Unravelling unexplored diversity of cercosporoid fungi (Mycosphaerellaceae, Mycosphaerellales, Ascomycota) in tropical Africa

Abstract Cercosporoid fungi (Mycosphaerellaceae, Mycosphaerellales, Ascomycota) are one of the largest and most diverse groups of hyphomycetes causing a wide range of diseases of economically important plants as well as of plants in the wild. Although more than 6000 species are known for this group, the documentation of this fungal group is far from complete. Especially in the tropics, the diversity of cercosporoid fungi is poorly known. The present study aims to identify and characterise cercosporoid fungi collected on host plants belonging to Fabaceae in Benin, West Africa. Information on their morphology, host species and DNA sequence data (18S rDNA, 28S rDNA, ITS and tef1) is provided. DNA sequence data were obtained by a simple and non-culture-based method for DNA isolation which has been applied for cercosporoid fungi for the first time in the context of the present study. Among the loci used for the phylogenetic analysis, tef1 provided the best resolution together with the multigene dataset. Species delimitation in many cases, however, was only possible by combining molecular sequence data with morphological characteristics. Based on forty specimens recently collected in Benin, 18 species are presented with morphological descriptions, illustrations and sequence data. Among these, six species in the genus Cercospora and two species in Pseudocercospora are proposed as species new to science. The newly described species are Cercospora (C.) beninensis on Crotalaria macrocalyx, C. parakouensis on Desmodium tortuosum, C. rhynchophora on Vigna unguiculata, C. vignae-subterraneae on Vigna subterranea, C. tentaculifera on Vigna unguiculata, C. zorniicola on Zornia glochidiata, Pseudocercospora sennicola on Senna occidentalis and Pseudocercospora tabei on Vigna unguiculata. Eight species of cercosporoid fungi are reported for Benin for the first time, three of them, namely C. cf. canscorina, C. cf. fagopyri and C. phaseoli-lunati are new for West Africa. The presence of two species of cercosporoid fungi on Fabaceae previously reported from Benin, namely Nothopassalora personata and Passalora arachidicola, is confirmed.

Cercosporoid fungi belonging to Mycosphaerellaceae (Mycosphaerellales, Ascomycota) are one of the largest and most diverse groups of hyphomycetes and cause a wide range of diseases, on numerous economically important plants such as cereals, vegetables and fruits as well as on wild plants. Major Videira et al. 2017). Infections by these fungi are mostly evident by leaf spots, but cercosporoid fungi can also cause necrotic lesions on flowers, fruits, seeds and pedicels of numerous hosts in most climatic regions (Agrios 2005). Cercosporoid fungi are known from all parts of the world but they are more abundant and diverse in tropical and subtropical regions (Beilharz et al. 2002;Braun and Freire 2004;Dianese 2008, 2009).
Cercosporoid fungi are dematiaceous hyphomycetes with conidiophores formed singly or in groups, arranged in sporodochia or in synnemata, with integrated, terminal or intercalary conidiogenous cells Ávila et al. 2005;Braun et al. 2013). Most of the cercosporoid species were previously assigned to a single genus, Cercospora, which was later split into several smaller genera mainly by Deighton (1967Deighton ( , 1973Deighton ( , 1974Deighton ( , 1976Deighton ( , 1979, Braun (1993) and Crous and Braun (2003). Crous and Braun (2003) recognized four genera, namely Cercospora, Passalora, Pseudocercospora and Stenella Syd as important cercosporoid genera. Later, the genus Stenella was assigned to the Teratosphaeriaceae based on the phylogenetic placement of the type species. Stenella-like species remaining in Mycosphaerellaceae were classified in the genus Zasmidium Fr. (Arzanlou et al. 2007;Braun et al. 2013). In the present paper, we follow generic concepts defined by Crous and Braun (2003) and recently updated by Braun et al. (2013), Crous et al. (2013a), Groenewald et al. (2013) and Videira et al. (2017). However, according to recent molecular sequence analyses, most genera of the cercosporoid fungi are not monophyletic (Videira et al. 2017). As many cercosporoid fungi have a strong impact on cultivated plants, a better understanding and stabilisation of the taxonomy of these fungi are urgently needed.
The genus Cercospora was established by Fresenius in 1863 (Fuckel 1863) based on the type species Cercospora apii Videira et al. 2017). It is one of the most species-rich genera of the hyphomycetes and contains numerous important plant pathogenic fungi throughout the world . In 1954, the genus was monographed by Chupp (1954), who treated 1419 Cercospora-species using a broad generic concept. Later, several attempts have been made to split Cercospora s. lat. into smaller genera by using characteristics of conidiomatal structure, hyphae, conidiophores, conidiogenous cells, conidiogenous loci and conidia (Ellis 1971(Ellis , 1976Deighton 1973Deighton , 1979Deighton , 1983Braun 1995aBraun , 1998Crous and Braun 2003). Currently, Cercospora species are morphologically characterised by pigmented conidiophores, unpigmented conidia, as well as thickened and darkened conidiogenous loci and conidial hila Groenewald et al. 2013). A significant problem in the taxonomy of Cercospora is the host specificity of its species. Most Cercospora species are considered to be distinct based on the host and thus assumed to be specific to a host species or to a host genus (Chupp 1954;Braun 1995a). Some species, such as C. apii and C. beticola, however, were isolated from a high number of host species belonging to several families (Groenewald M et al. 2006). Moreover, phylogenetic approaches based on multi-locus sequences can be problematic for species delimitation in Cercospora due to a high level of conservation in DNA sequences of commonly used loci (i.e., ITS, tef1, actA, cmdA and his3) (Bakhshi et al. 2018).
The genus Pseudocercospora was introduced by Spegazzini (1910) based on the type species Ps. vitis (Lév.) Speg., a foliar pathogen of grapevine. The majority of Pseudocercospora species are known as pathogens occurring on many different plants, mainly in tropical and sub-tropical regions (Chupp 1954;Crous and Braun 2003;. In contrast to Cercospora spp., they are characterised by pigmented conidiophores and conidia, without thickened and darkened conidiogenous loci and conidial hila (Deighton 1976). The monophyly of the genus has not yet been fully resolved (Kirschner 2014). According to molecular sequence data, most species of Pseudocercospora appear to be host specific   (Videira et al. 2017). Species of Passalora are characterised by pigmented conidiophores and conidia as well as thickened and darkened conidiogenous loci and conidial hila .
Several molecular phylogenetic studies are available on species of cercosporoid fungi that are represented by strains in culture collections (Świderska-Burek et al. 2020). These, however, only represent a small fraction of several hundreds of taxa of cercosporoid fungi that are valid species defined by morphological characteristics Świderska-Burek et al. 2020). Therefore, the number of cercosporoid species known by detailed morphological characteristics as well as molecular sequence data has to be increased.
Although cercosporoid fungi cause a wide range of diseases on major agricultural crops, the study of cercosporoid fungi in West Africa is still at an early pioneer stage and only very incomplete information is currently available (Piepenbring et al. 2020). To date, approximately 320 species of cercosporoid hyphomycetes are known from 14 West African countries (Piepenbring et al. 2020, Suppl. materials 1, 2). Among these, 12 species of cercosporoid fungi have been reported for Benin (Turner 1971;Marley et al. 2002;Crous and Braun 2003;Houessou et al. 2011;Piątek and Yorou 2018;Soura et al. 2018;Meswaet et al. 2019;Farr and Rossman 2021). Morphological characteristics and molecular sequence data are lacking for most cercosporoid species known for Benin and other West African countries. Although cercosporoid fungi have been investigated for more than 150 years and are important in the agricultural sector, almost no, or only inadequate, studies have been carried out in most West African countries such as Benin. In addition to this lack of species knowledge in tropical regions, many species of cercosporoid fungi are characterised morphologically only. Since many cercosporoid species are known as pathogens on cultivated plants, an accurate diagnosis, identification and documentation of these fungi are a prerequisite and urgent for their control and epidemiological surveys.
As a first step towards a systematic documentation of cercosporoid fungi in tropical Africa, we focus on species infecting hosts belonging to the Fabaceae (Leguminosae) in the present publication. Fabaceae are the third largest family of angiosperms (Gepts et al. 2005). This family includes peas, lentils, beans, peanuts and other plants with pods and/or seeds that are consumed as food (Messina 1999). Several species belonging to Vigna originate from West Africa (Benin, Burkina Faso, Cameroon, Ghana, Niger, Nigeria and Togo) including two important cultivated crops Vigna unguiculata (L.) Walp. and Vigna subterranea (L.) Verdc. (Hepper 1963;Faris 1965;Padulosi and Ng 1990). They provide important nutrients such as proteins, low glycemic index carbohydrates, minerals and vitamins. Legumes are richer in protein than other cultivated plants because of nitrogen-fixing bacteria living in nodules of their roots (Kouris-Blazos and Belski 2016).
We apply an integrative approach that includes sampling in Benin, detailed descriptions and illustrations of collected specimens and herbarium specimens, examination of closely related known species on the same or closely related host species based on herbarium specimens and the isolation, sequencing and analysis of nuclear DNA sequence data. For the isolation of DNA, a new, simple method for DNA isolation has been developed and is presented for the first time for cercosporoid fungi.

Collections and morphological studies
Samples of leaves infected by cercosporoid fungi were randomly collected in farmlands and fallows in Benin from July-August 2016, July-September 2017 and August-September 2019. Infected leaves were dried in a plant press and deposited in the herbaria Botanische Staatssammlung München (M) and University of Parakou (UNIPAR).
Dried specimens were observed by stereomicroscopy and by light microscopy, using a Zeiss Axioscope 40 microscope. For light microscopy, leaf sections were made with razor blades and mounted in distilled water or 5% KOH without staining. Semipermanent preparations of sections of the infected leaves were made by a microtome (Leica CM 1510-1) and mounted in lactophenol with cotton blue. For approximately 50 ml lactophenol cotton blue solution we mixed 10 mg phenol, 0.025 mg cotton blue, 10 ml lactic acid, 20 ml glycerin and 10 ml distilled water. Measurements of 30 conidia, conidiophores and other structures have been made for each specimen at a magnification of ×1000. Measurements are presented as mean value ± standard deviation with extreme values in parentheses. Line drawings were made freehand on scaled paper. Images and drawings were edited with Photoshop CS5 (Adobe, San Jose, California). Critical taxa were determined with the help of type specimens and other specimens loaned from the US National Fungus Collections (BPI), the Herbarium of the University of Illinois (ILL) and the New York Botanical Garden (NY).

Host plant identification
Host plants were identified by morphological characteristics and in some cases by molecular methods. Morphological identifications were made by comparison with herbarium specimens, literature (e.g., Akoégninou et al. 2006) and with the help of local botanists. Molecular sequence data for species identifications were obtained by polymerase chain reaction (PCR) for the amplification of the partial region of chloroplast rbcL with the primer pairs rbcLa-F (Levin et al. 2003) and rbcLa-R (Kress et al. 2009). DNA was extracted from approx. 0.05 g of leaf tissue dried with silica gel using the innuPREP Plant DNA Kit (Analytik Jena, Germany) and following the manufacturer's instructions. Protocols for PCR were carried out as described by Fazekas et al. (2012).

DNA Extraction and PCR amplification of fungal DNA
DNA was isolated from caespituli taken with a needle from dry specimens using the E.Z.N.A Forensic DNA Extraction Kit following the manufacturer's instructions. Small pieces of leaves containing several clean caespituli, with as little contaminations as possible, were selected under the stereomicroscope. Precautions were taken to avoid picking cells of any other organism (fungi, algae) associated with the leaves. To extract total genomic DNA from caespituli, a small amount of clean hyphae from the leaf sur-face was transferred into a sterile Eppendorf tube using a sterilized needle or adhesive mini-tapes. The sample was homogenized for 7-10 min. using a Retsch Mixer Mill MM301 with TL buffer and 2.5 mm Zirconia beads. Isolated DNA was re-suspended in elution buffer and stored at -20 °C. DNA concentration was checked by a Nan-oDrop 2000c spectrophotometer (Thermo Fisher Scientific, USA).
Four partial nuclear gene regions (three ribosomal loci and one protein-coding gene) were amplified and sequenced: For the large subunit nuclear ribosomal DNA (nrLSU, 28S rDNA) the primers LSU1Fd and LSU3Rd (Crous et al. 2009a), for the small subunit nuclear ribosomal DNA (nrSSU, 18S rDNA) the primers SSU1Fd and SSU1Fd (Crous et al. 2009a), for the internal transcribed spacer region of ribosomal DNA (ITS) the primers V9G (de Hoog and van den Ende 1998) and ITS4 (White et al. 1990) and for the translation elongation factor 1-α (tef1) the primers EF1-728F and EF1-986R (Carbone and Kohn 1999) were used. PCR amplification and sequencing were conducted following the protocols of Hunter et al. (2006), Crous et al. (2009aCrous et al. ( , 2012 and Videira et al. (2017). The PCR mixtures consisted of 1 μL genomic DNA, 15× MgCl 2 reaction buffer (Bioline, Luckenwalde, Germany), 25 mM MgCl 2 , 25 μM of each dNTP, 10 μM of each primer and 5 U Taq DNA polymerase (VWR) in a total volume of 25 μL. Cycling parameters of the PCR for LSU, SSU and ITS were as follows: initial denaturation at 94 °C for 3 min, followed by 35 cycles of amplification [denaturation at 94 °C for 30 s, primer annealing at 52 °C for 30 s and primer extension at 72 °C for 45 s] and a final extension at 72 °C for 5 min, followed by storage at 8 °C. The PCR mixture for tef1 contained 2 μL of template DNA and the cycling parameters to obtain the partial tef1 were as follows: an initial denaturation at 96 °C for 2 min; followed by 35 cycles of amplification [denaturation at 94 °C for 30 s, primer annealing at 56 °C for 30 s and primer extension at 72 °C for 30 s] and a final extension at 72 °C for 7 min, followed by storage at 8 °C. PCR-products were checked on 1.5% agarose electrophoresis gels containing HDGreenPlus DNA stain. Amplified PCR products were purified with the Cycle Pure Kit (VWR-Omega, USA). Sequencing was performed at Seqlab GmbH, Germany.

Molecular phylogeny
Amplification of the SSU, LSU, ITS and tef1 gene regions for all isolates used in this study yielded fragments of approximately 1100 bp, 900 bp, 650 bp and 300 bp, respectively. Consensus sequences of trace files were generated with Geneious 10.2.2 (https:// www.geneious.com, Kearse et al. 2012) and searched against GenBank (https://www. ncbi.nlm.nih.gov/, Benson et al. 2014) with MegaBLAST. Sequences with a high similarity (65 sequences of LSU, ITS and tef1 regions) were retrieved (Table 1). A total of 148 sequences for 65 specimens were obtained from GenBank (Table 1) and 92 sequences for 28 specimens from Benin were generated in this study (Table 2). They were aligned with MAFFT v. 7 using the L-INS-i algorithm (Nakamura et al. 2018). The alignments were manually checked by using MEGA v. 7 (Kumar et al. 2016). Gblocks v. 0.91b (Talavera and Castresana 2007) was used to remove poorly aligned positions and divergent regions from the DNA alignment using the parameters for a less stringent  For Maximum Likelihood analyses one thousand nonparametric bootstrap iterations were used with the generalised time-reversible model with a discrete gamma distribution (GTRGAMMA) (Stamatakis et al. 2008). For Bayesian phylogenies, two parallel runs with eight chains of Metropolis-coupled Markov chain Monte Carlo iterations were performed with the heat parameter being set at 0.2. Analyses were run for 100 million generations, with trees sampled every 1000 th generation until the average standard deviation of split frequencies reached 0.01 (stop value). The first 25% of saved trees were discarded as the 'burn-in' phase. Posterior probabilities (PP) were determined from the remaining trees. Bayesian posterior probabilities (BPP) ≥ 94% and Bootstrap values (BS) ≥ 70% are considered as significant.

Phylogeny
We isolated DNA from a total of 28 specimens of cercosporoid fungi recently collected in Benin (Table 2). These specimens represent 18 species found on species of Fabaceae for which 76 sequences are provided: 20 sequences of 18S rDNA, 16 of 28S rDNA, 21 of ITS and 19 of tef1. The separately aligned data sets for each marker consisted of 35 sequences/893 base pairs for 18S rDNA, 60/719 for 28S rDNA, 82/437 for ITS and 74/160 for tef1.
For the four-locus data analysis, DNA sequence data from the 18SrDNA, 28SrD-NA, ITS and tef1 gene regions were combined and submitted to Bayesian and Maximum Likelihood (ML) analyses. The final concatenated alignment contained a total of 91 specimens including the out-group (65 specimens from NCBI and 26 specimens from this study) and had an aligned length of 2212 characters including alignment Figure 1. The Bayesian phylogenetic tree inferred from DNA sequence data from the multigene alignment (SSU rDNA, LSU rDNA, ITS and tef1) of cercosporoid species. Nodes receiving Bayesian PP ≥ 0.94 or ML BS ≥ 70% are considered as strongly supported and are indicated by thickened branches. Names of newly described species are written in bold and red. Species newly reported for Benin are indicated by green letters. Names of host plants are written with blue letters.
gaps. As the ML analyses produced tree topologies mostly identical to results of Bayesian analyses, bootstrap support values of the ML trees were incorporated into the tree that resulted from Bayesian analyses (Fig. 1). In this tree, the cercosporoid fungi are grouped in three major clades: Cercospora (86/76), Pseudocercospora (87/78) and Passalora together with other species of other genera (100/98) (Fig. 1). Phylogenetic analyses of individual loci are deposited in TreeBASE. Details of results concerning the delimitation of species are mentioned and discussed as part of species notes below.
Tef1 sequence data showed differences between closely related species in the genera Cercospora and Pseudocercospora and are more informative than ITS and LSU rDNA sequence data. Therefore, we provide molecular phylogenetic analyses based on new tef1 sequences as well as sequences from GenBank for some newly described species, namely Cercospora rhynchophora, C. parakouensis, C. zorniicola and Pseudocercospora tabei. For Ps. sennicola, we provide an analysis based on ITS sequence data, because we were not able to obtain Tef1 sequence data.

Taxonomy
Based on morphological, molecular phylogenetic and host evidence, the cercosporoid fungi recently collected in Benin are assigned to 18 different taxa belonging to four genera. Among these, eight species are proposed as new to science, six in the genus Cercospora and two in Pseudocercospora. Eight species represent new reports for Benin, three of them are new for the whole of West Africa, namely Cercospora cf. canscorina, C. cf. fagopyri and C. phaseoli-lunati. Two species of cercosporoid fungi were previously reported in Benin and are confirmed. Etymology. The epithet beninensis refers to the country of origin of the type specimens, Benin.
Notes. The present Cercospora sp. on Calopogonium sp. also occurs on Vigna subterranea with different leaf spot appearances and caespituli. The lesions on Calopogonium sp. appear to be associated with a species of Pleosporales, whereas the leaf lesions on V. subterranea apparently are not associated with any other fungus and are dark reddish brown to dark brown with a dark margin, which are typical symptoms caused by Cercospora spp. The lesions on V. subterranea are larger and more abundant than those on Calopogonium sp., with abundant, dense caespituli and with dark greyish brown pigmentation (Fig. 2C).
Cercospora canescens is the only species of Cercospora known for Calopogonium spp. (Farr and Rossman 2021) and has been reported from West Africa (Guinea) on Calopogonium mucunoides (Lenné 1990). Apart from having slightly narrower conidia [(3-)3.5-4(-4.5) μm versus 2.5-5.5(-6) μm in C. canescens] as described by Chupp (1954), Hsieh and Goh (1990) and Mulder and Holliday (1975), the present specimen from Benin is morphologically identical to C. canescens. In the phylogenetic analyses, however, DNA sequences of the two specimens from Benin cluster together but separately from sequences of C. canescens available from India. In the multi-gene tree ( Fig. 1), C. canescens is located on a branch in a clade together with sequences of Cercospora spp. YMM3SO and YMM48SO on Sorghum bicolor (Poaceae) from Benin. C. canescens is known to correspond to a species complex that shows diverse morphological characteristics and genetic diversity (Joshi et al. 2006;Groenewald et al. 2013). Although C. canescens is an economically important species, no sequence data from the type or a neotype specimen are available (e.g., Groenewald et al. 2013). These are indispensable to resolve the C. canescens species complex. The specimens collected in Benin are tentatively placed into the species complex of C. canescens until DNA sequence data from the type locality (USA) and from diverse host species are available. C. aff. canescens is cited here for the first time for Benin (Piepenbring et al. 2020). Description. Leaf spots amphigenous, subcircular to irregularly angular, 2.5-8 mm diam., brown to reddish brown, with a dark margin. Caespituli amphigenous, greyish brown to brown. Mycelium internal. Stromata lacking or formed by few substomatal aggregated swollen hyphal cells. Conidiophores in small, loose fascicles to moderately large and dense fascicles of up to approx. 22 conidiophores, arising from internal hyphae breaking through the adaxial epidermis of the leaves or penetrating through stomatal openings, sometimes solitary, erect, straight, subcylindrical, 1-2 times geniculate, unbranched, (12-)20.5-68(-72) × (3-)3.5-4.5 μm, 0-6-septate, brown to dark brown. Conidiogenous cells terminal, usually monoblastic, sometimes polyblastic; loci apical or sometimes located on the shoulders of geniculations, 1.5-2.5(-3) μm wide, thickened and darkened. Conidia solitary, acicular to narrowly obclavate, straight to curved, 22-76(-80) × 2.5-3.5 μm, 1-7-septate, hyaline, smooth, tip acute, base truncate to short obconically truncate, 2-3 μm wide, hila thickened and darkened.
Hosts and distribution. On Canscora diffusa (Gentianaceae) from Khandala, West India (Chiddarwar 1959 (Farr and Rossman 2021). The present species from Benin is morphologically identical to C. canscorina (Chiddarwar 1959;Bhat and Pratibha 2010) except for narrower conidiophores with (3-)3.5-4.5 μm versus 3-7 μm in C. canscorina as mentioned by Bhat and Pratibha (2010). The original specimen of C. canscorina was not available for morphological examination and no DNA sequence data are currently published for this species. Therefore, a reliable species identification is not possible. The application of the name for the collections from Benin is tentative and must be verified based on sequences derived from the Indian type specimen or similar samples. Type. South Korea. Suwon, on Fagopyrum esculentum Moench (Polygonaceae), Sep 1934, K. Nakata & S. Takimoto (holotype specimen, not located and not preserved according to Groenewald et al. (2013), neotype: CBS H-21008, n.v).
Our sequence of the tef1 region of the specimen YMM23A from Benin is 100% similar to a sequence of Cercospora fagopyri on Fallopia dumetorum (GenBank JX143353) (Identities 233/233, i.e., 100%) and 99% similar to a further sequence of C. fagopyri on Fagopyrum esculentum (GenBank JX143352; Identities; 233/234, i.e., 99%). The identification of the present specimen as C. cf. fagopyri is only based on molecular data. Morphologically, descriptions of specimens of C. fagopyri on diverse host species in the literature differ and are quite confusing (Hsieh and Goh 1990;Groenewald et al. 2013). In order to establish a morphological concept and to know the host range of C. fagopyri, fresh specimens need to be collected once again on Fagopyrum esculentum in Korea, where this species was originally collected and pathogenicity needs to be proven for diverse host species. Etymology. The epithet parakouensis refers to the city of the type collection, Parakou, Benin.
In the multi-gene tree (Fig. 1), the ITS and the tef1 phylogeny (see Suppl. materials 3, 4), C. parakouensis forms part of a polytomy with a relatively large genetic distance (branch length) in relation to other sequences considered in the analysis.
Hosts and distribution. On Phaseolus lunatus from USA, Alabama, Tuskegee (type locality) (Braun and Crous 2005). This species is cited here for the first time for Benin. Thereby, it is cited for the first time for West Africa. Vigna radiata is a new host species.
Diagnosis. Cercospora rhynchophora differs from other Cercospora spp. known on Vigna spp. by causing distinct leaf spots, often well-developed stromata and up to 4 times geniculate conidiophores with often polyblastic conidiogenous cells with irregular, often beak-shaped tips.
Cercospora sp. YMM297B on Phaseolus lunatus L. Fig. 10 Description. Leaf spots almost lacking to well-developed, amphigenous, subcircular to irregularly angular, 2.5-8 mm diam., reddish brown, later dark brown by abundant caespituli, finally sometimes greyish brown to dark reddish brown, surrounded by dark margins, often with diffuse whitish centres. Caespituli amphigenous, greyish brown to dark brown. Mycelium mainly internal. External hyphae branched, 2-3(-4) μm wide, septate, olivaceous brown to brown, smooth. Stromata lacking or small, up to 20 μm diam., immersed in the mesophyll or in substomatal cavities, subcircular to irregular, olivaceous brown to darker brown. Conidiophores in small and loose fascicles, breaking through the adaxial epidermis of the leaves or penetrating through stomatal openings, sometimes solitarily arising through stomatal openings, erect, straight to sinuous, or somewhat geniculate, unbranched, (13-)17.5-195(-220) × (3.5-)4-5 μm, with 2-6(-8) septa each, occasionally slightly constricted and darker at the septa, brown to dark brown. Conidiogenous cells integrated, terminal, mainly monoblastic; loci 2-3.5 μm wide, thickened and darkened. Conidia solitary, narrowly obclavate to subacicular, straight to curved, (27-)36-148(-164) × (2.5-)3-4(-4.5) μm, with 2-7(-9) somewhat indistinct septa each, hyaline to sub-hyaline, smooth, apex subacute or acute, base truncate to short obconically truncate, 2-3(-3.5) μm wide, hila thickened and darkened. Notes. The infection of leaves of Phaseolus lunatus by Cercospora sp. YMM297B was associated with the infection by Pseudocercospora griseola. Among the Cercospora spp. known on Phaseolus and Vigna, C. olivascens is morphologically close to Cercospora sp. YMM297B. C. olivascens, however, differs from Cercospora sp. YMM297B by hypophyllous caespituli, no external hyphae, conidiophores that are up to five times geniculate and paler (Saccardo 1878;Chupp 1954), as well as hyaline conidia. The present specimen from Benin presents amphigenous caespituli, external hyphae, less geniculate and brown to dark brown conidiophores and often sub-hyaline conidia. C. olivascens also differs from the present species by being originally described from Aristolochia clematitis (Aristolochiaceae). According to Chupp (1954), this species was wrongly reported on Phaseolus vulgaris by Saccardo (1886). This was confirmed by Crous and Braun (2003). In the ITS phylogeny (see Suppl. material 3), Cercospora sp. YMM297B forms part of a polytomy with a relatively large genetic distance (branch length) in relation to other sequences considered in the analysis. In the tef1 phylogeny (see Suppl. material 4), it is not possible to distinguish this collection from several other Cercospora spp. As the description and sequence data are obtained only from a single specimen, the data are not sufficient for a final conclusion and the description as a new species. A reliable species characterisation is not possible until more collections become available. Etymology. The epithet tentaculifera refers to the ramified and flexible hyphae. Diagnosis. Cercospora tentaculifera differs from other Cercospora spp. on Vigna and Phaseolus in causing inconspicuous or no leaf spots, well-developed external hyphae, mainly adaxial caespituli and up to 435 μm long conidiophores that are constricted at the septa.
In the multi-gene (Fig. 1) and in the ITS phylogeny (see Suppl. material 3), C. vignae-subterraneae forms part of a polytomy with a relatively large genetic distance (branch length) in relation to other sequences considered in the analysis. In the tef1 phylogeny (see Suppl. material 4), it is not possible to distinguish C. vignae-subterraneae from other Cercospora spp. Based on the results of our comparative study, we propose C. vignae-subterraneae as a species new to science. Etymology. The epithet zorniicola refers to the host genus Zornia and "-cola" (lat. colere = to dwell).
Diagnosis. Cercospora zorniicola is characterised by external hyphae, unbranched conidiophores that are uniform in colour and width, with mostly monoblastic conidiogenous cells (Fig. 13).
Notes. Except for the presence of external hyphae and mostly slightly shorter conidiophores, the present specimen from Benin is morphologically identical to Ps. cruenta as known by literature (Chupp 1954;Deighton 1976). This identification is confirmed by results obtained by phylogenetic analyses based on tef1 sequence data (see Suppl. material 4). Ps. cruenta is a well-known pathogen causing leaf spot diseases on species of Vigna and allied genera. It can cause serious yield losses of up to 40% in cowpea (Sivanesan 1990 For synonyms see Crous and Braun (2003), Crous et al. (2006) Videira et al. 2017: 401, MBT378593, n.v.;Epitype: CBS H-19683, designated by Videira et al. 2017: 401, MBT378594, n.v.).
Notes. Four species of Pseudocercospora, namely Ps. cruenta, Ps. glycines (Cooke) Deighton, Ps. griseola and Ps. stizolobii are known agents of leaf spot diseases on Phaseolus spp. (Farr and Rossman 2021). The present Pseudocercospora sp. is phylogenetically ( Fig. 1) and morphologically well distinguished from Ps. cruenta, Ps. glycines and Ps. stizolobii ) by forming synnematous fascicles, longer and broader conidiophores and broader conidia. The morphology of this collection from Benin on P. lunatus fits well with the description of Ps. griseola.

Pseudocercospora sennicola
In the multi-gene tree (Fig. 1), Ps. sennicola is located in a polytomy at the end of a long branch reflecting a long genetic distance to other species included in the analysis. Morphologically, Ps. sennicola is distinct from all Pseudocercospora species known on species of Fabaceae from Benin by longer conidiophores [(16.5-) Based on a MegaBLAST search using the ITS sequence data, the closest matches in NCBI's GenBank nucleotide database were Pseudocercospora fuligena on Lycopersicon sp. (Solanaceae) from Thailand (GenBank GU214675; Identities 674/687, i.e., 98%), Pseudocercospora chengtuensis on Lycium chinense (Solanaceae) from South Korea (Gen-Bank GU214672; Identities 674/687, i.e., 98%) and Pseudocercospora atromarginalis on Solanum nigrum L. (Solanaceae) from South Korea (GenBank GU214671; Identities 673/687, i.e., 97%). Based on the result of our comparative study, we consider the present Pseudocercospora species on Senna occidentalis from Benin to represent a distinct species, which is described here. However, as sequence data are only available for Ps. sennae-multijugae, more molecular sequence data are needed to clarify the species delimitations among these twelve Ps. species on Senna spp. Etymology. The epithet tabei refers to the person who collected the type specimen, Affoussatou Tabé, mycologist at the University of Parakou, Benin.
In the multi-gene phylogeny (Fig. 1), Ps. tabei forms part of a polytomy with a large genetic distance (branch length) in relation to other sequences considered in the analysis. In the tef1 phylogeny, Ps. tabei clustered together with the isolates of Ps. cruenta on Vigna and Phaseolus form Benin (see Suppl. material 4). However, morphologically, the present species is clearly distinct from specimens of Ps. cruenta by having darker and shorter conidiophores and above all, shorter conidia [(20.5-)24-82(-84.5) μm] (

Discussion
The present study aims to increase the knowledge on the diversity of cercosporoid fungi in tropical Africa. Therefore, cercosporoid fungi collected on fifteen species of plants belonging to ten genera of Fabaceae found in Benin, West Africa, were characterised concerning their morphology, host species and DNA sequence data (18S rDNA, 28S rDNA, ITS and tef1). The specimens of cercosporoid species collected in Benin are attributed to groups corresponding to Cercospora, Pseudocercospora and a heterogeneous group around Passalora. The four-gene phylogenetic tree yielded results consistent with the current knowledge of generic relationships as presented in previous studies Groenewald et al. 2013;Nakashima et al. 2016). Species of Cercospora and Pseudocercospora form morphologically distinct groups that receive moderate support in the phylogenetic analysis (Fig. 1). In the Cercospora and Pseudocercospora clades, the lengths of branches of most new species (C. beninensis, C. rhynchophora, C. vignaesubterraneae, C. zorniicola, Ps. sennicola and Ps. tabei) are quite long (Fig. 1). This indicates a relatively large genetic and evolutionary distance from neighbouring species included in the analysis. The partial gene sequences of the protein-coding region tef1 and the combined analysis of four loci provided better results than single gene analyses of ITS and LSU rDNA for the differentiation of species of Cercospora and Pseudocercospora. Consequently, these molecular sequence data only allow to measure phylogenetic distances between the species. A similar situation has been found for Cercospora spp. by Bakhshi et al. (2015Bakhshi et al. ( , 2018 and for Pseudocercospora spp. by Crous et al. (2013a) and Silva et al. (2016). Fortunately, most species included in this study differ from each other by their morphology and host range. For example, Cercospora tentaculifera (YMM75) on Vigna unguiculata causes inconspicuous leaf spots and produces adaxial caespituli with large conidiophores (up to 435 μm) that are constricted at the septa (Figs 2H, 11). Thereby, this species is easily distinguishable from other Cercospora spp. known on species of Vigna and Phaseolus. C. zorniicola (YMM299) on Zornia glochidiata produces external hyphae and conidiophores that are unbranched and uniform in colour and width with usually monoblastic conidiogenous cells (Fig. 13). This is the first species of Cercospora known for the host genus Zornia.
For the morphological identification of all species included in this study, we examined about 50 type specimens and other specimens loaned from BPI, ILL and NY. As result of these analyses, dichotomous keys to the species of Cercospora and Pseudocercospora infecting members of Fabaceae known for Benin are presented (see below). The following morphological characteristics are helpful to separate species of Cercospora and Pseudocercospora: characteristics of leaf spots (distinctiveness, colour, size, form) and sporulation (distinctiveness, position on the leaf ), the stroma (size, density), the external hyphae (present/absent), conidiophores (form, size, branching, number and position of conidiogenous loci, form of conidiogenous cells), and conidia (form, size range) (comp. Deighton 1976;Crous and Braun 2003;Crous et al. 2013;Groenewald et al. 2013;Videira et al. 2017).
In order to obtain DNA sequence data, up to now, only cercosporoid fungi available as cultures have been used . Due to the fact that most cercosporoid fungi are not available as cultures, molecular sequences are available only for a small fraction of the species diversity of cercosporoid fungi known by morphological characteristics. It is often difficult to cultivate cercosporoid fungi, as this requires living fungal cells and a sterile environment to avoid contamination. As it was not possible to cultivate cercosporoid fungi in Benin, a technique for DNA isolation from dry specimens has been developed and successfully applied in the context of the present study for cercosporoid fungi for the first time. This direct DNA extraction method opens interesting possibilities to obtain DNA data of cercosporoid and other fungal plant pathogens especially in tropical countries.
The present study is the first effort towards generating molecular and morphological data for cercosporoid fungi in Benin, West Africa. We found 18 taxa, representing only a small fraction of the yet unknown species diversity of cercosporoid fungi (Piepenbring et al. 2020;Farr and Rossman 2021). Eight taxa found in this study are proposed as species new to science. Ten known species have been identified, including taxa important for agriculture such as Pseudocercospora cruenta and Ps. griseola on Phaseolus lunatus as well as Nothopassalora personata and Passalora arachidicola on Arachis hypogaea. Eight species are reported for Benin for the first time, with three of them namely, Cercospora cf. canscorina, C. cf. fagopyri and C. phaseoli-lunati, being new for West Africa.
New scientific data, such as species new to science, new records of hosts and for geographic areas, will help plant pathologists to develop efficient and sustainable disease management programs to control these fungal diseases and quarantine officials to take decisions based on scientific evidence. The plethora of novel and newly reported taxa collected on Fabaceae in Benin confirms that mycologists and phytopathologists in Africa have so far not given much attention to the species diversity of fungi occurring on plants, including species of economic relevance, such as those belonging to Fabaceae. Benin and other tropical African countries are likely to harbour highly diverse mycobiomes including cercosporoid fungi that still await discovery (Piepenbring et al. 2020). It is important to investigate them, because these unknown plant pathogens are or may become relevant as agents of emerging diseases that may spread and threaten cultivated plants worldwide. We hope that this study motivates further mycologists to study cercosporoid fungi in Benin, as well as in other countries of tropical Africa, and help to get a better understanding of cercosporoid fungal diversity worldwide.

Conclusions
The present study is a first step for the investigation of the diversity of cercosporoid fungi by an integrative approach including morphological, phylogenetic and ecological information. Taxonomic studies in this work generated eight newly described species, eight new records and the confirmation of two species of cercosporoid fungi that were previously reported from Benin. Previously, 12 cercosporoid fungi were known for Benin. The present work expands this number by adding 16 species of Cercospora and Pseudocercospora to this list, with a total of 28 species. These records together with herbarium specimens and molecular sequence data form a baseline for further studies in the field of systematics, ecology and phytopathology referring to cercosporoid fungi. This information will help plant pathologists to develop effective disease management programs and evidence-based quarantine regulations. The results obtained for a single family (Fabaceae) in easily accessible vegetation close to settlements suggest that many more taxa of cercosporoid fungi remain to be discovered on plants belonging to other family of plants in diverse habitats. In the future, more attention should be directed towards collecting cercosporoid and other pathogenic fungi from Benin as well as other parts of tropical Africa.