Oecologia

, Volume 162, Issue 2, pp 435–442

Ascomycete fungal communities associated with early decaying leaves of Spartina spp. from central California estuaries

Authors

  • Justine I. Lyons
    • Department of Marine SciencesUniversity of Georgia
  • Merryl Alber
    • Department of Marine SciencesUniversity of Georgia
    • Department of Marine SciencesUniversity of Georgia
Community Ecology - Original Paper

DOI: 10.1007/s00442-009-1460-4

Cite this article as:
Lyons, J.I., Alber, M. & Hollibaugh, J.T. Oecologia (2010) 162: 435. doi:10.1007/s00442-009-1460-4

Abstract

Ascomycetous fungi play an important role in the early stages of decomposition of Spartina alterniflora, but their role in the decomposition of other Spartina species has not been investigated. Here we use fingerprint (terminal restriction fragment length polymorphism) and phylogenetic analyses of the 18S to 28S internal transcribed spacer region to compare the composition of the ascomycete fungal communities on early decay blades of Spartina species (Spartina alterniflora, Spartina densiflora, Spartina foliosa, and a hybrid (S. alterniflora × S. foliosa)) collected from three salt marshes in San Francisco Bay and one in Tomales Bay, California, USA. Phaeosphaeria spartinicola was found on all samples collected and was often dominant. Two other ascomycetes, Phaeosphaeria halima and Mycosphaerella sp. strain 2, were also common. These three species are the same ascomycetes previously identified as the dominant fungal decomposers on S. alterniflora on the east coast. Ascomycetes appeared to exhibit varying degrees of host specificity, demonstrated by grouping patterns on phylogenetic trees. Neither the exotic S. alterniflora nor the hybrid supported fungal flora different from that of the native S. foliosa. However, S. densiflora had a significantly different fungal community than the other species, and hosted at least two unique ascomycetes. Significant differences in the fungal decomposer communities were also detected within species (two clones of S. foliosa), but these were minor and may be due to morphological differences among the plants.

Keywords

AscomycetesMarine fungiSpartinaPhaeosphaeriaMycosphaerella

Introduction

The cordgrass Spartina foliosa is native to salt marshes on the west coast of the US, but a number of the other Spartina species have been introduced (e.g., Spartina alterniflora, Spartina densiflora) and are exotic, invasive plants (Ayres et al. 2004). Wind-dispersed pollen of the introduced S. alterniflora has resulted in the formation of a fertile, vigorous hybrid with the native S. foliosa (hereafter referred to as “hybrid Spartina”; Daehler and Strong 1997). Hybrid Spartina is of concern in San Francisco Bay because its population has expanded rapidly to cover previously unvegetated mudflats and it also threatens to invade habitat normally occupied by S. foliosa. S. densiflora, which is native to Chile, was also deliberately introduced to San Francisco Bay (in 1976) as part of a marsh restoration project (Kittelson and Boyd 1997).

Spartina decomposition is an important process in salt marsh ecosystems. Spartina biomass, which represents a large reservoir of organic matter, largely comprises recalcitrant lignins. Fungi, together with bacteria, represent the major mechanism for the breakdown of vascular plant lignins (Benner et al. 1986; Newell and Porter 2000). Previous studies of the fungal community of US salt marshes have focused on the east coast (Blum et al. 2004; Buchan et al. 2002; Lyons et al. 2005; Newell et al. 2000). Several species of ascomycetous fungi have been identified as major decomposers of S. alterniflora blades, based on traditional culture- and microscopy-based methods (Kohlmeyer and Volkmann-Kohlmeyer 2002; Newell 2001a, 2001b; Newell and Porter 2000) as well as molecular approaches (Buchan et al. 2002, 2003; Lyons et al. 2005). Although at least 28 marine ascomycetes have been identified on Spartina (Kohlmeyer and Volkmann-Kohlmeyer 2002), the two most common species are Phaeosphaeria spartinicola and Mycosphaerella sp. strain 2, both of which are involved in the breakdown of lignocellulosic components of the blades (Bergbauer and Newell 1992; Newell and Porter 2000; Newell et al. 1996a, 1996b). Additional species that are typical but less common members of the community include Phaeosphaeria halima, environmental isolate “4clt” (an ascomycetous species that does not yet have a formal taxonomic description), and Buergenerula spartinae (Buchan et al. 2002; Newell 2001a, 2001b).

Studies along the east coast have suggested that plant type is a primary determinant of fungal community (Blum et al. 2004; Kohlmeyer and Volkmann-Kohlmeyer 2001, 2002). Newell and Porter (2000) found the same two ascomycetes (P. spartinicola and a Mycosphaerella) on all samples of standing decaying S. alterniflora blades collected from eight marshes between Maine and Florida, and a third (P. halima) at all sites except those in the northernmost latitudes (Massachusetts and Maine). Blum et al. (2004) compared the fungal communities associated with five species of standing dead marsh plants [three Spartina species (S. alterniflora, Spartina bakeri, and Spartina patens), Phragmites australis and Juncus roemerianus] collected at ten locations along the eastern coast of the US from Maine to Florida. Differences in fungal community composition were strongest when different genera were compared, but slightly different fungal communities were also recorded on the different Spartina species. Latitudinal differences of fungal communities within a plant species were not detectable.

The fungal decomposer communities associated with west coast Spartina species have not been described, and information on the fungi associated with Spartina species other than S. alterniflora is limited. The purpose of this study was threefold: to explore whether the hybridization between exotic S. alterniflora and native S. foliosa affected the composition of the associated fungal assemblage, to investigate variations in the ascomycete communities within and among Spartina species, and to assess the similarity of ascomycetous fungal communities on Spartina alterniflora transplanted to the west coast with those on the east coast.

Materials and methods

Site description and sample collections

Samples of Spartina blades in the “early decay” stage (brown in color, remaining attached to the stem and not collapsed into the sediment) were collected in June 2005 from S. foliosa, S. alterniflora, hybrid Spartina, and S. densiflora plants at three marshes (Blackie’s Pasture, 37°53′43.07″N, 122°29′21.86″W; Cogswell Marsh, 37°38′18.30″N, 122°9′8.20″W; and Hayward Landing, 37°38′46.06″N, 122°9′19.51″W) that border San Francisco Bay. In each marsh, two samples of each grass species present were collected from separate, distinct clones identified previously by genotyping (Ayres et al. 2004). Each sample consisted of a group of 18 blades, which was subdivided into three sub-samples of six blades each. The lowest portion of each blade (6 cm, beginning where the blade attached to the stem) was cut, air-dried overnight, and stored at −20°C until analysis.

Additional samples of S. foliosa were collected in July 2005 from ten marshes around the head of Tomales Bay, an estuary located on the central California coast approximately 50 km northwest of San Francisco (38°06′04.2″N, 122°50′36.0″W). Tomales Bay has not been invaded by S. alterniflora, thereby making it an ideal site from which to collect samples of the native grass. No S. alterniflora, no Spartina hybrid, and no S. densiflora were observed in Tomales Bay, and the site is typically upwind of San Francisco Bay sites.

Polymerase chain reaction amplifications and terminal restriction fragment length polymorphism

Genomic DNA was extracted from the samples using DNA Plant Mini Kits (Qiagen, Valencia, Calif.). Although triplicate sub-samples were available, results of initial analyses were so consistent that only two of the three sub-samples were used. The distribution of ascomycete taxa among Spartina species was assessed by terminal restriction fragment length polymorphism (T-RFLP) analysis of the 18S to 28S internal transcribed spacer (ITS) region (Buchan et al. 2002, 2003).

Phylogenies of the fungal population and presumptive identification were based on DNA sequences of the ITS region determined in parallel using cloned amplicons. ITS regions (Bridge and Spooner 2001; Gardes and Bruns 1993) were amplified with the ascomycete-specific primers ITS1F (5′-CTTGGTCATTTAGAGGAAGTAA-3′) and ITS4A (5′-CGCCGTTACTGGGGCAATCCCTG-3′; Larena et al. 1999). The ITS1F primer used for T-RFLP analysis was fluorescently labeled on the 5′ end with carboxyfluorescein. These primers amplify a product ranging in size from 632 to 713 bp, including the ITS1, 5.8S, and ITS2 regions of the rRNA operon (Torzilli et al. 2006). A previous study (Buchan et al. 2002) found no evidence for genotype variation due to multiple ITS sequences within single individuals for the salt marsh fungi we encountered.

All polymerase chain reactions (PCRs) were carried out with Ready-To-Go PCR beads (Amersham Pharmacia, Piscataway, N.J.), using 0.2 μM of each primer and 50–100 ng of DNA. Thermal cycling reactions for ITS amplifications consisted of an initial 3 min at 95°C followed by 35 cycles of 1 min at 95°C, 30 s at 52°C, and 1 min at 72°C. A final step of 10 min at 72°C was included to complete any partial polymerizations. Products were recovered from a 1% agarose gel using the QiaSpin gel extraction kit (Qiagen). Restriction enzyme digestion of the PCR product was carried out in 20 μl total volume containing 17 μl of purified PCR product and 10 U of HaeIII at 37°C for 3 h. Digested DNA was precipitated in ethanol and suspended in 12 μl of deionized formamide with 1 μl of DNA fragment length standard GeneScan-2500 (TAMRA-labeled; Applied Biosystems).

Terminal restriction fragment (T-RF) lengths were determined on an ABI PRISM 310 (Applied Biosystems, Foster City, Calif.) in GeneScan mode. The automated sequencer detects all fluorescent DNA fragments. If there is a partial digestion by the restriction enzyme, a signal that does not correspond to a true T-RF can be detected. The fact that two sub-samples were analyzed independently (i.e., independent PCRs and independent digestions) helped to guard against such artifacts. In addition, all T-RFLP runs were duplicated. Besides allowing us to eliminate erratic peaks, these additional observations gave us the confidence to retain small T-RF peaks that were present consistently.

T-RFLP peaks were analyzed using a Visual Basic program that reconciles minor shifts in T–RF sizes and quantifies them in terms of the percent of total area under the chromatogram represented by each peak (Stepanauskas et al. 2003). Peaks comprising <1% of total area under the chromatogram were excluded from the analysis. The standardized T-RFLP data sets from each sample were compared quantitatively by principal components analysis (PCA), analysis of similarity (ANOSIM), and similarity percentage (SIMPER) procedures based on fourth-root-transformed data in a Bray-Curtis similarity matrix using Primer 5 software (Primer-E, Plymouth, UK).

Clone library construction and analysis

Clone libraries were constructed from selected samples of Spartina [S. alterniflora and S. densiflora from Blackie’s Pasture, and two S. foliosa clones (4 and 23) from Hayward Landing]. PCR was carried out as described above. The purified PCR product was cloned into an Escherichia coli vector using a TOPO TA cloning kit (Invitrogen, Carlsbad, Calif.) and plated onto LB plates. Inserts were sequenced in a 96-well format by SeqWright DNA Technology Services (Houston, Tex.) using the ITS1F primer.

Approximately 500 bp of sequence were obtained from each clone; 325 bp of each sequence were aligned using ClustalW (http://www.ebi.ac.uk/clustalw/index.html) and used: (1) to predict theoretical HaeIII restriction sites (to be matched to T-RFLP profiles); (2) to associate the ascomycete operational taxonomic units (OTUs) represented in the T-RFLP profiles with sequences in the GenBank database using Basic Local Alignment Search Tool (BLASTn) (http://www.ncbi.nlm.nih.gov/BLAST/); (3) to construct phylogenetic trees. Differences in the composition of the clone libraries were tested for statistical significance using a web-based version of the library shuffle (LIBSHUFF) analysis (Singleton et al. 2001). Sequences from nine ascomycete isolates from the Sapelo Island database (retrieved from Genbank http://www.ncbi.nlm.nih.gov/Genbank/) were also included in the tree as references points. The trees were created with the phylogeny inference package (PHYLIP), BioEdit (http://www.mbio.ncsu.edu/BioEdit/BioEdit.html) (Hall 1999), and MEGA version 4 (Tamura et al. 2007) using evolutionary distances (Kimura algorithm) and the neighbor-joining method.

Results

T-RFLP profiles

The number and percent of total chromatograph coverage of fungal OTU (defined here as a T-RFLP peak accounting for >1% of the area under the chromatogram) identified in this study are presented in Table 1. (Note that duplicate sub-samples gave consistent results, so these were averaged.) The number of T-RFs per profile ranged from 1 to 6 and varied little among collection sites or plant species. This is relatively low compared to previously described samples of S. alterniflora (Buchan et al. 2002; Kohlmeyer and Volkmann-Kohlmeyer 2002; Newell 2001b).
Table 1

Percent coverage of total chromatogram area covered by terminal restriction fragments from Spartina alterniflora, Spartina foliosa, a S. alterniflora × S. foliosa hybrid, and Spartina densiflora collected from four marshes (Blackie’s Pasture, Cogswell Marsh, Hayward Landing, Tomales Bay), San Francisco Bay, California, USA. Values are averages of duplicate samples. Minor peaks Combined values for fragments covering less than 5% of the total area in a single sample

 

63

72

81

91

463

502

528

632

Minor peaks

Blackie’s Pasture

         

 S. alterniflora 1

13

60

18

5

4

 S. alterniflora 2

14

68

15

3

 S. foliosa 1

10

71

7

12

 S. foliosa 2

95

5

 Hybrid 1

4

94

3

 Hybrid 2

19

81

 S. densiflora 1

37

17

46

 S. densiflora 2

7

23

17

29

20

4

Cogswell Marsh

         

 S. alterniflora

23

77

 S. foliosa 1

7

90

3

 Hybrid 1

4

72

9

11

5

Hayward Landing

         

 S. foliosa clone 23

57

43

 S. foliosa clone 4

30

2

53

15

 Hybrid clone 5

18

72

10

 Hybrid clone 25

13

60

28

Tomales Bay

         

 S. foliosa 1

57

39

5

 S. foliosa 5

6

89

6

 S. foliosa 10

15

67

3

10

5

All T-RFLP profiles contained a 72-bp T-RF (covering 23–95% of the total chromatograph area; Table 1). In the few cases where the 72-bp T-RF was not the dominant peak, it was still prominent (S. densiflora collected from two sites at Blackie’s Pasture and S. foliosa clone 4 collected at Hayward Landing). Other T-RFs that were also recovered consistently included a 63-bp (an unknown seen in 24 of 36 samples) and a 528-bp T-RF (Mycosphaerella, present in 20 of 36 samples; Table 1).

There were also differences among samples in the T-RFLP profiles. All S. densiflora samples had two major peaks that were unique to this species (81 and 91 bp). A 502-bp T-RF was present in all samples collected from Tomales Bay and in one sample from Cogswell Marsh. The two clones of S. foliosa sampled at Hayward Landing had different T-RF profiles. Clone 4 yielded three minor T-RFs (85, 123, and 129 bp) totaling 15% not observed in the profiles of S. foliosa clone 23.

We performed a PCA on peak relative abundance to further evaluate differences among TRFLP profiles (Fig. 1). The variables were: plant species, collection site, T-RF size, and percent coverage of each T-RF in each profile. The analysis showed overlap of ascomycete populations amongst all species and sites, but PCA loadings (not shown) indicated that S. foliosa clone 4 from Hayward Landing, and S. densiflora from Blackie’s Pasture contributed the most to diversity among the samples. To confirm that the two S. foliosa clone libraries were significantly different, we conducted a LIBSHUFF analysis (Singleton et al. 2001) as well as an ANOSIM test, which verified that they were different (P < 0.001 and R = 0.424, P < 0.02, respectively). SIMPER analysis showed that clone 4 and S. densiflora were responsible for the highest percentages of dissimilarities among samples, and that these dissimilarities are due to the presence of the three minor T-RFs in the former sample.
https://static-content.springer.com/image/art%3A10.1007%2Fs00442-009-1460-4/MediaObjects/442_2009_1460_Fig1_HTML.gif
Fig. 1

Principal component (PC) plot (PC1 × PC2) of scores of individual samples generated from fingerprint (terminal restriction fragment length polymorphism; T-RFLP) profiles of ascomycete communities associated with four species of early decay Spartina (Spartina alterniflora, Spartina densiflora, Spartina foliosa, and S. alterniflora × S. foliosa hybrids) collected from four sites around San Francisco Bay, California, USA in 2005. Input variables were expressed as percentage of total peak area. BP Blackie’s Pasture, CM Cogswell Marsh, HL Hayward Landing, TBay Tomales Bay, AltS. alterniflora, DensiS. densiflora, FoliS. foliosa, HybridS. alterniflora × S. foliosa hybrid, 25 site 25, 5 site 5, 23 site 23, 4 site 4

Hybrid Spartina

The 72-bp fragment was dominant on all hybrid samples at all sites. ANOSIM and PCA conducted on the T-RF data revealed that, within a site, there was not a significant difference between the hybrid and any other Spartina species sampled at Blackie’s Pasture (R = 0.3, α = 0.05, P < 0.01), Cogswell Marsh (R = 0.4, α = 0.05, P < 0.01), or Hayward Landing (R = 0.3, α = 0.05, P < 0.01; data not shown).

Phylogenetic analysis

Partial ITS sequences were obtained for cloned amplicons from four plant samples: S. alterniflora 1 and S. densiflora 1 from Blackie’s Pasture (92 and 76 sequences, respectively), and S. foliosa clone 23 and clone 4 from Hayward Landing (85 and 92 sequences, respectively). These samples were chosen because their T-RFLP profiles contained a relatively high number of T-RFs.

A surprisingly high percentage (91%) of the sequences obtained from the clone library matched sequences of the three ascomycetes previously identified as dominant decomposers on the east coast (P. spartinicola, Mycosphaerella sp., and P. halima; Buchan et al. 2002, 2003; Table 2) with high similarity (≥98%). Each of these fungal species exhibited a degree of host specificity (Fig. 2a–c). For example, of the 186 sequences that we identified as P. spartinicola, 33% of those collected from S. alterniflora and 16% of those from S. densiflora grouped together with 91% bootstrap support. Eighty-one percent of the sequences obtained from samples of S. foliosa clone 23 grouped together on a distinct branch (64% bootstrap support), and all of the reference sequences from east coast samples of S. alterniflora grouped separately with 98% bootstrap support. These sequences all contained an HaeIII restriction site (GG/CC) at 72 bp. All isolates of P. spartinicola from Sapelo Island, Georgia, also had a restriction site at 72-bp T-RF (Buchan et al. 2002, 2003). We therefore conclude that the dominant 72-bp T-RF in this study represents P. spartinicola.
Table 2

Phylogenetic identity of internal transcribed spacer sequences retrieved from clone libraries determined by BLAST search. Plant species/clone Origin of the sample from which the library was produced; Ascomycete species species of the closest BLAST hit; Percent similarity similarity to the reference sequence; Reference species reference sequence, accession number and origin of the sequence most similar to the query sequence

Plant species/clone

Ascomycete species

Percent similarity

Reference species

S. densiflora

Phaeosphaeria spartinicola

>98%

AF422954: LIF22 (Phaeosphaeria spartinicola)

 

Phaeosphaeria halima

>98%

AF422969: Phaeosphaeria halima strain SAP137

 

Mycosphaerella

>92%

EU301087: Mycosphaerella verrucosiafricana

 

Pleospora pelagica

>95%

AF422995: Pleospora pelagica strain SAP165

S. alterniflora

Phaeosphaeria spartinicola

>98%

AF422954: LIF22 (Phaeosphaeria spartinicola)

 

Phaeosphaeria halima

>98%

AF422991: Phaeosphaeria halima strain SAP161

 

Mycosphaerella

>85%

AF423023: SIF32 (Mycosphaerella sp.2)

 

Other

96%

EU003022: Ascomycete sp. Olrim 149

S. foliosa 23

Phaeosphaeria spartinicola

>98%

AF422954: LIF22 (Phaeosphaeria spartinicola)

 

Phaeosphaeria halima

NA

NA

 

Mycosphaerella

>86%

AF423023: SIF32 (Mycosphaerella sp.2)

 

Other

97%

AY035665: Entrophospora sp. JJ60 18S

S. foliosa 4

Phaeosphaeria spartinicola

>98%

AF422954: LIF22 (Phaeosphaeria spartinicola)

 

Phaeosphaeria halima

>99%

AF422991: Phaeosphaeria halima strain SAP161

 

Mycosphaerella

>93%

AF423023: SIF32 (Mycosphaerella sp.2)

 

Other

98%

EF565863.1: Alternaria sp.4F

  

98%

Expression vector pESPBAC

https://static-content.springer.com/image/art%3A10.1007%2Fs00442-009-1460-4/MediaObjects/442_2009_1460_Fig2_HTML.gif
Fig. 2

Evolutionary relationships of a 189 Phaeosphaeria spartinicola operational taxonomic units (OTUs), b 71 Phaeosphaeria halima OTUs, and c 64 Mycosphaerella OTUs retrieved from decaying Spartina blades. Shaded bars represent the percent of the sequences in the combined libraries from each plant that match the focal ascomycete with >98% similarity. The evolutionary history was inferred using the neighbor-joining method. The optimal tree with the sum of branch length = 0.83390794 is shown. Bootstrap values greater than 50% are shown. The tree is drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree. Evolutionary distances were computed using the maximum composite likelihood method and are in the units of the number of base substitutions per site (scale bar). All positions containing gaps and missing data were eliminated from the dataset. There was a total of 437 positions, b 376 positions and c 286 positions in the final dataset. Phylogenetic analyses were conducted in MEGA4 with a Mycosphaerella species as an outgroup in a, b and with Phaeosphaeria halima as an outgroup in c. SAP Previously identified sequences from Sapelo Island, Georgia

Twenty percent of the clone sequences were highly similar (>98%) to P. halima (Table 2). P. halima does not have a HaeIII restriction site within the ITS region. If present on the T-RFLP profiles, it would produce a peak corresponding to the full-length amplicon, which could range from 632 to 713 bp (Torzilli et al. 2006). Detection of P. halima on T-RFLP profiles is therefore difficult, as other sequences may also be present that lack a restriction site and as a result would also produce a peak within this range. Analysis of the clone sequences showed that the library from S. foliosa clone 4 contained the highest percentage of sequences related to P. halima (62%), followed by libraries from S. densiflora (9%) and S. alterniflora (5%). None of the sequences retrieved from the S. foliosa 23 library were related to P. halima. Clustering patterns on the P. halima phylogenetic tree (Fig. 2b) once again suggest host specificity, but also a high degree of minor sequence variation as indicated by the large number of clades shown on the tree. Although all sequences included in the figure were at least 98% similar to P. halima, many were less than 100% similar to each other (indicated by percentage bars on the tree) and hence did not group in the same clade.

Finally, 17% of the sequences collected from the clone library were most similar to the ITS sequence from an isolate of Mycosphaerella sp. strain 2 (SIF32), collected from standing dead S. alterniflora (Buchan et al. 2002; Fig. 2c; Table 2). Restriction sites within these sequences were identified at 85, 528, 529, 530, 531, 532, and 533 bp. Interestingly, sequences retrieved from S. densiflora consistently matched the sequence from Mycosphaerella verrucosiafricana (>92% similarity), which is a different species than that identified on the east coast. These sequences grouped together with >96% similarity on a distinct branch of the tree (Fig. 2c) and contained a restriction site at 91 bp. As was the case for P. spartinicola, reference sequences from east coast samples clustered separately.

A few other ascomycete taxa were also identified in the clone libraries (Table 2). Sixteen sequences retrieved from S. densiflora samples were >95% similar to the Pleospora pelagica sequence (SAP165) retrieved from a Sapelo Island sample. Two sequences retrieved from S. alterniflora were >96% similar to an ITS sequence from the ascomycete Olrim 149. One sequence retrieved from S. foliosa 23 was most similar to an Enthrophospora sp. sequence, and one sequence from S. foliosa 4 matched an Alternaria sp. Sequences that appeared only once or twice are likely extremely minor or even transient players in the microbial communities of the Spartina sampled here.

Comparison of clone libraries and chromatographs

Several fragments were observed in the T-RFLP data that were not represented in the clone library. These include T-RFs with lengths of 63, 81, 123, 129, 463, and 632 bp (Table 1). Absence of sequences with these cut sites from libraries may result from biases against these sequences being incorporated into clones, which would skew amplicon representation in the clone library. This discrepancy is unlikely to result from PCR biases since the same primers were used for T-RFLP analysis and library generation. It is unlikely that complex mixtures of amplified rRNA genes are cloned with uniform efficiency, and it is therefore often assumed that cloning systems generally influence the relative abundance of individual sequences in gene libraries (Osborn et al. 2000). Finally, it is possible that these restriction fragments represent rare species that were simply missed in the random selection of clones chosen for sequencing.

Discussion

This paper used analysis of T-RFLPs and clone libraries to evaluate the ascomycete fungal communities associated with several Spartina species found on the west coast of the United States. Our analysis indicates that there are no significant differences among the ascomycete communities of S. alterniflora, S. foliosa, and hybrids of these two species at any of the sites where all three were growing. Instead, we saw the most variance in fungal communities within and among Spartina species. S. densiflora hosted the most divergent ascomycete communities among the Spartina species sampled (Tables 1, 2), containing the only sequences identified as P. pelagica and Mycosphaerella verrucosiafricana. PCA of all samples (Fig. 1) showed that S. densiflora samples group separately along the PC2 axis, indicating that it harbors a distinct ascomycete community.

Both T-RFLP results (Table 1) and sequence analysis (Table 2; Fig. 2a, c) showed that the ascomycete communities from the two clones of S. foliosa sampled from the same site (Hayward Landing, clone 4 and clone 23) were distinct. We observed differences in the growth forms of these two clones in the field. The blades of clone 23 (as well as all other samples of S. foliosa collected) were short (<6 cm), thin, and very dark brown in color. Leaf blades from clone 4 were longer (>6 cm), broader, and lighter brown in color. Because the standard length of leaf processed for DNA extraction was 6 cm, the entire leaf collected from clone 23 was used for analysis, whereas only the bottom 6 cm of leaves from clone 4 was processed. Also, although areas of dense monocultures of S. foliosa were noted along Hayward shoreline, the two clones sampled from this area were growing in discrete circular patches separated by unvegetated mud. Clone 23 was growing surrounded by many other clones at the end of a small peninsula protruding from the Hayward shoreline. Conversely, clone 4 was completely isolated, away from the shoreline in a mudflat approximately ~400 m north of clone 23. It is possible that the lighter color and different morphology that were observed in the leaves of clone 4 may indicate different chemical composition, a factor that could influence the decomposer community (Newell et al. 1989). Their locations in proximity to the shoreline would almost certainly affect the drying and wetting cycles that the plants experience, as well.

Perhaps our most significant finding was that more than half of the ITS sequences retrieved from California Spartina samples shared >98% sequence similarity to an isolate of P. spartinicola (LIF22) from Sapelo Island, Georgia. Moreover, all T-RFLP profiles contained a prominent peak at 72 bp (Table 1). P. spartinicola is the dominant ascomycete on east coast S. alterniflora (Buchan et al. 2002; Newell et al. 2000) and has also been found on Spartina anglica in Ireland (Bacic et al. 1998). Many of the remaining clones (129; 38%) shared sequence similarities (>98 and >85%, respectively) to two other ascomycetes, Phaeosphaeria halima (Fig. 2b) and Mycosphaerella sp. strain 2 (Fig. 2c), that were also previously described as dominants on east coast Spartina.

We find it intriguing that not only are the same major species (P. spartinicola, Mycosphaerella, and P. halima) of ascomycetous fungi found on various species of Spartina across such a broad geographic area, but that those same species dominate. We acknowledge there could well be variations in the relative importance of different fungal OTUs or even OTUs that were not present in these samples, and that fungal assemblages may vary over time and with environmental condition. However, several seasonal studies have been conducted by other researchers (e.g., Gessner 1977; Newell 2001a, 2001b; Buchan et al. 2003) and none have found a significant temporal difference in the dominant members of the fungal decomposer community associated with different seasons or stages of decay. In addition, Buchan et al. (2003) found no significant correlations between the dominant fungal taxa and temperature, salinity, or rainfall. Since our collections were all done in summer (June and July) on blades at a similar stage of decay, we believe that samples from different sites are comparable with respect to the dominant members of the community.

Few animal or plant communities experience such consistency across such a large biogeographic range. Our results suggest that the dominant ascomycetes are extremely successful and/or competitive. P. spartinicola in particular which has been shown to be an important degrader (Bergbauer and Newell 1992) appears able to live and thrive in a very wide range of locales. It would be quite interesting to further investigate these results, testing different Spartina species across a wider range to determine whether these ascomycete species dominate their decomposition globally.

Acknowledgments

The authors would like to thank D. Ayres for her help in locating and identifying Spartina clones in San Francisco Bay. We would also like to thank J. Campbell and M. A. Moran for comments on earlier drafts of this manuscript. This is a contribution from Georgia Coastal Ecosystems Long-Term Ecological Research and was supported by NSF OCE 99-82133 and OCE-0620959.

Copyright information

© Springer-Verlag 2009