Specimens examined were collected in a hilly forest near the cemetery of Sousa, in Puerto Plata Province, Dominican Republic, and are deposited in the Herbarium of Jardín Botánico Nacional of Santo Domingo, Dr. Rafael Ma. Moscoso (JBSD) (acronym from Thiers 2019), while “ANGE” and ‘‘MG’’ refer to the personal herbarium of Claudio Angelini and Matteo Gelardi, respectively. Herbarium numbers are cited for all collections from which morphological features were examined. Author citations follow the Index Fungorum, Authors of Fungal Names (www.indexfungorum.org/authorsoffungalnames.htm).

Macroscopic descriptions and ecological information, such as habitat notations, time of fruiting and associated plant communities accompanied the detailed field notes of the fresh basidiomata. Color terms in capital letters (e.g. Myrtle Green, pl. VIII) are from Ridgway (1912). Photographs of collections were taken in the natural habitat using a Nikon Coolpix 8400 digital camera. Microscopic anatomical features were observed and recorded from revived dried material; sections were rehydrated either in water, 5% potassium hydroxide (KOH) or in anionic solution saturated with Congo red. All anatomical structures were observed and measured from preparations in anionic Congo red. Colors and pigments were described after examination in water and 5% KOH. Measurements were made at 1000× using a calibrated ocular micrometer (Nikon Eclipse E200 optical light microscope). Basidiospores were measured directly from the hymenophore of mature basidiomes, dimensions are given as (minimum) average ± standard deviation (maximum), Q = length/width ratio with the extreme values in parentheses, Qm = average quotient (length/width ratio) ± standard deviation and average spore volume was approximated as a rotation ellipsoid [V = (π.L.W2)/6 ± standard deviation]. The notation [n/m/p] indicates that measurements were made on “n” randomly selected basidiospores from “m” basidiomes of “p” collections. The width of each basidium was measured at the widest part, and the length was measured from the apex (sterigmata excluded) to the basal septum. Metachromatic, cyanophilic and iodine reactions were tested by staining the basidiospores in Brilliant Cresyl blue, Cotton blue and Melzer’s reagent, respectively. Line drawings of microstructures were traced in free-hand based on digital photomicrographs of rehydrated material.

Genomic DNA was isolated from 10 mg of four dried herbarium specimen (Table 1), by using the DNeasy PlantMini Kit (Qiagen, Milan Italy) according to the manufacturer’s instructions. PCR amplifications were performed with the primers ITS1F/ITS4 for the nrITS region (White et al. 1990, Gardes and Bruns 1993), LR0R and LR5 for the nrLSU region (Vilgalys and Hester 1990) and the reverse complement of bRPB2-6R2 (Matheny et al. 2007) and bRPB2-7.1R2 (5'– CCCATNGCYTGYTTVCCCATDGC –3') or RPB2-B-F1 and RPB2-B-R (Wu et al. 2014) for partial rpb2. Amplification reactions were performed in a PE9700thermal cycler (Perkin-Elmer, Applied Biosystems) following Vizzini et al. (2015). The PCR products were purified with the AMPure XP kit (Beckman Coulter) and sequenced by MACROGEN Inc. (Seoul, Republic of Korea). The sequences were submitted to GenBank (http://www.ncbi.nlm.nih.gov/genbank/) and their accession numbers are reported in Table 1.

The sequences obtained in this study were checked and assembled using Geneious v. R 11.1.4 (Kearse et al. 2012) and compared to those available in GenBank database (https://www.ncbi.nlm.nih.gov/genbank/) by using the BLASTN algorithm (Altschul et al. 1990). A general combined Maximum likelihood tree including all the Boletaceae sequences present in GenBank and UNITE (http://unite.ut.ee/) databases was generated to detect the phylogenetic position of our collections in the major clades of Boletaceae as circumscribed by Wu et al. (2014) (tree not shown). Consequently, phylogenetic analyses were restricted to the major clade including Neoboletus sequences (Pulveroboletus group, Fig. 1).

Figure 1. 

Phylogeny of the Pulveroboletus group based on a Bayesian and Maximum-likelihood inference analysis of a matrix of concatenated sequences from three nuclear gene regions (nrLSU/28S, rpb1 and rpb2). Zangia erythrocephala was used as outgroup taxon. Values for clades that are supported in either the Bayesian (posterior probabilities, BPP) and Maximum likelihood (ML bootstrap percentage, MLB) analyses are indicated. BPP ≥ 0.95 and MLB ≥ 70% are given above clade branches. Newly sequenced collections are boldfaced in black. For each collection, the specific epithet (as present in GenBank), the herbarium code and GenBank accession numbers of the nrLSU/rpb1/rpb2 sequences are reported.

Our datasets consist of sequences of Neoboletus and other sequences with greatest similarity available in GenBank selected based on BLASTN search and previous molecular studies including Neoboletus collections (Wu et al. 2014, 2016a, b; Smith et al. 2015; Urban and Klofac 2015).

Sequences were aligned with MAFFT v. 7.017 (Katoh et al. 2002) and then manually adjusted using Geneious v. R 11.1.4 (Kearse et al. 2012). Two phylogenetic analyses were performed: the first phylogenetic analysis, based on a combined nrLSU/rpb1/rpb2 dataset, was focused on the intergeneric position of the new species in the Pulveroboletus group of the Boletaceae, as delimited by Wu et al. (2014). According to the results by Wu et al. (2014), Zangia erythrocephala was chosen as outgroup taxon for the three-loci combined dataset. The second phylogenetic analysis based only on a nrITS sequence dataset was restricted to the taxa closely related to the new species (genus Neoboletus). Costatisporus cyanescens was used as outgroup taxon for this dataset following Smith et al. (2015).

The GTRGAMMA model of sequence evolution was selected for both analyses. The two phylogenetic analyses were inferred with three partitions: nrLSU(28S)/rpb1/rpb2 and ITS1/5.8S/ITS2, respectively. The datasets were analyzed using Bayesian inference (BI) and Maximum likelihood (ML) criteria. The BI was performed with MrBayes v.3.2 (Ronquist et al. 2012) with four incrementally heated simultaneous Monte Carlo Markov Chains (MCMC) run for 10 million generations, under the selected evolutionary model. Trees were sampled every 1000 generations, resulting in overall sampling of 10001 trees; the first 2500 trees were discarded as “burn-in” (25 %). For the remaining trees, a majority rule consensus tree showing all compatible partitions was computed to obtain estimates for Bayesian Posterior Probabilities (BPP). The ML was performed with RAxML v.7.2.8. (Stamatakis 2006) and a total of 1000 bootstrap replicates (Felsenstein 1985) were computed to assess the relative robustness of the branches. Only BPP values ≥ 0.95 and MLB (Maximum likelihood bootstrap) values ≥ 70 % have been reported in the phylogenetic trees (Figs 1, 2). Pairwise % identity values of the sequences were calculated using Geneious v. R 8.1.2 (Kearse et al. 2012). Alignments and phylogenetic trees are available at Tree-BASE (www.treebase.org, submission number S24011).

Figure 2. 

Bayesian phylogram obtained from the nrITS sequence alignment of Neoboletus species. Costatisporus cyanescens was used as outgroup taxon. Values for clades that are supported in either the Bayesian (posterior probabilities, BPP) and Maximum likelihood (ML bootstrap percentage, MLB) analyses are indicated. BPP ≥ 0.95 and MLB ≥ 70% are given above clade branches. Newly sequenced collections are boldfaced in blue. For each collection, the specific epithet (as present in GenBank), the herbarium code, GenBank accession number of the nrITS sequence and geographical origin (country) are reported.

The combined nrLSU/rpb1/rpb2 data matrix (focused on the Pulveroboletus group) comprised 47 sequences and is 2381 bp long. The nrITS data matrix (focused on Neoboletus) comprised 41 sequences and is 830 bp long. As the topology and branches support values of all the analyses are consistent, only the Bayesian trees with both BPP and MLB values are shown (Figs 1, 2). In the combined analysis (Fig. 1) a major clade is recognizable (BPP = 1, MLB = 82%), here named as the Sutorius clade, where the two sister (BPP = 1, MLB = 89%) genera Sutorius and Costatisporus are sister (BPP = 1, MLB = 82%) to the genus Neoboletus. The two collections of the new species clustered together within the genus Neoboletus forming a strongly supported clade (BPP = 1, MLB = 100%) which is sister (with no support) to N. magnificus. In the nrITS analysis (Fig. 2) the three collections of the new species (N. antillanus) clustered together in a strongly supported clade (BPP = 1, MLB = 100%) which shows no clear phylogenetic affinities with other species. The three nrITS sequences (677 to 683 bp) and the two nrLSU sequences (840 to 857 bp) of N. antillanus show a pairwise % identity value of 99.7 and 100, respectively. The type specimen of Boletus brunneopanoides from Belize forms: i) a strongly supported clade (BPP = 1, MLB = 89%) with also two collections of B. vermiculosoides, one collection of Boletus cf. vermiculosoides and one of B. vermiculosus, in the combined analysis; ii) a strongly supported clade (BPP = 1, MLB = 100%) with also two collections of B. vermiculosoides, one collection of Boletus cf. vermiculosoides and one of Boletales sp. (KY826093), in the nrITS analysis.

Phylogenetic analyses corroborate the proposal of the new species N. antillanus (Figs 1, 2). It forms an independent evolutive line within Neoboletus with no evident phylogenetic relationships (it is sister to the Chinese N. magnificus in the combined analysis, but without statistical support) to allied congeneric taxa. According to the same analyses, B. brunneopanoides, a Belizean red-pored bolete species phylogenetically nested in Neoboletus, clustered in the same clade with collections named B. vermiculosus/B. vermiculosoides from North America. Should future molecular work prove conspecificity among these three species, B. vermiculosus Peck would have priority.

The genus Neoboletus currently encompasses fewer than ten species geographically restricted to the northern hemisphere and essentially distributed in temperate and tropical regions. However, judging from morphological traits, there might be an additional number of species, up to three times as many in fact, belonging to the same genus, most of which have not yet been molecularly investigated. It is worth noting that a group of Chinese researchers after having firstly accepted Neoboletus as an independent genus (Wu et al. 2016a), have subsequently reduced it in synonymy with Sutorius Halling, Nuhn & Fechner based on a wider interpretation of the generic boundaries within the Boletaceae (Wu et al. 2016b). As previously pointed out by Gelardi (2018), we presently disagree with this broad circumscription of Sutorius since it is, judging from the original description, easily separated from Neoboletus based on the overall dark colors, different stipe ornamentation pattern, different spore print color, pores stuffed in early developmental stages like those of Boletus s. str. and Butyriboletus Arora & J.L. Frank and non-bluing tissues (Halling et al. 2012b). Enough, in our opinion, to state they are not the same thing especially because they cluster in two different (although with a low statistical support) sister clades. Moreover, molecular studies carried out by Smith et al. (2015) on false-truffle fungi from north-eastern South America (Guyana) and our nrLSU/rpb1/rpb2 analysis (Fig. 1) indicate the sequestrate genus Costatisporus T.W. Henkel & M.E. Smith as the sister taxon to Sutorius. Costatisporus, Neoboletus and Sutorius form the Sutorius clade (Fig. 1).

Neoboletus antillanus is easily identifiable among other species of the same genus based on the following set of unique morphologically informative features: 1) medium-small size, 2) reddish to pinkish red then pinkish cream pileus surface, 3) pores orange red to yellowish orange, 4) stipe ornamented over the lower three fourth by purple-red to reddish orange punctuations on a yellow background, 5) lowermost part of the stipe prominently strigose with golden yellow to brownish yellow hairs, 6) yellow context, 7) tissues bruising dark blue when injured, 8) ellipsoid-fusiform, smooth basidiospores, 9) ixocutis pileipellis consisting of gelatinized, repent filamentous hyphae and 10) occurrence in neotropical lowland mixed broadleaved forests. To date, N. antillanus has never been found with host species other than local autoctonous broadleaved trees and does not appear to be associated with either Pinaceae or Fagaceae (the latter plant family is not present in Dominican Republic). Moreover, such a purported ECM association of N. antillanus with the endemic C. uvifera might implicate a neotropical origin. Further suggestion supporting a symbiotic relationship between N. antillanus and C. uvifera is the co-occurrence at the same locality with Cantharellus coccolobae Buyck, P.-A. Moreau and Courtec., which is strictly associated with seagrape in tropical America (Buyck et al. 2016).

Among the other endemic red-pored boletes reported from Central America, Boletus pyrrhosceles Halling, B. guatemalensis R. Flores & Simonini, B. dupainii Boudier and B. paulae J. García, Singer & F. Garza-Ocañas superficially resembles N. antillanus. However, B. pyrrhosceles is easily separated by the reddish brown to brownish orange pileus surface, adnate to subdecurrent hymenophore, shallow tubes (up to 5 mm deep), brownish red pores, tomentose and reticulate stipe that is entirely brownish red to deep red, slightly smaller basidiospores (9.1–11.2 × 4.2–4.9 μm, Qm= 2.3), trichodermal pileipellis and association with Quercus humboldtii Bompl. in Colombia (Halling 1992). Boletus guatemalensis has a whitish to pale yellow context with yellowish green spots towards the stipe base, radially elongated angular pores, stipe surface with brownish green fibrils in the lower half and a smooth, whitish base, white basal mycelium, unchanging tissues, mostly 2- or 3-spored basidia, a cutis pileipellis consisting of non-gelatinized hyphae, hymenophoral trama intermediate between the “Boletus-type” and the “Phylloporus-type” and is associated with Pinus caribaea Morelet in Guatemala, Belize and Mexico (Flores Arzù and Simonini 2000, Ortiz-Santana et al. 2007, García-Jiménez et al. 2013). Boletus dupainii s. Ortiz-Santana et al. differs in the larger size (pileus up to 13 cm broad), polish and shiny, carmine red to crimson red pileus surface, deep red pores, stipe base devoid of strigosity, longer spores (12.8–14.4 × 4–5.6 μm, Qm= 2.9), smaller basidia (24–29.6 × 9.6–10.4 μm), shorter pleuro-, cheilo- and caulocystidia (26.4–47.2 × 7.2–8.8 μm, 16–30.4 × 4.8–8 μm and 16–36.8 × 5.6–11.2 μm, respectively), thinner pileipellis hyphae (up to 6.5 μm diam.) and growth in symbiosis with Quercus spp. in Belize (Ortiz-Santana et al. 2007). This species has recently been assigned to Rubroboletus Kuan Zhao & Zhu L. Yang on account of morphological and molecular evidence (Zhao et al. 2014). It was originally described from Europe (Boudier 1902) where it appears to be widespread although uncommon, but in recent times it has also been reported from the New World (McConnell and Both 2002, Ortiz-Santana et al. 2007, Both et al. 2009, García-Jiménez 2013, Bessette et al. 2016). However, the conspecificity of the European material with that from the western hemisphere is yet to be confirmed and a comparative analysis is currently under examination. Finally, B. paulae exhibits a deep red, vinaceous red to strawberry red pileus surface, smooth stipe base, whitish gray basal mycelium, pale whitish yellow and erratically bluing context on exposure, hymeniform pileipellis consisting of chains of inflated to subglobose elements up to 34 μm broad and ECM association with oaks in Mexico (García-Jiménez et al. 2013).

Although N. antillanus exhibits some superficial morphological affinities with Boletus vermiculosus Peck, B. vermiculosoides A.H. Smith & Thiers and B. brunneopanoides B. Ortiz, these three species have larger basidiome size (pileus 7–18 cm broad and stipe 9–14 cm long in B. vermiculosus, pileus up to 12 cm and 16 cm broad in B. vermiculosoides and B. brunneopanoides, respectively), subtomentose to velvety, yellowish brown or grayish brown to dark brown pileus surface, brownish orange to amber brown or dark brown pore surface fading brownish yellow with age, extremely fine brownish punctuations on stipe surface and stipe base without hairs. B. vermiculosus also differs from N. antillanus in the trichodermal pileipellis devoid of gelatinous matter, longer basidiospores [(11) 12.6–14 (15) × (4) 4.9–5.6 (6) μm, Qm= 2.6] and the occurrence under Fagaceae. B. vermiculosoides is further distinguished by the paler, whitish-yellow stipe surface, narrower basidiospores [9–12 × 3–3.5 (4) μm], smaller basidia (20–26 × 7–9 μm) and association with Fagaceae, whereas B. brunneopanoides is also separated by the whitish stipe surface, narrower basidiospores (8.8–12.8 × 4 μm), smaller basidia (20.4–32× 8–8.8 μm) and the occurrence with Pinaceae (P. caribaea) (Coker and Beers 1943, Smith and Thiers 1971, Both 1993, Bessette et al. 2000, 2016; Halling and Mueller 2005, Ortiz-Santana et al. 2007). Boletus vermiculosus and B. vermiculosoides were originally described from eastern North America but the former is also encountered in Central America south to Belize and Costa Rica (Bessette et al. 2000, 2016; Halling and Mueller 2005), while B. brunneopanoides was only found in Belize (Ortiz-Santana et al. 2007). Up to now, neither of these species has been reported from the Dominican Republic.

At least two additional North American boletes might be confused with N. antillanus, namely Boletus subluridus (Murrill) Murrill and B. fairchildianus (Singer) Singer. The combination of yellowish orange, orange-pink to purplish red pileus surface, dark red pores, non-strigose stipe base, slightly longer basidiospores [(8.5) 9–14(14.5) × (3.5) 4–6(7) μm], smaller basidia (20–25.5 × 7.5–10 μm and occurrence with oaks and pines in south-eastern USA differentiate B. subluridus from N. antillanus (Murrill 1938, Singer 1945, 1947, both as B. miniato-olivaceus var. subluridus Singer; Both 1993, Bessette et al. 2000; 2016), whereas B. fairchildianus is distinguished by the larger size (pileus up to 15 cm broad), stipe base without strigosity, larger basidiospores [(12.5) 13–18.8 (19.7) × (4.5) 5–8 μm] and the association with Quercus spp. in south-eastern USA and Mexico (Singer 1945, 1947, both as B. rubricitrinus var. fairchildianus Singer; Both 1993, Bessette et al. 2000, 2016; García-Jiménez 2013).

Neoboletus luridiformis (Rostk.) Gelardi, Simonini & Vizzini (= Boletus erythropus Pers. s. Fr. et auct. p.p. non s. Pers.) differs significantly from N. antillanus in the large sized basidiomes (pileus up to 25–30 cm in diam.), dark chocolate brown to umber brown, velvety pileus surface, bright red pores, stout, fleshy stipe (up to 15 × 8 cm), non-strigose stipe base, longer basidiospores [(12.8) 13.3–15.5 (16.5) × 4.2–5.5 μm, Qm= 2.95], trichodermal pileipellis with interwoven erect, non-gelatinous hyphae and occurrence in Europe in temperate regions (Pilát and Dermek 1974, Alessio 1985, Breitenbach and Kranzlin 1991, Lannoy and Estadès 2001, Muñoz 2005, Watling and Hills 2005, Klofac 2007, Knudsen and Taylor 2012; pers. obs.).

The eastern Asian species N. brunneissimus (W.F. Chiu) Gelardi, Simonini & Vizzini and N. antillanus share some common features such as basidiome size, presence of golden yellow to brownish yellow strigosity at the stipe base, yellowish context and dark blue staining of tissues by auto-oxidation but the former is readily separated by the velvety and rusty brown to umber-brown pileus cuticle, rusty brown to reddish-brown pores, denser and rusty-brown punctuation on stipe surface, trichoderm pileipellis consisting of non-gelatinized erect hyphae with slightly shorter and narrower terminal elements [23–45 (58) × 3.5–5 (7) μm] and the occurrence in East Asia in association with Fagaceae and Pinaceae (Chiu 1948, 1957; Bi et al. 1997, Mao 2000, 2009; Wang and Liu 2002, Wang 2004, Wang et al. 2004; Zang 2006, Wu et al. 2016a, Gelardi 2018).

The Chinese N. magnificus (W.F. Chiu) Gelardi, Simonini & Vizzini, Sutorius sanguineoides G. Wu & Zhu L. Yang and S. sanguineus G. Wu & Zhu L. Yang are three additional eastern Asian species that may be confused with N. antillanus. Aside from the different geographical distribution and the ECM deciduous and coniferous host associates (Fagaceae and Pinaceae), the former species is also delimited by the dark red to reddish brown pores in the early developmental stages, a decidedly clavate to bulbous stipe base (up to 6 cm broad) that is devoid of or sometimes with inconspicuous strigosity and non-gelatinized trichodermal pileipellis with broader end elements (up to 16 μm wide) (Chiu 1948, 1957; Bi et al. 1997, Mao 2000, 2009; Wang 2004, Wang et al. 2004, Zang 2006, Wu et al. 2016a), whereas S. sanguineoides and S. sanguineus are both separated from N. antillanus on account of the deep red, blood red to brownish red pileus surface, dark red to brownish red pores, non-strigose stipe base, non-gelatinized trichodermal pileipellis and the occurrence in subalpine forests at very high elevations (over 3000 m alt.) (Wu et al. 2016b). Furthermore, S. sanguineoides differs in its decidedly larger basidiospores [13.5–17 (21) × 5–7 μm, Qm= 2.56] while S. sanguineus also exhibits an evenly red stipe surface, slightly longer basidiospores [10–14 (15) × 5–6 (7) μm, Qm= 2.14] and broader cystidioid pileipellis terminal cells (9–15 μm wide) (Wu et al. 2016b).