Introduction

In the early 1970s, three large wooden river barges and three smaller Roman-era vessels were discovered during construction work in the village of Zwammerdam in the Netherlands. At the time, the care facility ‘Ipse de Bruggen’ was being built in Zwammerdam on the Hooge Burch estate and by chance the workers came across the Zwammerdam vessels. The vessels were all made of oak and the river barges were exceptionally large, ranging from 20 to 34 m in length (De Weerd 1988).

In the years leading up to the discovery, the University of Amsterdam conducted archaeological research in the area and investigated a Roman fort that was marked on a Roman road map as the military point Nigrum Pullum (De Weerd 1988). This fort was part of the Roman Limes, a border where a series of forts along the river Rhine formed a defensive line in the Lower Germanic region (Fig. 1A). This defence line also functioned as a transport corridor, a strictly controlled long-distance waterway from the interior regions inland towards the coast.

Fig. 1
figure 1

A The Roman Limes with the fort of Nigrum Pullum situated in today’s Zwammerdam, the Netherlands (drawing by T. Hazenberg). B The position of the vessels of Zwammerdam, the Netherlands, in front of the fort of Nigrum Pullum. C The Zwammerdam 6 ship during the excavation (photo by G. Verkuyl, collection ACASA, University of Amsterdam, the Netherlands). D Fungal growth of Microascus melanosporus (dominant) and Acremonium species on wood from an L-shaped chine girder. E. White patches with Blastobotrys nigripullensis and Microascus paisii on wood from a portside wale

Full size image

The vessels found near the site of Nigrum Pullum were scattered along a section of timber shoring in front of the fort, where the Rhine once flowed, and were located at various depths (Fig. 1B). They were essentially found in the riparian zone of a silted up river bed. To document and salvage the ships, construction work had to be halted and, under great time pressure, only the ships were properly examined. At that time, the relationship to the fort could not be further explored (De Weerd 1988). Why the ships sank or were scuttled in the vicinity of the fort thus remains unknown.

An early form of crowdfunding, supported by former Queen Juliana of the Netherlands, allowed the vessels to be salvaged and conserved to enable future display in a museum. The conservation of the wood with polyethylene glycol (PEG, 4000 g/mol), a water-soluble waxy solid, was carried out in batches starting in 1976 and took an average of two years per load of timbers. When heated, high-molecular weight PEG can penetrate wood tissues and replace water molecules. The PEG also bulks the deteriorated wood cells and thus reinforces the structure of the wood, improving dimensional stability. After PEG treatment, the wood was stored stacked in shipping containers. Decades later in 2017, a project started, organized by Hazenberg Archeologie and Museumpark Archeon in the Netherlands, that enabled the restoration of the vessels.

Unfortunately, an undetected leak in one of the shipping containers with remains of the sixth ship of Zwammerdam (Fig. 1C) had resulted in the preservative being washed out of some of the timbers. These timbers had started to dry out, crack and warp in some areas and had developed visible fungal growth (Fig. 1D, E). To seek advice for the removal of this growth, the project staff contacted the Westerdijk Fungal Biodiversity Institute (WI) in Utrecht and send in some samples to investigate further.

During the study of the fungal diversity present, a putative new Blastobotrys species was isolated. In the dual nomenclature era, Blastobotrys was considered the asexual genus (or anamorph) associated with the sexual genus (or teleomorph) Trichomonascus. This relationship was confirmed by Smith et al. (2011a, b). Following the abandonment of dual nomenclature for fungi (Hawksworth et al. 2011; McNeill et al. 2012), a decision on what generic name to apply to the Blastobotrys/Trichomonascus lineage remains to be made. This was discussed in a recent revision published for the lineage by Visagie et al. (2023) in which B. davinci was described, a species that has been detected from house dust and historically important artifacts including the self‐portrait of Leonardo da Vinci which dates from ca 1512. The goal of the current study is to report on the fungi isolated from Zwammerdam vessel wood samples, to compare the putative new species with known Blastobotrys and Trichomonascus species and to formally introduce B. nigripullensis sp. nov. For this, we provide a phylogenetic analysis based on the internal transcribed spacer (ITS1-5.8S-ITS2) rDNA region (ITS) and D1/D2 regions of the large subunit (LSU), and describe the new species based on morphological and physiological characters.

Materials and methods

Sampling and isolations

A large oak knee detached from a floor timber with fungal growth as well as small pieces of a few centimetres of oak from hull planks with small white patches of suspected fungal growth on the surface were sent from Archeon (Alphen aan den Rijn, the Netherlands) to the Westerdijk Fungal Biodiversity Institute (WI, Utrecht, the Netherlands). Here, the suspected fungal material was plated directly on dichloran 18% glycerol agar (DG18), malt yeast 50% glucose agar (MY50G) and malt extract agar (MEA, Oxoid) supplemented with the antibiotics penicillin (6600 U per plate) and streptomycin (20,000 U per plate) to inhibit bacterial growth. The plates were incubated for 10–14 days at 25 °C in the dark. After incubation, colonies were transferred to fresh agar media. The strains were preserved in 30% glycerol, stored at − 80 °C and accessioned in the working culture collection of the Food and Indoor Mycology research group (DTO) housed at WI. Some of these strains were also preserved in the CBS culture collection and fungarium of WI.

DNA extraction, sequencing and phylogenetic analysis

DNA was extracted from 7 d old colonies grown on MEA using the UltracleanTM Microbial DNA isolation Kit (MoBio Laboratories Inc., Solana Beach, USA). The ITS was amplified for all strains using the primer pair V9G and LS266 and used to make initial identifications to genus level. For species level identifications, additional regions were amplified as needed. These included LSU, translation elongation factor 1-α (TEF) and beta-tubulin (BenA) using primer pairs LR5 and LROR, EF1 and EF2 and Bt2a and Bt2b, respectively. All PCR amplifications were made following Samson et al. (2019). Amplification profiles were as follow: initial denaturing 94 °C for 5 min, 36 cycles of 94 °C for 45 s, 55 °C for 45 s, 72 °C for 1 min, followed by a final elongation step of 72 °C for 10 min. Sequencing reactions were prepared using the BigDye Terminator Cycle Premix Kit (Applied Biosystems, Waltham, USA) and were run on an ABI PRISM 3730xl genetic analyser with the same primer pairs used for PCR amplification. Sequence contigs were assembled in SeqMan Pro v.15.0.0 (DNASTAR) and newly generated sequences deposited in GenBank (https://www.ncbi.nlm.nih.gov/genbank/) (accessions OQ876811–OQ876820, OQ876860, OQ876861 and OQ918213). Initial identifications were made by performing similarity searches using BLAST against the nucleotide sequence database. To compare the putative new Blastobotrys strains with known Blastobotrys/Trichomonascus species, the same DNA sequence reference dataset (Table 1) and phylogenetic approaches published by Visagie et al. (2023) were used. Alignments and treefiles were uploaded to the University of Pretoria’s research data repository hosted on Figshare (https://doi.org/https://doi.org/10.25403/UPresearchdata.23815032).

Table 1 Strains used for phylogenetic analyses
Full size table

Morphological and physiological characterization of the yeast isolates

Two strains (CBS 17879 and CBS 17880) were morphologically and physiologically characterised as described by Visagie et al. (2023) using MEA, yeast morphology agar (YMA), potato dextrose agar (PDA) and glucose yeast peptone agar (GYPA), 5% glucose in yeast nitrogen broth and Dalmau plates on yeast morphology agar. The physiological growth profile of the strains was assessed using methods commonly used in yeast taxonomy (Kurtzman et al. 2011).

Results

Isolations and identifications

Three wood samples with visible fungal growth were studied and isolations made from these resulted in five strains representing four species. Light brown to beige fungal spots were present on the wood from an L-shaped chine girder sample (a combined bottom/side timber in the hull of a Roman river barge) on port side (timber code ZW6-K2 bb), and isolations revealed the presence of Microascus melanosporus (DTO 457-E5: ITS = OQ876812, LSU = OQ876817, BenA = OQ876861) and a member of the Acremonium chrysogenum-clade (DTO 457-E6: ITS = OQ876813, LSU = OQ876818). The other two samples (knee of a floor timber [timber code ZW6-S18] and part of a wale on port side [timber code ZW6-BE2 bb]) showed white spots (Fig. 1E) and direct microscopic analysis of these spots showed (pseudo)hyphae and yeast cells. Initial isolations only showed growth of yeast-like colonies on the MY50G plates; however, subculturing on MEA/PS and DG18 showed that these colonies could also grow on these media. A Blastobotrys species (CBS17879 = DTO 457-E3: ITS = OQ876814, LSU = OQ876819; CBS17880 = DTO 455-G9: ITS = OQ876815, LSU = OQ876820) was growing on the two investigated wood samples, and the new species B. nigripullensis is introduced below. Besides B. nigripullensis, also Microascus paisii (DTO 457-E4: ITS = OQ876811, LSU = OQ876816, BenA = OQ876860, TEF = OQ918213) was isolated from wood sample ZW6-BE2 bb, but no structures of this species were observed during the direct examination of the material.

Microascus species are both widespread and prefer substrates with cellulose, while Acremonium species occurs in soils and on decaying plant matter (Barron et al. 1961, Domsch et al. 1993; Hou et al. 2023). Considering the context of archaeological wood taken from the ground and then decomposed in parts, the occurrences of these species fit the picture. The surfaces of samples ZW6-K2 bb and ZW6-BE2 bb were steam cleaned and seven months later no regrowth was evident. However, the knee of the floor timber (ZW6-S18) was steam cleaned twice after Blastrobotrys reappeared after the first cleaning attempt, but seven months later the fungus again formed visible growth on the wood surface. Next, a treatment with 70% alcohol is planned in the hope that this will be sufficient in the long term.

Species delineation and phylogenetic placement of B. nigripullensis sp. nov

Phylogenetic analyses based on ITS and LSU included all species currently accepted in Blastobotrys/Trichomonascus (Visagie et al. 2023). The ITS and LSU alignments were 669 and 576 bp long, respectively, with TVM + I + G being the best model for the ITS partition and GTR + G the best for the LSU partition. Blastobotrys nigripullensis resolved on a fully supported branch distinct of all other species (Fig. 2). The tree had a general lack of backbone support, and its closest relative(s) could not be determined as a result. Two DNA reads in Genbank, namely clones CMH114 (KF800205) and S23T-59 (KU164259), represent the same species. KF800205 originated from house dust and/or indoor air collected from Kansas City (USA) (Rittenour et al. 2014), and KU164259 originated from sea water collected from an unknown locality. Similarities between B. nigripullensis (CBS 17879 T) and other species were low with LSU that differed by 42nt or more (based on alignment from 5’-ATTGCCCTAGTAACGGCGAGCG-3’) and ITS by 130nt or more (based on alignment from 5’-AAGGATCATTACCGAA-3’) nucleotide substitutions, thus warranting recognition as a distinct species.

Fig. 2
figure 2

Combined phylogenetic tree of described Blastobotrys and Trichomonascus species based on ITS and LSU, indicating the phylogenetic placement of Blastobotrys nigropullensis. The tree is rooted to Wickerhamiella domercqiae CBS 3451 T. Posterior probabilities (≥ 0.95) and Bootstrap support values (≥ 80%) are given above branches. The new species is indicated in bold text and grey box, T = ex‐type strain, c = sequence from cloned DNA fragment

Full size image

The new species lacks the typical morphology of most Blastobotrys as it did not form the branched conidiophores ending in conidiogenous cells producing blastoconidia as seen in e.g., B. davincii and other Blastobotrys (and those formerly classified in Trichomonascus) species (Smith et al. 2011a, b, Visagie et al. 2023). It does, however, resemble conidiophores produced by most Sporothrix species (Fig. 3). The nutritional growth requirements of strains CBS 17879 and CBS 17880 are presented in Table 2 and can be summarized as follows: the strains ferment glucose, but no other sugars tested; they utilize many of the tested carbon sources, but no maltose, and do not utilize nitrate and nitrite; the species is mesophilic and able to grow from 18 to 30 °C, but not at 35 °C; no growth in 0.1% cycloheximide, no production of starch or starch-like compounds; growth in substrates with elevated sugar (50%) and NaCl (36%) concentrations. Salient growth characteristics of the new species, and species that belong to the two most closely related clades are given in Supplemental Table 1. Like B. baotianmenensis the species does not utilize maltose, but can be separated from this species by growth on L-sorbose, D-xylose, L-arabinose, D-ribose, ethanol, galactitol, mannitol, glucitol, and myo-inositol.

Fig. 3
figure 3

Blastobotrys nigripullensis (CBS 17879 T). A Colonies on, from left to right, YMA and GYPA. Colony close‐ups on MYA B, PDA C, MEA D and YMA E. F–O. Conidiophores and conidia from YMA observed under light microscope. Scale bars 10 µm

Full size image
Table 2 Growth characteristics of two isolates of Blastobotrys nigripullensis
Full size table

Here we adopt a phylogenetic species concept to delineate our new species and formally describe it below as B. nigripullensis.

Taxonomy

Blastobotrys nigripullensis Boekhout, M. Groenew., Meijer, and Visagie, sp. nov.—MycoBank MB 847655; Fig. 3.

Etymology: Latin, nigripullensis, after Nigrum Pullum, the Roman fort located at Zwammerdam, province Zuid-Holland, the Netherlands, along the Rhine River [now Old Rhine River].

Type specimen: The Netherlands, Zwammerdam, isolated from wood of Roman shipwreck, 15 September 2021, collected by Y. Vorst, (holotype CBS H-25194 (preserved in a metabolically inactive state in the fungarium), culture ex-type CBS 17879 = DTO 457-E3).

DNA barcodes: ITS = OQ876815, LSU = OQ876820.

UNITE species hypothesis: https://doi.org/10.15156/BIO/SH1135984.09FU.

Description: Sporothrix-like. After two weeks at 25 °C in 5% glucose in yeast nitrogen broth, white floating flocks and white sediment occur with short hyphal cells, 6–16 × 2.5–3.5 μm, with sympodial and lateral budding. After two weeks at 25 °C on yeast morphology agar, colonies ca. 7 mm wide, flat, dry, and tough, dull white to cream, under lens surface venose, with entire margin. Hyphae branched, hyaline, 1.5–3 μm wide, with lateral, clavate blastoconida, 3–7 × 1–3.5 μm, on short to longer denticles forming a sympodial rachis of blastoconidia; thicker, inflated regions, 7–130 × 7–20 μm, may be present with finger-like outgrowths. Dalmau plate on yeast morphology agar with short adhering filaments. Similar growth on GYPA, YMA, PDA and MEA. Fermentation, assimilation, results from other growth and biochemical tests, and growth at different temperatures are listed in Table 2.

Additional strains examined: The Netherlands, Zwammerdam, isolated from wood of Roman shipwreck, 15 September 2021, collected by Y. Vorst, CBS 17880 = DTO 455-G9.

Discussion

The morphology of B. nigripullensis resembles that of the genus Sporothrix, more than that of Blastobotrys/Trichomonascus (Smith et al. 2011a, b, Visagie et al. 2023) as the typical inflated conidiogenous cells of the latter are absent. The molecular phylogeny, however, clearly shows that B. nigripullensis belongs to the core clade of Blastobotrys/Trichomonascus that include the generic type species, Blastobotrys nivea.

Blastobotrys species are common and previously found on straw, rotting/dead wood, mushrooms, nuts and black pepper; in soil compost, the indoor environment (dust and air), the Antarctic Ocean and open waters. Species of the genus are usually osmotolerant, xerotolerant and thermotolerant, and some are applied in biotechnology (Visagie et al. 2023). Blastrobotrys nigripullensis was isolated from wood of a Roman shipwreck that was preserved in alluvial clay deposits near the former Roman fort of Nigrum Pullum, located in the present-day village of Zwammerdam, province Zuid-Holland, the Netherlands. The oak of the shipwreck was preserved with PEG 4000, but in this case, due to an undetected leak, the preservative washed out and, consequently, the wood slowly dried. This resulted in some visual growth of fungi that may be tolerant to high PEG concentrations. Two DNA clones that originated from house dust in the USA and seawater, had the same ITS sequences and might indicate the presence of B. nigripullensis in those samples as well. It indicates a rather broad ecological and, likely, also a broad geographical preference. The new Blastobotrys species was isolated together with Microascus melanosporus (predominant), Microascus paisii and a member of the Acremonium chrysogenum-clade. Both Microascus species are widespread and prefer substrates with cellulose, while Acremonium species occurs in soils and on decaying plant matter (Samson et al. 2019). It will therefore be interesting to investigate further whether the novel species is also able to assimilate substrates such as cellulose. Preliminary experiments did not observe growth of B. nigripullensis strains on agar media with lignin and ferulic acid (R.P. de Vries, unpubl. observations).

Blastobotrys davincii was recently described from house dust collected in Canada and has previously been detected in environmental DNA from several historically important artifacts, including a mummy (Palermo, Italy), Leonardo da Vinci’s self-portrait and cave paintings (Maijishan Grottoes, China) (Visagie et al. 2023). Together with B. nigripullensis isolated from a Roman shipwreck, albeit treated with PEG, it seems that Blastobotrys species can occur on historical artefacts and may be more common than previously thought when dealing with such substrates. The discovery of B. nigripullensis raised the question of whether the yeast was already present when the wood was processed and is therefore a species from Roman times or whether it arrived via infiltration of groundwater into the archaeological site or after excavation, entering the conservation process. Since it is a species of a genus with representatives that has also been found on other historical and ancient objects, might a contemporary origin not be as likely as later contamination? Could B. nigripullensis have flourished again under the right circumstances, in this case pieces of PEG-washed out degrading wood? More research is needed to answer this question. The initial isolation of B. nigripullensis from the wooden remains was done on MY50G, an agar medium containing 50% glucose, which is only used to isolate extremely xerophilic fungi. However, this medium is not routinely used by most laboratories dealing with fungi on archaeological remains. Extreme xerophiles, such as Aspergillus halophilicus that grows on climate-controlled stored heritage artefacts, can only be detected and grown on media with low water activity (e.g. MY50G) (Bastholm et al. 2022), while B. nigripullensis can also grow on media with higher water activity. The species was able to grow on agar media with up to 36% NaCl, indicating salt tolerance. Further investigations are needed to understand why this yeast grew on PEG-contaminated remains of a Roman shipwreck.