environment-derived assemblage of cultured gram-positive bacteria whose chemical and biosynthetic diversity can be investigated.

In addition to actinomycetes from the families Micromonosporaceae, Nocardiaceae, and Streptomycetaceae, spore-forming strains from the Pseudonocardiaceae and Thermomonosporaceae and a large and diverse assemblage of unicellular and/or non-spore-forming gram-positive bacteria were cultured. While a diverse assemblage of bacteria within the Actinomycetales was cultured, no strains from other orders within the Actinobacteria were isolated despite the fact that bacteria from other orders have been identified in the marine environment using culture-independent methods (54).

Within the actinomycetes, the highest level of sequence divergence was observed within the Nocardioidaceae (Table 1), with all strains sharing <98% sequence identity to currently described species. CNJ-780 and CNJ-872 were most closely related to Marmoricola aurantiacus, the only described species within the genus Marmoricola. Their percent identities with the type strain (94.8% and 97.0%, respectively) suggest that they may represent new species and, perhaps in the case of CNJ-872, a new genus within the Nocardioidaceae. Significant phylogenetic novelty was also observed among strains most closely related to the genera Bacillus, Pontibacillus, Paenibacillus, and Laceyella. These strains appear to represent multiple new species and, in the case of the Paenibacillus and Laceyella strains, which share only 91.9% and 92.3% sequence identity with their respective nearest type strains, possibly higher-level taxa.

Of the potential new taxa observed, 7 of the 11 Bacillales OTUs and 3 of the 18 Actinomycetales OTUs required seawater for growth (Fig. 3). While it is possible that strains belonging to these OTUs also occur in nonmarine environments, it is equally plausible that the seawater-requiring OTUs represent obligate marine taxa. Both the number and phylogenetic distribution of these seawater-requiring actinomycete and Bacillales strains was intriguing as they were clearly scattered throughout the phylogenetic tree (Fig. 3). Thus, it remains possible that the requirement of seawater for growth either evolved rapidly and independently in these groups, was acquired by horizontal gene transfer, or represents a highly plastic phenotype.

The most remarkable intraclade diversity observed in the present study occurred within the genus Bacillus. This genus has been generally recognized to be among the most heterogeneous within the bacterial domain and in need of division into multiple genera (10). The present study recovered 45 strains most closely related to 17 described Bacillus species. These strains shared, in some cases, <88% 16S rRNA gene sequence identity, far outside the sequence diversity associated with most bacterial genera. While a taxonomic reevaluation of the genus Bacillus in the near future is improbable, the results clearly indicate that considerably diverse Bacillus populations can be readily cultured from marine sediments.

Another noteworthy observation from this study was the recovery of 11 strains from six separate OTUs that share 100% 16S rRNA sequence identity with a type strain. While it was not surprising to culture Salinispora arenicola and Serinicoccus marinus, species previously reported to have been isolated from marine sediments and seawater, respectively, the recovery of a strain with 100% sequence identity to Kocuria palustris, isolated originally from a cattail rhizosphere sampled at the Soroksar tributary of the Danube river, Hungary (Table 1), suggests that some bacterial strains exhibit remarkably broad geographical and environmental distributions.

There is presently much to learn about gram-positive bacteria in marine sediments. Like their terrestrial relatives, marine gram-positive bacteria may play a significant role in the breakdown of recalcitrant organic matter and therefore in the ocean's biogeochemical cycle. Additionally, even as spores, marine gram-positive bacteria have the capacity to impact their surrounding chemical environment, as evidenced by their capacity to oxidize metals (20). It is clear from this single survey that considerable new examples of gram-positive bacterial diversity can be readily cultured from marine sediments. The continued use of cultivation-dependent techniques will undoubtedly lead to the discovery of additional gram-positive diversity and provide a direct means to learn more about their ecophysiology and applications in biotechnology.