First Description of Akanthomyces uredinophilus comb. nov. from Hemipteran Insects in America

: Filamentous fungi of the genera Lecanicillium and Akanthomyces (Ascomycota: Hypocreales: Cordycipitaceae) have been isolated from a variety of insect orders and are of particular interest as biological control agents for phloem-sucking plant pests. Three aphid-and whiteﬂy-pathogenic fungal strains that had been isolated from naturally infected Trialeurodes vaporariorum and Myzus persicae in Argentina were assigned to the species Lecanicillium uredinophilum by combined analyses of morphology and ITS, LSU, EF1A, RPB1 and RPB2-based molecular taxonomy, giving rise to both the ﬁrst description of this fungus from hemipteran insects and its ﬁrst report from outside South-East Asia, especially from the American continent. A combination of phylogenetic reconstruction and analysis of pair-wise sequence similarities demonstrated that—reﬂecting recent changes in the systematics of Cordycipitaceae—the entire species L. uredinophilum should be transferred to the genus Akanthomyces . Consequently, the introduction of a new taxon, Akanthomaces uredinophilus comb. nov., was proposed. Moreover, extensive data mining for cryptic A. uredinophilus sequences revealed that (i) the fungus is geographically widely distributed, including earlier unrecognized isolations from further American countries such as the USA, Mexico, and Colombia, and (ii) entomopathogenic and mycoparasitic lifestyles are predominant in this species.


Introduction
The sap-sucking insects (Hemiptera) such as aphids and whiteflies have emerged as major agricultural pests with high economic relevance in numerous cropping systems [1,2]. These insects are exclusive phloem feeders and are among the most economically important pest insects of temperate agriculture [3]. Aphids and whiteflies are considered important pests on many crops, such as cotton, cucurbits, lettuce, potato, eggplant, pepper, soybean, sunflower, and tomatoes [3][4][5]. Aphids and whiteflies feed on plant nutrients essential for plant growth and development and induce phytotoxic effects by injecting saliva into the plant [6,7]. In addition to the impact of feeding, aphids also transmit plant viruses; more than 275, i.e., nearly 50% of all insect-borne plant viruses, are vectored by aphids [8,9].
The most prevalent approach to the management of hemipteran pests is the application of classical chemical insecticides [10,11]. However, insecticides commonly lose efficacy with the development of insecticide resistance, most notably in aphids and whiteflies [12][13][14].
Microbial control agents can serve as environmentally friendly components of integrated pest management (IPM) programs due to their selectivity, safety, and compatibility with other natural enemies [15]. Entomopathogenic fungi (EPF) are the most abundant group comprising app. 60% of all microbial insect pathogens [16]. Filamentous fungi proach, and introduction of the taxon designation Akanthomyces uredinophilus comb. nov. is proposed. Moreover, the Genbank database was mined for cryptic data giving proof of unrecognized previous isolations of L. uredinophilum fungi. Taken together, these findings confirmed that both A. uredinophilus and its versatile entomopathogenic-mycoparasitic lifestyle are globally distributed and do not just represent a regional South-East Asian variety or adaptation. Moreover, systematic reorganization of Lecanicillium uredinophilum in the genus Akanthomyces corroborates that the entomopathogenic and mycoparasitic fungi formerly organized around the "core species" of the genus Lecanicillium remain phylogenetically tightly linked to each other as part of a presumably monophyletic group even after being transferred to a new genus. The identification of hemipteran-pathogenic A. uredinophilus fungi in America and especially in Argentina paves the way for the development of endemic biological control agents against these highly relevant agricultural pests.

Morphological and Microscopic Characterization
Fungal species were identified based on both macroscopic and microscopic features. Morphological characterizations were made from in vitro cultures grown for 15 d at 24 • C on SDAY plates. Macroscopic features included the aspect, color and mycelium appearance of the fungal colonies. Microscope observations were made from fungal structures such as conidiophores, phialides and conidia. Fungal mycelia were mounted on glass slides and stained in lactophenol/cotton blue (0.01% w/v). Phialides and conidia were measured at magnifications of 400X using a model Axiostar Plus microscope (Zeiss, Germany). Furthermore, fungal structures were photographed with Moticam 3.0 MP Color Digital Camera (Motic, Xiamen, China). Semi-permanent slides were mounted, according to Humber [38]. Fungi were initially identified according to the taxonomic keys of Humber [38].

Virulence Bioassays
The virulence of three isolates, CEP054, CEP057 and CEP108, towards hemipteran insects was evaluated. The pathogenicity of fungal isolates CEP 057 and CEP 108 was tested on M. persicae apterous adults, as described in Manfrino et al. [34], while the activity of fungal isolates CEP 054 and CEP 057 was tested on T. vaporariorum nymphs according to Scorsetti et al. [39]. In all bioassays, the target insects were reared under laboratory conditions. For each isolate, 15 apterous adults of M. persicae and 25 fourth-instar nymphs of T. vaporariorum were used per replicate. The inoculation was performed by aspersion of suspensions of 1 × 10 7 conidia/mL in 0.01% (v/v) Tween 80 (sodium polysorbate). Controls were sprayed with solutions of 0.01% Tween 80. Three replicates and a control were performed for each isolate. The whole experiment was replicated twice. Cumulative mortality determined by the presence of mycelial masses was recorded daily over a period of 7-10 days [40]. Dead insects were removed daily and placed in plastic Petri dishes (60 mm) with water agar (1%). Petri dishes were incubated at 25 • C for 3-5 days to allow for fungal development. Dead insects were mounted in lactophenol/cotton blue (0.01% v/v) to check for fungal infection.

DNA Extraction, PCR Amplification and Sequencing
DNA extraction was carried out according to Manfrino et al. [36] using the DNeasy Plant kit (Qiagen). Purified DNA was finally eluted in 100 µL elution buffer (10 mM Tris-Cl, pH = 8.5). The following phylogenetic markers were amplified from fungal DNA samples using GoTaq DNA polymerase (Promega) with the PCR primers and conditions indicated in Table 1: internal partial sequences of the gene encoding translation elongation factor 1-alpha (EF1A) (primer pair EF1A-983F/EF1A-2218R), of the genes encoding the largest (RPB1) and second-largest (RPB2) subunit of RNA polymerase II (primers RPB1Af/RPB1Cr and RPB2-5f/RPB2-7r, respectively), of the gene encoding the large (28S or LSU) ribosomal RNA (primers LR0R/LR5), as well as the complete ribosomal RNA operon internal transcribed spacer (ITS) region (ITS1-5.8S-ITS2) (primers ITS4/ITS5). The generalized PCR protocol employed for marker amplification consisted of one initial denaturation step of 95 • C for 2 min, 35 cycles of 30 s at 95 • C, 30 s at the primer-specific annealing temperature, and a 72 • C elongation step of amplicon-specific time, followed by a 5 min final elongation step at 72 • C. PCR product size was confirmed by agarose gel electrophoresis, and DNA was purified using a QIAquick ® PCR purification kit (Qiagen, Germany). Additional primers used in combination with PCR primers for sequencing by StarSeq (Mainz, Germany) are indicated in Table 1. Raw sequence data were combined into a single consensus sequence for each fungal isolate and marker using version 11 of the MEGA software package [41]. Sequences were submitted to the GenBank database under accession numbers indicated in Table S1.

Phylogenetic Analysis
Fungal strains and marker sequences employed as a reference for phylogenetic reconstruction are identified in Table S1. Marker sequences were aligned at the nucleotide sequence level using the CLUSTAL W function [47], as implemented in the MEGA 11 software package. For a comprehensive analysis of protein-encoding sequences, a "meta-gene" sequence generated by concatenation of markers EF1A, RPB1 and RPB2 was analyzed. Pairwise sequence similarity percentages were assessed from p-distance matrices calculated in MEGA 11 from unfiltered nucleotide sequence data under pair-wise deletion of alignment gaps and missing data. Phylogenies were reconstructed using a p-distance matrix-based neighbor joining (NJ) method as implemented in MEGA 11. Tree topology confidence limits were explored in non-parametric bootstrap analyses over 1000 pseudo-replicates.

Database Mining for Cryptic Identifications of L. uredinophilum
Consensus sequences obtained from isolate CEP 057 for the five markers were used as queries in unfiltered BlastN searches across the Genbank database, allowing for up to 1000 similarity percentage sorted hits to be retained for analysis and applying a general 90% cut-off value for sequence coverage and the following marker specific sequence similarity cut-offs: ITS (98.0%), EF1A (94.0%), RPB1 (94.1%), RPB2 (94.0%), LSU (99.6%). Identified Genbank entries were aligned to the set of reference sequences (Table S1) and used for phylogenetic reconstruction in order to identify sequences clustering with references for L. uredinophilum.
Fungal strains and marker sequences employed as a reference for phylogenetic reconstruction are identified in Table S1. Marker sequences were aligned at the nucleotide sequence level using the CLUSTAL W function [47], as implemented in the MEGA 11 software package. For a comprehensive analysis of protein-encoding sequences, a "metagene" sequence generated by concatenation of markers EF1A, RPB1 and RPB2 was analyzed. Pair-wise sequence similarity percentages were assessed from p-distance matrices calculated in MEGA 11 from unfiltered nucleotide sequence data under pair-wise deletion of alignment gaps and missing data. Phylogenies were reconstructed using a p-distance matrix-based neighbor joining (NJ) method as implemented in MEGA 11. Tree topology confidence limits were explored in non-parametric bootstrap analyses over 1000 pseudoreplicates.

Database Mining for Cryptic Identifications of L. uredinophilum
Consensus sequences obtained from isolate CEP 057 for the five markers were used as queries in unfiltered BlastN searches across the Genbank database, allowing for up to 1000 similarity percentage sorted hits to be retained for analysis and applying a general 90% cut-off value for sequence coverage and the following marker specific sequence similarity cut-offs: ITS (98.0%), EF1A (94.0%), RPB1 (94.1%), RPB2 (94.0%), LSU (99.6%). Identified Genbank entries were aligned to the set of reference sequences (Table S1) and used for phylogenetic reconstruction in order to identify sequences clustering with references for L. uredinophilum.

Morphological and Microscopic Characterization
Colonies grown on SDAY medium reaching 20-25 mm after 12 days at 24 °C, white to cream, velvety-cottony, irregular ( Figure

Virulence Bioassays
The three fungal isolates under study were pathogens to both Trialeurodes vaporariorum and Myzus persicae. The mortality caused by strains CEP 057 and CEP 054 when assayed against the original host, T. vaporariorum, was 52.6% ± 8.3 and 65.0% ± 23.0, respectively, at seven days post-treatment [39]. Against M. persicae, the mortality caused by CEP 57 and CEP 108 was 38.8% and 25.5%, respectively, at 10 days post-treatment [36]. Mortality in controls was 10%.

Molecular Taxonomic Identification
For the three fungal isolates under study, consistent consensus sequences were obtained for the five molecular taxonomic markers used. Comparisons with reference sequences gave rise to alignments comprising 685 bp (EF1A), 515 bp (RPB1), 792 bp (RPB2), 819 bp (LSU), and 470 bp (ITS), respectively, of these marker sequences. Concatenation of EF1A, RPB1 and RPB2 sequences thus gave rise to a meta-gene alignment comprising 1992 bp. Moreover, the ITS sequence of the A. dipterigenus type strain CBS 126.27 as well as the EF1A, RPB1, RPB2 and LSU sequences of the A. attenuatus type strain CBS 170.76 were determined for comparison. All new sequences reported in this study were submitted to the Genbank database under accession numbers indicated in Table S1. Pair-wise p-distances were calculated from alignments of ITS, LSU, and concatenated EF1A, RPB1 and RPB2 sequences (Tables S2 and S3). Alignments of the same markers were used to reconstruct the phylogeny of Akanthomyces fungi (Figures 2-4).
All marker sequences determined were at least 99.9% similar across the three Argentine isolates under study (Tables S2 and S3). In the EF1A-RPB1-RPB2 phylogeny (Figure 3), CEP 054, CEP 057 and CEP108 formed an independent branch within a 100% bootstrapsupported clade, uniquely comprising L. uredinophilum reference strains from Korea and China, including the specific type strain. Pair-wise sequence similarities across this presumed L. uredinophilum clade were ≥99.3%. Sequence similarities between the Argentine isolates and Akanthomyces reference strains outside this clade were ≤95.8% (as calculated for both the A. muscarius and A. neocoleopterorum type strains), thus giving rise to a supposed taxon gap of about 3.5%.
In the ITS phylogeny (Figure 2), the Argentine isolates formed a 91% bootstrapsupported clade together with L. uredinophilum references from China displaying 99.8% sequence similarity; no ITS sequence data were available for comparison for the L. uredinophilum type strain. Sequence similarities with reference strains outside this clade were ≤98.5% (for both the A. lepidopterorum and A. neocoleopterorum type strains). A qualitatively similar picture arose from the comparison of LSU sequences (Figure 4, Table S3).
In all reconstructed phylogenies, the respective presumed L. uredinophilum clade was found in a neighboring position to clades representing Akanthomyces species derived from former Lecanicillium core species. In the EF1A-RPB1-RPB2 tree, L. uredinophilum appeared most closely related to A. muscarius, A. attenuatus, A. pissodis, A. neocoleopterorum, and A. lepidopterorum (98% bootstrap support), with the over next neighbors being A. lecanii and A. sabanensis. Independent from the exact neighboring relationships, the L. uredinophilum clade appeared tightly integrated into the genus Akanthomyces as reorganized by Kepler et al. [24] for all three phylogenies.
These results from the molecular taxonomic analysis were consistent with both (i) the assignment of the three Argentine isolates CEP 054, CEP 057 and CEP108 to the taxonomic species Lecanicillium uredinophilum and (ii) the reorganization of this species in the genus Akanthomyces.

Database Mining for Cryptic Identifications of L. uredinophilum
Using the ITS sequence of isolate CEP 057 as a query for BlastN search across the Genbank nucleotide database, a total of 261 entries passed the 90% query coverage, and 98% pair-wise sequence similarity thresholds were applied. Out of these entries, 19 were found to associate with the presumed L. uredinophilum clade in phylogenetic reconstruction ( Figure 2). Elimination of redundancies, e.g., entries from the same origin or study, led to the identification of 11 unrecognized descriptions of L. uredinophilum. Analogous database searches for the four other markers revealed two L. uredinophilum-associated LSU sequences ( Figure 4); for both fungal specimens identified, ITS sequences were available in the database, but only one of these had already been identified in our ITS search. No further L. uredinophilum-associated entries were identified for the EF1A, RPB1 and RPB2 markers.     In all reconstructed phylogenies, the respective presumed L. uredinophilum clade was found in a neighboring position to clades representing Akanthomyces species derived from former Lecanicillium core species. In the EF1A-RPB1-RPB2 tree, L. uredinophilum appeared most closely related to A. muscarius, A. attenuatus, A. pissodis, A. neocoleopterorum, and A. lepidopterorum (98% bootstrap support), with the over next neighbors being A. lecanii and A. sabanensis. Independent from the exact neighboring relationships, the L. uredinophilum The identified cryptic reports of L. uredinophilum sequence data (Table 2) spanned a wide range of geographic origins and isolation sources. Without being perceived as such, L. uredinophilum had previously been reported from South-East Asia (Tibet, China), America (Colombia, Mexico, USA), Africa (Kenya, South Africa), Europe and the Middle East (Germany, Turkey, Iran) and twice from New Zealand. Four reports associated the fungi described with insect hosts, three with basidiomycete fungi, one with the entomopathogenic ascomycete Ophiocordyceps sinensis, two with plant hosts, and one specimen had been isolated from a sweet water sample. At the moment of description, most specimens were assigned to existing Lecanicillium or Akanthomyces species different from L. uredinophilum/A. uredinophilus, but at least two reports discuss-at the level of Genbank entries-the introduction of a new fungal taxon to be termed "Verticillium zealandica" or "Lecanicillium zaquensis", respectively.

Discussion
The molecular taxonomic analysis presented above firstly demonstrated that the L. uredinophilum type strain and further reference strains form a distinct, presumably monophyletic clade firmly located within the fungal species Akanthomyces, as reorganized by Kepler et al. [24]. In conclusion, following the example of the former core species of the genus Lecanicillium, the species L. uredinophilum should be reorganized into a new taxon to be named Akanthomyces uredinophilus comb. nov.; see the taxonomic description below. Moreover, three Argentine fungal isolates naturally infecting hemipteran insects were morphologically, microscopically and molecular-taxonomically characterized. The ITS, EF1A, RPB1, RPB2 and LSU marker-based molecular analysis firmly located strains CEP 054, CEP 057 and CEP 108 in the L. uredinophilum/A. uredinophilus clade.
Morphologically, the description of the isolates was consistent with the characteristics used to describe L. uredinophilum, i.e., being similar to the former Lecanicillium core species in having verticillate conidiophores, gradually tapering phialides, and ellipsoidal to oblongoval aseptate conidia [17,20,29], but differing from A. lecanii, A. dipterigenus, A. attenuatus and A. muscarius by the longer phialides. Therefore, taking molecular and microscopic results together, isolates CEP 054, CEP 057 and CEP 108 were conclusively assigned to the new species Akanthomyces uredinophilus.
Former reports on L. uredinophilum, now A. uredinophilus, were exclusively from South-East Asia: Park et al. [20] introduced the species as a mycoparasite of rust fungi (Uredinales) in Korea, Wei et al. [34] described entomopathogenic L. uredinophilum isolates from China, and Meng et al. [35] concluded that one component (isolate QHLA) of a fungal entomopathogen complex from Tibet was L. uredinophilum. The Argentine strains characterized in this study represent the first report of this species from other parts of the world, especially from the American continent. Moreover, it is the first time the natural infection of hemipteran insects has been related to L. uredinophilum.
With respect to strain QHLA assigned by Meng et al. [35] to L. uredinophilum, the molecular-taxonomic analysis presented here has lent only weak support to this assignment. Whereas in the LSU phylogeny strain QHLA co-localized-under insufficient bootstrap support-with L. uredinophilum reference strains (Figure 4), it appeared unconnected to the L. uredinophilum clade in both the ITS and EF1A-RPB1-RPB2 trees (Figures 2 and 3). However, in all phylogenies, strain QHLA appeared tightly linked to a further fungal specimen, termed strain HMAS 246917. As strains QHLA and HMAS 246917 share both a rather specific source of isolation, i.e., the Ophiocordyceps sinensis complex, and their geographic origin, i.e., Qinghai province in Tibet, and have identical ITS and RPB1 marker sequences, one might expect them to stem from the same isolation event. However, for both specimens, LSU contradicts ITS, EF1A, RPB1 and RPB2 sequence data; a conclusive species level assignment is not possible on the basis of the currently available sequence data. There does not appear to date to be a formal description available for strain HMAS 246917, but the corresponding Genbank entries (MT789698, MT797809, MT797811) considered assignment to a new species to be termed "Lecanicillium zaquensis", an option much more in line with above molecular-taxonomy results than assignment to L. uredinophilum.
Beyond the above-mentioned explicit descriptions of fungal isolates, such as L. uredinophilum, data mining has revealed eleven cryptic reports of A. uredinophilus that had not been recognized as such (Table 2). Most authors of these cryptic reports assigned the fungus under study to either A. muscarius or A. attenuatus. Interestingly, the earliest identified report considered the introduction of a new species under the (invalidly published) designation "Verticillium zealandica" [48]. These cryptic reports demonstrated beyond the limited previously available knowledge that mycoparasitism and entomopathogenicity are the predominant, globally distributed ecophyiological lifestyles of A. uredinophilus. Moreover, together with earlier studies, including strains CEP 054, CEP 057 or CEP 108 [36,39], the reports contain valuable information with respect to the potential application of A. uredinophilus for biological control of insect pests or fungal phytopathogens in agriculture. The results of virulence bioassays, for instance, demonstrated that CEP 054, CEP 057 or CEP 108 were pathogens to aphids and whiteflies. Mortalities produced on whiteflies were higher than those on aphids, reaching a 65% mortality rate in T. vaporariorum and 38% in M. persicae [34,37]. Marshall et al. [48] tested the activity of L. muscarium isolates, revealed by data mining to be A. uredinophilus, against Scolypopa australis (Hemiptera: Ricaniidae) and showed that the isolates were highly pathogenic to S. australis. More-over, a further ecophysiological trait linked here for the first time to L. uredinophilum or A. uredinophilus, namely endophytism, might open a future option for the development of respective biocontrol strategies.
Future attempts to tap into the potential of A. uredinophilus fungi, especially for hemipteran biocontrol, can rely on experience made with related Akanthomyces species. Wang et al. [53] found that L. attenuatum species showed high efficacy against nymphs and adults of Acyrthosiphon pisum Harris. Askary et al. [52] and Kim et al. [54,55] tested the activity of three Lecanicillium spp. isolates against aphids and Sphaerotheca fuliginea (the causal agent of cucumber powdery mildew). These authors observed that strains may have potential for development as a single microbial control agent effective against several plant diseases, pest insects and plant parasitic nematodes due to their antagonistic, parasitic and disease resistance-inducing characteristics. Broumandnia et al. [56] studied the potentials of four Iranian isolates of A. lecanii and A. muscarius to control B. tabaci on cucumber under laboratory conditions. The authors found that all isolates, especially A. muscarius (AGM5), exhibited appropriate potential as a biological control agent against B. tabaci. Lu et al. [57] tested the virulence of four strains assigned to L. longisporum, L. attenuatum and A. lecanii against B. tabaci and found that all isolates were pathogenic for this insect species. Akanthomyces fungi, including potentially versatile A. uredinophilus, should therefore be evaluated as powerful components of next-generation sustainable agriculture [58].

Conclusions
Molecular taxonomy demonstrated that the fungal species Lecanicillium uredinophilum has to be transferred to the genus Akanthomyces, giving rise to the new taxon combination Akanthomyces uredinophilus. Morphological and molecular-taxonomic identification of three hemipteran-associated fungal isolates from Argentina introduced a new host and a new country record to A. uredinophilus. Database mining for cryptic A. uredinophilus sequence reports revealed a wide geographic distribution of this fungus and its entomopathogenic or mycoparasitic lifestyles. Analysis of earlier reports and virulence bioassays with the Argentine strains against aphids and whiteflies indicated the potential of A. uredinophilus for the biological control of insect pests and fungal phytopathogens in agriculture. Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/d14121118/s1: Table S1: Fungal strains and marker gene sequences used in this study. Table S2: Pair-wise p-distance values for aligned ITS (lower left triangle) and EF1A-RPB1-RPB2 (upper right) marker sequences. Table S3: Pair-wise p-distance values for aligned LSU marker sequences.  Data Availability Statement: Sequence data analyzed in this study are publicly available from the Genbank database (https://www.ncbi.nlm.nih.gov) under nucleotide sequence accession numbers listed in Table 2 and Table S1 to this study.

Conflicts of Interest:
The authors declare no conflict of interest.