Four new species of Pyropia (Bangiales, Rhodophyta) from the west coast of North America: the Pyropia lanceolata species complex updated

Abstract Recent molecular studies indicate that the Pyropia lanceolata species complex on the west coast of North America is more speciose than previously thought. Based on extensive rbcL gene sequencing of representative specimens we recognize seven species in the complex, three of which are newly described: Pyropia montereyensis sp. nov., Pyropia columbiensis sp. nov., and Pyropia protolanceolata sp. nov. The new species are all lanceolate, at least when young, and occur in the upper mid to high intertidal zone primarily in winter and early spring. Pyropia montereyensis and Pyropia columbiensis are sister taxa that are distributed south and north of Cape Mendocino, respectively, and both occur slightly lower on the shore than Pyropia lanceolata or Pyropia pseudolanceolata. Pyropia protolanceolata is known thus far only from Morro Rock and the Monterey Peninsula, California; it occurs basally to the other species in the complex in the molecular phylogeny. A fourth newly described species, Pyropia bajacaliforniensis sp. nov., is more closely related to Pyropia nereocystis than to species in this complex proper. It is a thin species with undulate margins known only from Moss Landing, Monterey Bay, California, and northern Baja California; it also occurs in the high intertidal in spring. Porphyra mumfordii, a high intertidal winter species that has frequently been confused with species in the Pyropia lanceolata complex, has now been confirmed to occur from Calvert Island, British Columbia, to Pescadero State Park, California.


Introduction
Th e foliose Bangiales are one of the best-studied groups of marine red algae occurring on the west coast of North America. Th e fi rst two species to be named from the region were two of the most common, Porphyra perforata J. Agardh (1883) and Porphyra nereocystis C.L. Anderson (Blankinship & Keeler, 1892). Hus (1900Hus ( , 1902 summarized knowledge of the genus on the Pacifi c Coast, recognizing eleven species and describing three new forms of Porphyra C. Agardh, the genus to which all foliose Bangiales belonged at the time. One of those new forms, P. perforata f. lanceolata Setchell & Hus in Hus (1900), was erected to accommodate lanceolate forms that were dioecious; this taxon was later raised to specifi c status in Smith and Hollenberg (1943: 213), who also added two more species of Porphyra to the fl ora. It was Krishnamurthy (1972) who signifi cantly revised the genus in the region and added seven new species, mostly from Washington State. A summary of knowledge at the time was provided by Conway et al. (1975), with detailed descriptions of Pacifi c Northwest species of Porphyra with emphasis on British Columbia and Washington State; their work was updated by Garbary et al. (1981).
Studies up to then mostly utilized thallus morphology and the pattern of reproductive cell disposition and division as defi ning features for species. Mumford and Cole (1977) added chromosome numbers as a useful feature, and Cole (1990, 1992a, b, c) and Lindstrom (1993) utilized isozymes in addition to morphology, chromosome numbers, biogeography and habitat as characters for separating and recognizing even more species.
Th e taxonomy of foliose Bangiales entered a new phase with the application of DNA sequencing methods. Lindstrom and Fredericq (2003) sequenced the chloroplast rbcL gene of many West Coast species, and Lindstrom (2008) included numerous additional specimens, indicating the need to describe even more species, as did Kucera and Saunders (2012) utilizing the mitochondrial 5´end of the COI gene. Sequencing also indicated that a wholesale revision of the order was needed (fi rst suggested by Oliveira et al. 1995). Th is led to a revision of the genera of foliose Bangiales by Sutherland et al. (2011), redefi ning, resurrecting or creating eight genera of bladed Bangiales. Among these eight genera, four (Boreophyllum S.C. Lindstrom, Fuscifolium S.C. Lindstrom, Porphyra and Pyropia J. Agardh) occur on the west coast of North America, and among these Pyropia is by far the most speciose.
Th e resurrected genus Pyropia contains a number of clades that are resolved with substantial support, and many of these clades are biogeographically circumscribed (Sutherland et al. 2011). One such clade is the northeast Pacifi c P. lanceolata-P. pseudolanceolata complex, fi rst identifi ed as such by Lindstrom and Cole (1992b), who recognized that a number of species were confused under these names. Members of this clade, like other species of Pyropia, have monostromatic blades. As resolved by Sutherland et al. (2011) Lindstrom,and Pyropia sp. 485, indicating that at least two species are as yet undescribed and suggesting uncertainty over the identity of Py. lanceolata (Setchell & Hus) S.C. Lindstrom. In the present study, we analyzed rbcL and 18S rRNA (SSU) gene sequences from recently collected specimens belonging to this clade from the west coast of North America extending from Baja California to Alaska. We also include the closely related northeast Pacifi c species Py. nereocystis and Py. kanakaensis (Mumford) S.C. Lindstrom (Lindstrom 2008, Sutherland et al. 2011, and we analyzed short DNA sequences from the type sheets of P. lanceolata and P. hiberna to resolve their relationship, and to determine whether any of the undescribed species could be the same as one of these species. Th ese new data support the recognition of at least four additional species. Below we discuss these species, their relationships to each other, and the characters that distinguish them.

Materials and methods
Specimens were collected by the authors or by those named in the Acknowledgments (Table 1, Suppl. material 1). Collections were made along the west coast of North America from Baja California, Mexico, to the western tip of the Aleutian Islands, Alaska, between 1992 and 2014. Upon collection, the specimens were damp-dried and then desiccated in silica gel. Pieces or separate specimens were pressed to make herbarium vouchers, which are deposited in UBC or UC. Silica-gel dried specimens were returned to the lab, where they were extracted following the CTAB protocol as implemented by Lindstrom and Fredericq (2003). PCR amplifi cation and sequencing of the rbcL gene was carried out as described in Lindstrom (2008) except that KitoF1 (5' ATGTCTCAATCCGTAGAATCA 3') was used as the forward primer rather than F57. DNA from type material of P. lanceolata and P. hiberna was extracted, amplifi ed and sequenced following the protocol described in Lindstrom et al. (2011), except for using 3X the primer concentration used previously. Th e type fragments were extracted in a separate laboratory (Hartnell College) and processed employing the precautionary steps proposed by Hughey and Gabrielson (2012). For amplifi cation of type material, primers F625 (5'CTCACAACCATTTATGCGTTGG 3') and R900 (5'GCGAGA-ATAAGTTGAGTTACCTG 3') were cycled together.
Sequences of the rbcL gene of Pyropia sp. FAL from Playa Saldamando, Baja California, Mexico, HQ687535, and Pyropia sp. MIG from Faro de San Miguel, Baja California, Mexico, HQ687536, were also included in the analyses because of their close relationship to P. kanakaensis and P. nereocystis (Sutherland et al. 2011) and because of the identity of Pyropia sp. MIG with one of our unknown specimens. We selected two specimens of Pyropia sp. (AB118586 and AB287965) as outgroups based on their close genetic identity to Py. nereocystis using the GenBank blastn algorithm (accessed 06 Sept 2014).
Sequences were aligned using BioEdit version 7.0.9.1 (Hall 1999). Maximum parsimony (MP) analysis was performed using PAUP* 4.0b10 (Swoff ord 2002) as Table 1. Specimens for which the rbcL gene was sequenced in this study and used in the phylogenetic analyses. All herbarium vouchers are deposited in UBC unless noted otherwise. Numbers indicate the total number of specimens with the identical sequence (see Suppl. material 1). Specimens in brackets were included in initial analyses but excluded from the analysis shown in Fig. 1. Specimen P814 in Fig. 1  implemented by Lindstrom and Fredericq (2003). Maximum likelihood (ML) was performed using RAxML 7.2.6 [as implemented on the T-rex website (http://www. trex.uqam.ca/index.php?action=raxml; Stamatakis 2006, Buc et al. 2012, and data were partitioned by codon position. Bayesian phylogenetic analyses were performed on the Bio-Linux7 platform (Field et al. 2006) with MrBayes 3.2.1 (Huelsenbeck et al. 2001, Ronquist andHuelsenbeck 2003). We followed the MrBayes 3.2 manual, which recommends continuing analyses by increasing the number of generations until the average standard deviation of split frequencies drops below 0.01. All runs were performed using a sample frequency of 10 with two independent analyses. To calculate the Potential Scale Reduction Factor and posterior probabilities, the sump and sumt burn-in values were set to discard 25% of the samples.

Results
Th irty-seven rbcL gene sequences (Table 1) were included in the phylogenetic analyses that generated Fig. 1. Both MP and ML generated the same tree topology, as did Bayesian analysis. Several unique sequences were omitted from the analyses after it was determined that their omission did not alter the topology of the phylogenetic tree. In addition to these sequences, the Suppl. material 1 includes 186 additional specimens that were identical to those in Fig. 1. With AB118586 and AB287965 as outgroup species, three major clades are apparent, the Py. nereocystis clade, the Py. kanakaensis clade and the Py. lanceolata clade (formerly called the P. lanceolata-P. pseudolanceolata complex). Within the Py. lanceolata clade, Py. protolanceolata diverges fi rst. Th is species is sister to Py. pseudolanceolata, then Py. lanceolata, but this order of divergence is without support. Th e clade is terminated by two pairs of sister taxa, the closely related Py. montereyensis and Py. columbiensis species pair, and the somewhat more distantly related Py. conwayae and Py. fallax pair. Both of these species pairs represent a southern and northern species, as is also the case for Py. lanceolata and Py. pseudolanceolata. For the most closely related pair, Py. montereyensis and Py. columbiensis, the former has to date only been found south of Cape Mendocino whereas the latter has only been collected from Cape Mendocino north; thus these species do not appear to overlap in their distributions. In the case of Py. conwayae and Py. fallax, the species overlap in distribution between southern Vancouver Island and southern Oregon. Of these species pairs, the former pair is more constrained in its distribution, occurring only between southern California and central British Columbia whereas the latter pair extends from central California to at least the westernmost Aleutian Island. For Py. lanceolata and Py. pseudolanceolata, this older species pair shows an even wider area of overlap, between Sitka Sound, AK, and Crescent City, CA. All species in the Py. lanceolata clade occur on strongly supported branches, and all but Py. protolanceolata show some intraspecifi c variation (to 0.4%) in their rbcL sequences (only two specimens of Py. protolanceolata were sequenced due to the infrequency of collection). Th e nonoverlapping intraspecifi c versus interspecifi c divergence, also referred to as the ''barcode gap'', allows specimens to be assigned unambiguously to genetic clusters that constitute putative genetic species (Le Gall and Saunders 2010).
In the Py. nereocystis clade, Py. nereocystis is sister to two divergent species. Pyropia sp. has been collected several times in early winter from the uppermost intertidal on the Monterey Peninsula; it is the subject of a separate study and will be described there. Pyropia bajacaliforniensis, the other species, has been collected in late spring on the central California and northern Baja California coasts. Th e type specimen, described below, diverges from two other collections by 0.3% (4 base pairs); this level of divergence is within the typical species variation exhibited by the rbcL gene in foliose Bangiales of up to 0.4% Broom 2010, Mols-Mortensen et al. 2012) although levels up to 1% have been reported for a few species (Lindstrom 2008).
Pyropia kanakaensis terminates its own long branch, suggesting a long evolutionary history separate from its closest relatives. It also shows signifi cant within species variation.
We also sequenced the 18S rRNA gene in representatives of these species (Table 1, Suppl. material 1) to complement the data in Sutherland et al. (2011). Th ere was relatively little variation among species and little structure to the phylogenetic tree except for weak support for sibling relationships between Py. nereocystis and Py. kanakaensis and between Py. lanceolata and Py. pseudolanceolata.
Characters of the species in the Py. lanceolata clade are summarized in Table 2. Most specimens are lanceolate with slightly undulate margins. All are monostromatic with one chloroplast per cell although chloroplast division prior to cell division can give the appearance of cells being vegetatively diplastidial. Among the species, only  Py. fallax is monoecious, with spermatangial patches or streaks among pale red zygotosporangia, which occur in submarginal patches, mottles, streaks or hieroglyphs. Th e remaining species are almost invariably dioecious, with spermatangia occurring along cream-colored margins and with the red zygotosporangia occurring along the margin and across the distal end of the thallus in patches usually intermixed with vegetative cells, giving the appearance of red hieroglyphs. All species occur on rock, often near sand. Below we describe in detail the previously unnamed species in this clade, as well as a new species in the Py. nereocystis clade. Fig. 2 Description. Th alli lanceolate and acuminate (occasionally oblanceolate) when young, becoming ovate to nearly orbiculate and often cleft when post-reproductive, base cuneate to strongly umbilicate when old; 50-75 mm thick when dried and young, 90-110 mm thick when old; males to at least 2.3 cm wide and 69 cm long; females to at least 4.8 cm wide and 68 cm long (although usually narrower; to 10 cm broad when old); color uniform throughout the thallus except for reproductive areas, olive green when fresh, drying to grayish or brownish purple. Th alli dioecious. Spermatangia in packets of 2-4 × 2-4 × 8-16. Zygotosporangia in packets of 2-4 × 2-4 × 4-8. Habitat: mid to high intertidal rock, usually associated with sand. Phenology: Winter to mid spring. Distinguished from other species of Pyropia by unique rbcL and 18S rRNA gene sequences.
Isotypes. UBC A90632. Etymology. Th is species is named for the biogeographic region in which it is found following the boundaries of Croom et al. (1995) more closely than those of Valentine (1966).
Distribution. Fort Bragg to just south of Ventura Beach, California, USA. We did not obtain an SSU sequence from type material of this species. Th e SSU sequence in GenBank (KP903907) for this species is from another Monterey Peninsula site: Carmel River State Beach.
Pyropia columbiensis S.C. Lindstrom, sp. nov. Fig. 3 Description. Th alli lanceolate when young, becoming somewhat ovate (rarely obovate) when mature; base cuneate, becoming umbilicate; 50-115 mm thick; males to at least 5.5 cm wide and more than 31 cm long; females to 12 cm wide and more than  28 cm long, but thalli mostly narrower; color uniform throughout the thallus except for reproductive areas, olive-green when fresh, drying to grayish or brownish purple. Th alli dioecious. Spermatangia in packets of 2-4 × 2-4 × 8. Mature zygotosporangia in packets of 2-4 × 2-4 × 2-4. Habitat: mid to high intertidal rock, usually associated with sand. Phenology: winter to early spring (a few thalli may persist as late as mid summer). Distinguished from other species of Pyropia by unique rbcL and 18S rRNA gene sequences.
Holotype. Saxicolous in the upper mid intertidal on rocks partially buried in sand at the south end of West Beach, Calvert Island, British Columbia, Canada (51°39.14'N 128°08. Etymology. Th is species is named for the biogeographic region in which it is found, using the terminology of Valentine (1966), but with a modifi cation of the boundaries to extend from Cape Mendocino, California, to the central coast of British Columbia. It also commemorates the centenary of the University of British Columbia herbarium, which was established in early 1916.
Distribution. Calvert Island, British Columbia, Canada, to Cape Mendocino, California, USA.
Pyropia montereyensis and Py. columbiensis are essentially morphologically identical and represent the southern and northern species of a vicariant pair, respectively. Pyropia protolanceolata S.C. Lindstrom & J.R. Hughey, sp. nov. Fig. 4 Description. Th alli linear to lanceolate, base cuneate; 28-65 mm thick; to 1.2 cm wide and 16 cm long; color uniform throughout the thallus except for reproductive areas: dusky rose. Th alli dioecious. Spermatangia in packets 2 × 2 × 8. Zygotosporangial thalli not observed Habitat: very high intertidal, above Py. lanceolata and Py. montereyensis when they co-occur. Phenology: Winter to early spring. Distinguished from other species of Pyropia by unique rbcL and 18S rRNA gene sequences.

Discussion
Molecular phylogenetic analysis of the foliose Bangiales indicates that Pyropia is the most speciose genus in the order; it also displays the most morphological variation and the widest geographical distribution. Still, there are many geographically restricted clades (Fig. 1, Sutherland et al. 2011). Th is indicates that much speciation in the order has occurred in particular geographical regions. Th e Py. lanceolata clade and its close relatives (Py. kanakaensis, the Py. nereocystis clade) are an example of a geographically restricted clade, with species known thus far only from the northeast Pacifi c, from Baja California, Mexico, to the Aleutian Islands, Alaska. Several of the species are highly restricted geographically: Py. bajacaliforniensis (in the related Py. nereocystis clade) is known only from the Moss Landing area of Monterey Bay, CA, and northern Pacifi c Baja California. Other species are limited to particular areas of coastal California: Py. protolanceolata thus far known only from Morro Bay and Spanish Bay, California, and Py. montereyensis from southern to northern California south of Cape Mendocino. In contrast, Py. lanceolata and especially Py. pseudolanceolata are widely distributed, occurring from California to Alaska although Py. lanceolata is replaced by Py. pseudolanceolata at many sites from British Columbia north. As with all geographic records, these are based on collections to date and are subject to revision due to both more intense collecting eff orts in the region as well as changes in distributions due to changing environmental conditions. Th e phylogeny of this group of related species suggests a number of patterns that have occurred in the evolution of some of the species. For example, the diplastidial condition in vegetative cells of Py. kanakaensis has also been observed in species in the Py. lanceolata complex (Smith andHollenberg 1943, Lindstrom andCole 1992b), where division of the chloroplast seems to precede by days or even weeks cell division associated with reproductive cell formation. In species of the Py. lanceolata clade, the two chloroplasts remain close together (Smith and Hollenberg 1943, Fig. 10) whereas they move to opposite ends of the cell in Py. kanakaensis (Mumford 1973).
Although the habitat of Py. nereocystis as an obligate epiphyte on the kelp Nereocystis is unique, and Py. kanakaensis occurs primarily in the lower mid intertidal, the remaining species have adapted to the rigors of the mid to high intertidal. In the Py. lanceolata clade proper, Py. lanceolata, Py. pseudolanceolata, and Py. protolanceolata are mostly restricted to the high intertidal and are among the highest-occurring species of seaweeds, as are Py. bajacaliforniensis and Pyropia sp. in the Py. nereocystis clade. Pyropia fallax can occur in the high intertidal but also extends into the mid intertidal, where its sister taxon, Py. conwayae, is found. Where they co-occur, Py. conwayae usually occurs at a slightly lower elevation than Py. lanceolata. Pyropia columbiensis and Py. montereyensis also occur primarily in the upper mid to high intertidal although perhaps not as high as Py. lanceolata and others. Exact elevation of occurrence depends on many factors such as wave exposure, direction the rock is facing as well as season and latitude (and longitude for northern populations). Although thalli can be common on bedrock, when that is the predominant habitat in an area, all of the species can also be abundant on rock protruding from wave-swept sandy shores.
Whereas Py. bajacaliforniensis and Py. kanakaensis are spring and spring-summer species, respectively (appearing on the shore ~April, disappearing in June in the case of the former, and persisting as late as November for the latter), the remaining species, including Py. nereocystis, appear to be winter-spring species, reaching their peak abundance from February to April, and then depending on the species and the location, disappearing from the shore from April to August or later (these later dates occurring for populations near the northern limits of the species).
Because of their similar morphologies, habitats, seasonalities and overlapping distributions, species in this complex have been frequently confused. Much of what has been published on Py. pseudolanceolata in particular has actually applied to diff erent species. For example, the haploid chromosome number reported by Mumford and Cole (1977) for this species was actually for Porphyra mumfordii, and the culture conditions for conchocelis growth and maturation reported by Waaland et al. (1990) were probably for Py. conwayae. Moreover, the rbcL sequence reported for this species by Lindstrom and Fredericq (2003) was that of Py. lanceolata, as were the culture conditions reported for conchospore release .
Th ere have also been problems with the identity of Pyropia lanceolata. Krishnamurthy (1972) lectotypifi ed Porphyra perforata f. lanceolata, the basionym of Porphyra lanceolata, with UC 95720 (collected by Setchell in Carmel Bay, California on 11 Jan 1899), but Lindstrom and Cole (1992b) felt that the specimens on the sheet did not accord with Smith's description or with the major portion of Setchell & Hus' description. Th ey therefore designated MO 24356 in UC (Lindstrom and Cole 1992b, Fig. 8), collected by H.T.A. Hus at Land's End, San Francisco, California, as lectotype since that collection better fi t with the original description. Th e latter contains two outer specimens that are linear in habit, and four inner specimens that are lanceolate. Since modern DNA methods allow the sequencing of historic material, we sequenced a 251 bp region of the rbcL gene for the two outer and two inner specimens on MO 24356, a single specimen on UC 95720 (https://ucjeps.cspace.berkeley.edu/ucjeps_project/imageserver/blobs/c68a16ad-5ba5-4c15-8d8d/derivatives/OriginalJpeg/content), which all showed a similar morphology, as well as fi ve of the six specimens on the type sheet of Py. hiberna (UBC A80269: http:// bridge.botany.ubc.ca/herbarium/details.php?db=ubcalgae.fmp12&layout=ubcalgae_ web_details&recid=210219&ass_num=A80269), a species closely related if not identical to Py. lanceolata (Sutherland et al. 2011). Seven of the specimens fell within the variation observed for contemporary collections of Py. lanceolata (Table 3). Specifi cally, UC 95720 from Carmel Bay, Monterey Peninsula, and four of the UBC A80269 specimens, all from Pacifi c Grove, Monterey Peninsula, had sequences identical to the two contemporary specimens from Pacifi c Grove. Th e contemporary Monterey Peninsula specimens diff ered from specimens of Py. lanceolata from other geographical regions by 0.3%, an amount insuffi cient to recognize them as a separate species. Th e distinctness of Monterey Peninsula genotypes within a species has been observed for other organisms (e.g., Mastocarpus papillatus (C. Agardh) Kützing, Lindstrom et al. 2011). Th e two identical inner specimens (one female and one male) on the lectotype sheet of MO 24356 from Land's End diff ered by 2 bp from the fi ve Monterey Peninsula specimens noted above, but were identical to other Py. lanceolata specimens from outside of the Monterey Peninsula. In contrast, the two outer specimens on the same sheet (female far left and male far right) diff ered from these two inner specimens by 5 bp over the 251 bp region (but by only 3 bp from Monterey Peninsula Py. lanceolata and by only 2 bp from all Py. conwayae sequenced). Th us, at this time, we are unable to assign a name to the two outer linear specimens on the sheet of MO 24356. Since MO 24356 is heterotypic, we therefore narrow the lectotypifi cation of MO 24356 to the middle four specimens. Our results also confi rm that Py. hiberna S.C. Lindstrom & K.M. Cole, 1992: 435 is a heterotypic synonym of Py. lanceolata. Th e fi fth specimen on the type sheet of Py. hiberna did not match among any described foliose Bangiales sequences but did match a recent collection we recognize here as Pyropia sp. (to be described later in a separate paper). Pyropia lanceolata was identifi ed as Unknown #3 in Lindstrom (2008).
In the earlier paper on the Py. lanceolata complex (Lindstrom and Cole 1992b), they included Porphyra mumfordii as one of the species. Th is was in part because this entity had previously been misidentifi ed as P. pseudolanceolata (Conway et al. 1975, Mumford andCole 1977). Subsequent DNA sequencing studies have shown that these species are unrelated Fredericq 2003, Lindstrom 2008), despite the fact that P. mumfordii continues to be easily confused with species in the Py. lanceolata complex in the fi eld because of similar habitat, seasonality and habit (see Lindstrom and Cole 1992b for a detailed comparison of these species). In conjunction with the present study, we have extended the range of P. mumfordii south to Pescadero State Park, California, and north to Calvert Island, British Columbia (Suppl. material 1).
As noted above, the species in the Py. lanceolata clade show little morphological diff erentiation. Th erefore, the following key to species in this clade relies heavily on geographic distribution and on modest diff erences in seasonality and elevation on the shore.