Insights of the Neofusicoccum parvum–Liquidambar styraciflua Interaction and Identification of New Cysteine-Rich Proteins in Both Species

Abstract

Neofusicoccum parvum belongs to the Botryosphaeriaceae family, which contains endophytes and pathogens of woody plants. In this study, we isolated 11 strains from diseased tissue of Liquidambar styraciflua. Testing with Koch’s postulates—followed by a molecular approach—revealed that N. parvum was the most pathogenic strain. We established an in vitro pathosystem (L. styraciflua foliar tissue–N. parvum) in order to characterize the infection process during the first 16 days. New CysRPs were identified for both organisms using public transcriptomic and genomic databases, while mRNA expression of CysRPs was analyzed by RT-qPCR. The results showed that N. parvum caused disease symptoms after 24 h that intensified over time. Through in silico analysis, 5 CysRPs were identified for each organism, revealing that all of the proteins are potentially secreted and novel, including two of N. parvum proteins containing the CFEM domain. Interestingly, the levels of the CysRPs mRNAs change during the interaction. This study reports N. parvum as a pathogen of L. styraciflua for the first time and highlights the potential involvement of CysRPs in both organisms during this interaction.

1. Introduction

A current challenge in the area of plant–pathogen interactions is to comprehensively understand the molecular mechanisms involved in the defense response of plants against a specific pathogen. Improved prevention and diagnostic strategies should therefore be designed for more effective disease management. Fungi are important causal agents of plant diseases, and different species among the Botryosphaeriaceae family are recognized as endophytic and latent pathogens of trees and other woody plants, and are responsible for causing cankers, dieback, and even blight [1]. Neofusicoccum parvum belongs to the Botryosphaeria genus and has been recognized as a widespread pathogen worldwide [2], causing damage to a wide range of hosts, including agricultural, horticultural, and forestry plants. Its hosts of economic importance include grapevine (Vitis vinifera) [3], avocado (Persea americana Mill.) [4,5,6], blueberry (Vaccinium spp.) [7], pomegranate (Punica granatum) [8], peach (Prunus persica) [9], and walnut (Juglans regia) [10]. Its forestry host plants include the coast redwood (Sequoia sempervirens) [11], Norfolk Island pine (Araucaria heterophylla) [12], Eucalyptus globulus [13,14], and Ginkgo biloba [15]. Most of the relevant studies published to date have focused on isolation of the fungus from symptomatic or asymptomatic tissue, followed by molecular identification and pathogenicity testing using Koch’s postulates. Genomic analysis of UCR-NP2 isolated from grapevine in which the presence of glycoside hydrolases, polysaccharide lyases, cutinases, lignin peroxidases, and cytochrome P450 monooxygenases that might be involved in lignin degradation, and likely also in pathogenesis, were revealed [16]. It is already known that N. parvum is able to produce a variety of phytotoxic compounds that belong to the chemical families of dihydrotoluquinone, epoxy lactone, dihydroisocoumarin, and hydroxybenzoic acid, such as (-)-terremutin, (R)-mellein, and 6-methyl-salicylic acid [17]. Recently, exogenous application of (R)-mellein and (-)-terremutin in V. vinifera produced upregulation of the salicylic acid (SA)-responsive genes PR1 and GST1. In contrast, the expression of jasmonic acid (JA)/ethylene (ET)-responsive genes remained weak or was even downregulated by (-)- terremutin [18]. Necrosis and ethylene-inducing proteins (NLPs) have recently been identified in this fungus and, using biotechnological approaches, pure recombinant proteins were determined that are toxic to both plant and mammalian cells [19]. To date, there has been no effective control strategy against N. parvum, and horizontal transmission of this pathogen [20] has been considered to increase the risk of infection in different hosts. This study reports for the first time that N. parvum is a pathogen of L. styraciflua and, through molecular approaches, provides insights into this interaction. Liquidambar styraciflua L. (Altingiaceae) is a deciduous timber tree native to the Americas that is distributed in localized regions in North and Central America [21], and has been widely introduced in eastern and central China [22]. This tree, commonly known as American sweetgum, is an attractive hardwood species for potential bioenergy production [23] and is often used for reforestation, agroforestry, and landscaping [24].

Study of specific proteins in both host and pathogen is a good approach to gain knowledge of the interaction. Small proteins with a high percentage of cysteine residues that possess novel functions during plant pathogen interaction are an attractive group of proteins [25]. Examples of these kind of cysteine-rich proteins (CysRPs) include SCR96 of Phytophthora cactorum that trigger cell death in Nicotiana benthamiana and tomato, silenced transformants of scr96 lose their pathogenicity on host plants [26]. The cysteine-rich small protein SsSSVP1 in Sclerotinia sclerotiorum induce plant cell death when interacting with QCR8, a subunit of cytochrome b-c1 complex, and disturbs the localization of QCR8 in mitochondria, suggesting that SsSSVP1 manipulate plant energy metabolism to facilitate the infection of S. sclerotiorum [27]. For Verticillium dahliae, three CysRPs: VdSCP27, VdSCP113, and VdSCP126 expressed in Nicotiana benthamiana caused cell death, a reactive oxygen species burst, and the induction of defense genes [28].

In turn, during the defense response, the plants can produce proteins rich in cysteines with antifungal activity, such as defensins, lipid-transfer proteins, snakins, hevein-like peptides, knottins, hairpinins [29], and cyclopeptides [30]. Snakins/GASA proteins StSN1 from potato and MsSN1 from alfalfa, showed antifungal activity against many fungi and overexpression of StSN1 in potato and wheat improved plant resistance to commercially important pathogens [31]. Hevein-like peptides of Ginkgo biloba inhibited the hyphal growth of Aspergillus niger, Curvularia lunata, Fusarium oxysporum, and Rhizoctonia solani [32], and finally cyclotides cyO2, cyO3, cyO13, cyO19 of Viola odorata showed a potent activity against Fusarium oxysporum, Fusarium graminearum, Fusarium culmorum, Mycosphaerella fragariae, and Botrytis cinerea [30].

Here, we established a reproducible in vitro pathosystem using foliar tissue of L. styraciflua and characterized the infection process using scanning electron microscopy (SEM). We also identified cysteine-rich proteins (CysRPs) of both the plant and the fungus that may be important to the defense response and infection process. Evidence indicates that effector proteins are highly specific and inherent to each pathogen. Finally, no information has been available regarding the CysRPs in any of the species of the Botryosphaeriaceae family.

2. Materials and Methods

2.1. Plant Material

Liquidambar styraciflua leaves with disease symptoms such as chlorosis and necrosis were sampled from five different areas with trees located within the cloud forest of Santuario del Bosque de Niebla (SBN) in Xalapa, Veracruz, Mexico (19°30′37″ N 96°56′34″ W). This forest is a protected area belonging to the Institute of Ecology (INECOL A.C.) and is used to collect samples for research purposes. Identification of Liquidambar specimens was performed by the curators Carlos Duran and Sergio Avendaño in the Herbarium-XAL of INECOL A.C. Leaves and stems used for artificial inoculation were collected from healthy trees in SBN.

2.2. Isolation of Fungi from L. styraciflua Leaves

Complete leaves were cut with gardening scissors, deposited in plastic bags, and kept at 4 °C. The leaves were cut into small pieces of 0.5 cm × 0.5 cm and disinfected with ethanol (96%) and chlorine solution (20%) for 2 min each rinsed five times with sterile water. For fungal isolation, the disinfected tissue was placed in Petri dishes containing potato dextrose agar (PDA) and incubated for 2–4 days at 25 or 29 °C under two different light conditions: darkness or a 12 h light/12 h dark. Monosporic cultures were grown for all of the isolated fungi that showed different morphologies. The isolates were preserved in 10% glycerol at −80 °C.

2.3. DNA Extraction and Molecular Identification of the Liqui 1–3–Isolate by PCR

Each isolate was grown on PDA in Petri dishes of diameter 10 cm for 9–10 days at 28 °C in darkness. The mycelium was collected with a scalpel and pulverized to isolate genomic DNA according to the protocol described by Tapia-Tussell et al. [33], with minor modifications. Specifically, in the lysis step, we added 1 mL of SDS buffer, while the purification step was replaced with the addition of 500 µL of phenol:chloroform:isoamyl alcohol (Sigma-Aldrich, St. Louis, MO, USA). For molecular identification of the Liqui 1−3 isolate, previously published primers that amplify regions of the β-tubulin gene [34]; the internal transcribed spacer (ITS) regions; a portion of the RNA polymerase II subunit (RPB2) [35]; the BotF15 locus, an unknown locus containing microsatellite repeats [36]; and the portion of the gene encoding translation elongation factor 1 alpha (EF-1α) [35] (see Table S1) were used. Genes from these five regions were amplified from genomic DNA using PCR with Platinum^® Taq DNA Polymerase, High Fidelity (Invitrogen™ Thermo Fisher Scientific, Waltham, MA, USA), following the manufacturer’s recommendations. The optimal PCR conditions on a SureCycler 8800 (Agilent Technologies, Inc., Santa Clara, CA, USA) were as follows: initial denaturation at 94 °C for 1 min, followed by 32 cycles of 94 °C for 15 s, 58 °C (β-tubulin, RPB2 and Botf15)/55 °C (ITS and EF-1α) for 30 s, and 72 °C for 1 min, with a final extension step of 72 °C for 5 min. The size and concentration of the amplicons were visually determined by electrophoresis on a 1% agarose gel, and PCR products were purified using the Wizard SV Gel and PCR Clean-up System (Promega, Madison, WI, USA). The DNA concentration was measured using a NanoDrop 2000c spectrophotometer (Thermo Fisher Scientific, USA). Samples were sent to Langebio, Cinvestav, for sequencing. DNA sequences were analyzed using the nucleotide collection (nr/nt), optimized for highly similar sequences (MegaBLAST) in NCBI (https://blast.ncbi.nlm.nih.gov/Blast.cgi, 26 June 2019).

2.4. Pathogenicity Assay in Arabidopsis Thaliana Seedlings

Seeds of A. thaliana ecotype Col-0 were surface-disinfected with 96% (v/v) ethanol for 7 min and 20% (v/v) bleach for 7 min, rinsed five times for 5 min with distilled water and stratified for 2 days at 4 °C. Seeds were grown on each agar plate (10 seeds per plate) containing 0.2 × Murashige and Skoog medium (MS basal salt mixture, PhytoTechnology Laboratories, Lenexa, KS, USA), 0.6% sucrose (Merck, Darmstadt, Germany) and 1% Agar Plant TC at pH 7. Plates were incubated at 21 ± 1 °C under a cycle of 12 h light/12 h dark for 7 days. The seedlings were then inoculated with 0.5 cm × 0.5 cm agar plugs containing the isolate, and the plates were incubated for seven additional days under the same conditions. A photograph was then taken with a camera D3200 (Nikon, Tokyo, Japan). All of the experiments were carried out in triplicate.

2.5. Pathogenicity Assay in L. styraciflua Leaves and Stems

Healthy L. styraciflua leaves and stems were collected from the cloud forest of the SBN at Xalapa, Veracruz, Mexico. Once collected, the samples were immediately placed in plastic bags that contained sterile water to preserve the moisture. The leaves were collected and sterilized on the same day that the experiment was conducted. The leaves were sterilized with 2% sodium hypochlorite for 1 min and then washed five times with sterile distilled water. Three sterilized leaves were placed in a humid chamber. The humid chambers were prepared by placing a circle of sterile filter paper in the bottom of a Petri dish (150 × 20 mm) that had been previously sterilized with UV light for 15 min. Approximately 4–5 mL of sterile distilled water was added to the filter paper. The leaves (3) were mechanically damaged at the base with a sterile scalpel and then inoculated with a plug (approximately 0.5 × 0.5 cm) of N. parvum (Liqui 1–3) previously cultured on PDA medium. Leaves with damaged base were used as controls. Finally, the control and inoculated leaves were incubated at 25 ± 2 °C with 80% relative humidity in darkness in a plant growth chamber (Thermo Scientific, USA). Photographs were taken at 0, 1, 3, 8, and 18 days post infection (dpi) with a camera D3200 (Nikon, Japan). Stem infection experiments were carried out as follows: previously sterilized young branches of an adult L. styraciflua tree were cut with a scalpel, or single-edged knife, into fragments of approximately 10 cm, and then cut longitudinally. The L. styraciflua stems were placed facing upward in quadruplicate inside a humid chamber prepared as previously described. The plant stems were inoculated with 0.5 × 0.5 cm mycelium plugs of N. parvum (LSH1-083) previously cultured on PDA medium, at the center of the vascular tissue. A PDA plug with no fungus was included as a negative control. Finally, the sample was incubated at 25 ± 2 °C at 80% relative humidity in darkness in a plant growth chamber (Thermo Scientific, USA).

2.6. DAB Staining

Infected and uninfected L. styraciflua leaves were stained following a previously reported protocol [37].

2.7. SEM

Sections of infected and uninfected L. styraciflua leaves, at 1, 3, and 8 dpi, were fixed in 2.5% glutaraldehyde for 12 h and then washed in 0.1 M Sorensen’s phosphate buffer. The samples were dehydrated in an ethanol solution series from 30 to 100% for 40 min each, dried in a critical point drier K850 (Quorum Technologies Ltd., Laughton, UK), coated with gold in a rotary pumped sputter coater Q150RS (Quorum Technologies Ltd., UK) [38], and then examined under a scanning electron microscope Quanta^TM FEG 250 (Thermo Fisher Scientific, USA).

2.8. CysRP Identification

For identification of the CysRPs of L. styraciflua and N. parvum, the Hardwood Genomics Project database (https://www.hardwoodgenomics.org/; accession: PRJNA273273) and the genomic information already published by Blanco et al. [16] (accession: PRJNA244484) were consulted, respectively. Prediction of subcellular localization for all proteins, as well as signal peptide cleavage, were performed with the TargetP 1.1 server [39] and Protter [40]. The predicted localization was evaluated with SignalP-5.0. [41] and DeepLoc-1.0 [42]. BLAST (National Center for Biotechnology Information; https://blast.ncbi.nlm.nih.gov/Blast.cgi), Motif search (GenomeNet; https://www.genome.jp/tools/motif/MOTIF.html) and DeepGOPlus [43] approaches were used to infer the possible functions of the CysRPs. Theoretical molecular weight and isoelectric point were predicted with compute pI/Mw tool (https://web.expasy.org/compute_pi) [44].

2.9. Phylogenetic Trees

Phylogenetic analysis was performed with Mega X software [45] using the maximum likelihood method and JTT matrix-based model [46]. Jalview software was used to perform the alignments [47] with the T-Coffee method [48] and, finally, the BLAST tool was used to find similar sequences in other fungi.

2.10. RNA Extraction and Gene Expression

Total RNA of whole leaves at 24 and 72 h infected and no infected (but mechanically injury) was extracted using the Plant/Fungi Total RNA Purification Kit (Norgen BioTek Corp., Thorold, ON, Canada), following the manufacturer’s recommendations. The concentration of RNA was determined with a NanoDrop™ 2000c spectrophotometer (Thermo Fisher Scientific, USA), and RNA integrity was evaluated based on the A260 nm/A280 nm ratio and by electrophoresis in a 1.5% agarose gel. One µg of the RNA was then treated with deoxyribonuclease I from Invitrogen™ (Thermo Fisher Scientific, USA), following the manufacturer’s instructions, after which the RNA was used as a template for synthesis of cDNA with the reverse transcriptase Superscript III Invitrogen™, again following the manufacturer’s protocols.

Quantitative real-time PCR (qRT-PCR) was carried out using SYBR™ Green PCR Master Mix (Thermo Fisher Scientific, USA) in a final reaction volume of 20 μL, containing 10 μL of SYBR™ Green (Thermo Fisher Scientific, USA), 1.0 μL of each primer (10 mM), and 8 μL of cDNA (50 ng or 5.0 ng). For determination of oligo efficiency, curves with 50.0, 5.0, 0.5, 0.05, and 0.005 ng of cDNA were generated. Real-time PCR was performed in a Stratagene Mx300P system (Agilent Technologies, USA) under the standard thermal profile design: 10 min at 95 °C, followed by a total of 40 cycles of 30 s at 95 °C, 1 min at 55 °C and 1 min at 72 °C. Oligo specificity was determined by a melting curve analysis with continuous fluorescence data acquisition during the 55–95 °C melt. The software Primer3 was used for oligo design, following the parameters suggested by Thornton and Basu [49]. Relative expression levels for validated genes were calculated by the equation [50];

$$ratio=\frac{{\left(E_{target}\right)}^{\mathsf{\Delta }C{P}_{target}\left(control-Sample\right)}}{{\left(E_{ref}\right)}^{\mathsf{\Delta }C{P}_{ref}\left(control-Sample\right)}}$$

and oligos for Tip41, actin and ubiquitin of L. styraciflua were evaluated as reference genes (all primers are listed in Table S1). The geometric mean [51] between Tip41 and actin was used as a references [52,53].

3. Results

3.1. Identification of N. parvum as a Pathogen of L. styraciflua

Eleven fungi were isolated from L. styraciflua leaves with visible symptoms, such as necrosis and discoloration. The potential pathogenicity of the fungi was tested in L. styraciflua leaves and in seedlings of the model plant, A. thaliana (Figure 1 and Figure S1). In L. styraciflua, Liqui 1–3 provoked clear necrosis and leaf discoloration, including zones at the main veins and petiole, at 8 dpi. Liqui 1–02 and Liqui 2–2 also triggered disease symptoms in L. styraciflua; however, in contrast to the effect produced by Liqui 1–3, no noticeable change was observed in petiole coloration. The calculated percentage of damage (relationship between the size of the leaf and the size of the lesion) showed no statistically significant differences among the three strains: a value of 26.8 ± 4.7% damaged leaf area was found for Liqui 1–3, while the values were 15.1 ± 8.5% and 20.1 ± 1.2% for Liqui 1–02 and Liqui 2–2, respectively.

In A. thaliana seedlings, Liqui 1–3 covered all of the plant tissue, inducing severe leaf discoloration at 7 dpi. Interestingly, in A. thaliana seedlings, Liqui 1–02 and Liqui 2–2 did not have a critical pathogenic effect. For example, while Liqui 2–2 covered more than 50% of the seedlings, the foliar tissue of the infected plants showed greater vigor than that of the control. Liqui 1–2–01, Liqui 1–2–03, and Liqui 3–2 at 7 dpi showed a different pathogenic effect: Arabidopsis thaliana plants developed shorter primary roots but an increased number of secondary roots during the interaction with these isolates were observed. The isolates Liqui 1–04, Liqui 1–01, Liqui 3–3, and Liqui 3–1 had no effect on L. styraciflua leaves or A. thaliana seedlings. Since Liqui 1–3 was the only strain that showed a clear pathogenic effect on both L. styraciflua and A. thaliana plants, we identified this strain at molecular level and established the pathosystem in L. styraciflua leaves. To identify the Liqui 1–3 strain at the molecular level, universal and specific primers were used (Figure S2 and Table S1). The analysis revealed that Liqui 1–3 belongs to the Botryosphaeriaceae family as a species N. parvum.

3.2. Establishment of the L. styraciflua–N. parvum Pathosystem

Since we were interested in studying the interaction between L. styraciflua and N. parvum, we established a pathosystem using the L. styraciflua leaves. The graph representing the infected leaf area per dpi shows that the infection process progresses rapidly. The inoculated leaves displayed a brownish discoloration at 1 and 3 dpi, which was accentuated over time (Figure 2), and there was visible distinct petiole necrosis at 8 and 16 dpi. The presence of a whitish mycelium in the necrotic leaf area was also evident.

Since it was evident that the disease symptoms appear at early stages, diaminobenzidine (DAB) staining to indirectly detect the hydrogen peroxide production in L. styraciflua leaves at 1 and 3 dpi was used. The presence of a dark brown precipitate was detected in the infected leaves at early time points (1 and 3 dpi), indicating the presence of H₂O₂ (Figure 3 and Figure S3).

The SEM images revealed that the fungus grew robustly on the leaf (adaxial) surface, forming a hyphal mass and causing tissue degradation in which the cuticle and wax integrity were compromised (Figure 4a–d). In addition, the infection provoked petiole degradation. A transverse cut of the leaf base showed that the fungus developed pycnidia, asexual reproductive structures. These pycnidia appeared individually or as aggregates embedded in the plant tissue with thick walls (Figure 4i–l). A longitudinal section of pycnidium showed mature conidia. The conidiogenic oval cells were presented perpendicular to the walls of the pycnidium.

Finally, since the Botryosphaeriaceae family members are characterized as woody-plant pathogens, the pathogenicity of the Liqui 1–3 strain on fresh stems of L. styraciflua was tested. As shown, N. parvum at 7 dpi triggered distinct symptoms of disease, including discoloration and necrosis, covering a zone beyond the site of inoculation (Figure 5).

3.3. Identification and Description of CysRPs in L. styraciflua and N. parvum

CysRPs have been widely studied for their important functions in plant–pathogen interactions, in both the host and pathogen. To identify CysRPs in L. styraciflua and N. parvum, two databases with transcriptomic and genomic information were analyzed. For each organism, five sequences encoding CysRPs were identified (named as LsCysRP1-5 and NpCysRP1-5). All of these contained a putative start and stop codon, apart from the LsCysRP3 sequence, in which a stop codon was absent (

Table 1. Amino acid sequences of L. styraciflua and N. parvum CysRP. Predicted putative function was inferred with BLAST and Motif prediction tools, molecular weight, and isoelectric point are theoretical predictions (see Materials and Methods). Cysteines are underlined in the amino acid sequences.

Assigned Name	Target Name	Protein	Length	Total Number of Cysteines	Molecular Weight (kDa)	pI	Predicted Putative Function
LsCysRP1	comp30850_c0_seq1	MDRLSSMRSVFWILVSFGTNAAVSELINVACPQSAEVDHCMCNCRAIDDAWNVDWHNDLRMFGLKTEIQRPQNEVCCNLVTSCVFIHWDLVSHRKSRNLR	100	7	8.9	6.04	No significant similarity found
LsCysRP2	comp28855_c0_seq1	MGMALFCPVSGLQQQITAATSKFDGKGTSGVGACCTKPCPCQHQMKKLCCISSYGYCNRFRFKELSYHHLVAKANHSWLSCIEVIFAIPYSSADLGWLEMVKP	103	9	9.5	8.84	No significant similarity found
LsCysRP3	comp18453_c1_seq1	MASSGILNLACVVLVSMVLVAPHAEAAMTCGQVTSSVSPCLPYLKGASPLQPGCCNGIRSLNSAAKTTPDRQAACRCLQQAASSISGINLGLASGLPGKCGISI	104	8	7.7	9.03	Lipid-transfer protein type 1
LsCysRP4	comp37785_c0_seq1	MGFTGMSPMSFLCAWCCGQPTHHLEALVSWTAALLQSLKLKVAEDTDFFTNSFDPVEVWIISLYKYLYHKTLRKTNIGKEAALKRAKPMMIACVVEMLLIDLQS	104	4	8.9	8.8	No significant similarity found
LsCysRP5	comp31840_c0_seq3	MAIYKALLASLLVSLFVLHLVVADGMGMQTTDAATSSPPKSSPPKKIDCGGACSSRCQLSSRPNLCKRACGTCCARCNCVPPGTSGNEDVCPCYASMTTHGGRHKCP	107	12	8.6	8.67	Gibberellin-regulated protein 1-like
NpCysRP1	EOD50423.1	MQFKFALVSLFTALALAAPGIDLEKRCTANGGSCTQLSQCCSGNCEYDSSIGLVCKAASKKEKRCTANGGACTQLAQCCSGNCEYDSSIGLVCKP	95	12	7.8	7.57	Hypothetical protein UCRNP2_2803
NpCysRP2	EOD50922.1	MRPLYSHLVHTLPPLFSILLDPRPASAQDDHVPRCYYPDGTLASNDYACRLNTTESFCCTTNVTCLDNKICQVLAPTQYEYNRGTCTDKTWTSDECPKFCQAQSPSYGSGVIRPVY	116	9	10	4.79	Hypothetical protein UCRNP2_2308
NpCysRP3	EOD52838.1	MRFFVAISVAATALFSTGLAASHAKCACQIASDRGTNDAATAGMCSYVGGVMSKSNVAFEKQVATWSLTSILADNTITTGVLYPGKYCQGIGGHELGGDAVYEACRSQKCGDTYCQDSTCV	121	8	10.5	5.48	Hypothetical protein UCRNP2_338
NpCysRP4	EOD51105	MRVSAFAAILGYSSLALAQTQLLPTCAQKCVGTDFGGCSTVDVKCICANKELLTGLACCVSTGCNAADQETVINFAQNLCKPQGVTDLPTTATCASGASSATSSASSGTASTTASSAASASASATGSAASSASSAASSVASSAASAASSAAASASAATTAGSGAGSCQGNAAGMGFGLVLAGLLGAL	187	11	15.5	4.53	Putative CFEM domain protein
NpCysRP5	EOD44996.1	MKTSFAAVAFSLASVAYSQNISDLPSCSLNCFVTTLGSDGCSSLTDFECHCKVPGLTDEITPCVQKACSAADQQTVAEQVQALCAQYGVSISVPEASSTAAPSSTAAATSAPASSSAAATTAASSAASSAASSASASATSTETASPSATDVVISTPVATPTTPGASTPAGSTPTPSQFQGAAGKTGVVGGVVGFAAAVAAVAAL	204	8	17.6	4.02	Putative CFEM domain protein

and Table S2). The amino acid sequence length varied between 95 and 204, and the molecular weight ranged from 7.7 to 17.5 kDa. Interestingly, with the exception of LsCysRP1, which had a calculated isoelectric point (pI) of 6.04, the calculated pI values of all of the LsCysRPs exceeded 8.67. In contrast, apart from NpCysRP1 with a pI of 7.57, the calculated pI values of all the NpCysRPs were lower than 5.48 (Table 1). To predict the disulfide bonds in L. styraciflua and N. parvum CysRPs, the servers Cyscon, Disulfind, DiANNA, CYS_REC, and SCRATCH were used. The results revealed that LsCysRP1, LsCysRP2, LsCysRP3, LsCysRP4, and LsCysRP5 have the potential to form 3, 4, 4, 0, and 6 disulfide bonds, respectively. Meanwhile, in N. parvum, 6, 4, 4, 5, and 4 disulfide bridges were estimated for NpCysRP1, NpCysRP2, NpCysRP3, NpCysRP4, and NpCysRP5, respectively. The connectivity pattern of the cysteine residues forming the disulfide bonds varied between prediction methods (Table S3). To determine whether the CysRPs have the potential for secretion, an analysis was conducted with the TargerP-2.0 server. All sequences, apart from the LysCysRP2 sequence, have a peptide signal between 17 and 27 amino acids. To corroborate these results, additional analyses were performed using the Protter and DeepLoc-1.0 servers, and the results indicated that all of the CysRPs have a signal peptide and are extracellular proteins (

Table 2. Predicted cellular location of L. styraciflua and N. parvum CysRPs.

Assigned Name	TargetP-2.0 Server Prediction	Cleavage Site between Pos	Likelihood	Protter	Cleavage Site between Pos	DeepLoc-1.0 Prediction	Likelihood	Soluble/Membrane
LsCysP1	Signal peptide	24–25	0.9931	Signal peptide	21–22	Extracellular	0.8321	0.9813/0.0187
LsCysP2	Other	⎻⎻⎻⎻	0.7782	Signal peptide	18–19	Extracellular	0.8088	0.9946/0.0054
LsCysP3	Signal peptide	26–27	1	Signal peptide	26–27	Extracellular	0.9998	1/0
LsCysP4	Signal peptide	26–27	0.7135	Signal peptide	26–27	Extracellular	0.7316	0.9415/0.0585
LsCysP5	Signal peptide	23–24	1	Signal peptide	23–24	Extracellular	0.9973	1/0
NpCysP1	Signal peptide	17–18	1	Signal peptide	20–21	Extracellular	1	1/0
NpCysP2	Signal peptide	27–28	0.9962	Signal peptide	27–28	Extracellular	0.8773	0.9997/0.0003
NpCysP3	Signal peptide	20–21	1	Signal peptide	20–21	Extracellular	0.999	0.998/0.002
NpCysP4	Signal peptide	18–19	0.999	Signal peptide	18–19	cell membrane/extracellular/Endoplasmic reticulum	0.2925/0.2815 /0.2539	0.3246/0.6754
NpCysP5	Signal peptide	18–19	0.9986	Signal peptide	18–19	Extracellular/Endoplasmic reticulum/Cell membrane	0.3539/0.2639 /.2186	0.4355/0.5645

).

To identify possible functions and regions of similarity in the CysRPs, BLAST, and MOTIF tools were used. No significant similarity was found for LsCysRP1, 2, and 4, or for NpCysRP1, 2, and 3. In contrast, LsCysRP3 showed similarity with an LTP and LsCysRP5 with a gibberellin-regulated protein 1-like protein. Interestingly, a CFEM domain was identified for both NpCysRP4 and 5 (Table 1).

3.4. CysRP Phylogenetic Analyses

To obtain more information about the NpCysRPs, a phylogenetic analysis was conducted for each of these proteins (Figure 6 and Figure 7). All of the NpCysRPs were clearly grouped in the Botryosphaeria lineage, including Lasiodiplodia, Diplodia, and Macrophomina species. However, the species in the subclade were not always the same. NpCysRP1, 2 and 5 were closely related to Lasiodiplodia theobromae, while NpCysRP3 and 4 shared a branch with Macrophomina phaseolina.

The NpCysRP1 phylogeny revealed the existence of few orthologous sequences for this protein in databases, and this protein occurred in the Botryosphaeriaceae family, only in the species L. theobromae (85.00% identity), Diplodia corticola (85.42% identity) and Diplodia seriata (82.29%, identity), as well as in the family Cordycipitaceae of the order Hypocreales, characterized by entomopathogenic fungal species, such as Beauveria bassiana and Cordyceps confragosa, but with low identities (29.67% and 31.36%, respectively). However, the cysteines remained at the same site in the sequences (Figure S4). Other species, such as Aspergillus leporis and Rhizoctonia solani, were included in the alignment; however, these sequences had a lower number of cysteines and shared lower identity.

In addition to the Botryosphaeria lineage, the NpCysRP2 phylogenetic tree contained orthologs of the family Hypocreaceae, order Hypocreales, including seven different species of the Trichoderma genus with identities that ranged from 33.04 to 38.18%. There were also members of the family Glomerellaceae present, including species of the well-known phytopathogenic genus Colletotrichum, such as Colletotrichum asianum, Colletotrichum nymphaeae, and Colletotrichum orchidophillum, with 33.04, 36.28, and 37.19% amino acid identity, respectively. Interestingly, the tree also included the human pathogen Madurella mycetomatis of the order Sordariales, with 33.90% identity. The alignment revealed that NpCysRP2 introduces a new cysteine-rich domain with the consensus motif C1[Y/F]xPx9-10C2x6-8C3C4x4C5x2Nx2C6x10-23C7Tx8C9x3C10 at the N-terminus. Moreover, the multiple-sequence alignment of NpCysRP2 with the ortholog sequences of L. theobromae (67.52% identity) and D. corticola (62.07% identity) showed that NpCysRP2 is a larger protein, since the L. theobromae and D. corticola sequences contain 310 and 296 amino acids, respectively (Figure S4). The proteoforms of L. theobromae and D. corticola presented a transmembrane domain crossing at the carboxyl-end, while the predicted proteoforms for all of the sequences used in the phylogenetic analysis (26 in total) showed the transmembrane helix (Figure S5). A particular phylogeny was noted in the case of NpCysRP3 (Figure 6C); this protein did not present a well-defined clade distribution within the Botryosphaeriaceae family, and only the M. phaseolina orthologs (MpCysRPs A, B, and C) were found, with 64.46, 29.17 and 30.83% identity, respectively. The Nectriaceae family was represented in the phylogenetic tree with some members of the unique and fascinating Ambrosia Fusarium clade, represented by Fusarium euwallaceae and Fusarium kuroshium, which has been recognized recently as an emerging fungal pathogen [54,55]. NpCysRP5 has a CFEM domain characterized by eight cysteines with the specific consensus motif PxC¹[A/G]x₂C²x₈-12C³x_1-3[x/T]Dx_2-5C⁴xC⁵x_9-14C⁶x_3-4C⁷x_15-16C⁸ [56] and where P62 (in N. parvum sequence) was conserved in this group, in addition to the conserved amino acids of the CFEM motif (Figure S6). NpCysRP4 and its orthologs have a CFEM domain (we refer to this as a CFEM-like domain) with a conserved extra cysteine (C58 in the N. parvum sequence). Outside the CFEM domain, another cysteine (C94 in N. parvum) was conserved, forming the consensus motif PxC¹[A/G]x₂C²x_8-12C³x_1-3[x/T]Dx_2-5C⁴xC⁵x_8-13C⁶C⁷x_3-4C⁸x_15-16C⁹x_12-13C¹⁰. NpCysRP4 and NpCysRP5 are well represented in the clade of the Botryosphaeriaceae family, but NpCysRP4 shares a subclade with M. phaseolina, with 78.07% identity, while NpCysRP5 presents 83.01% identity with L. theobromae, the species Cenococcum geophilum and Glonium stellatum belong to the Gloniaceae family and present orthologs for both NpCysRP4 and NpCysRP5. Various orthologs were identified for NpCysRP4 in the Nectriaceae family, including in the order Hypocreales represented by the well-known phytopathogenic Fusarium species Fusarium verticillioides and Fusarium oxysporum, which were not found in the NpCysRP5 phylogenetic tree using 33 sequences. In the case of NpCysRP5, a clade represented by the order Eurotiales was clearly identified, grouping different Penicillium species with identities that ranged from 34.3 to 41.78%.

In the case of the L. styraciflua CysRPs, no similar sequences were found for LsCysRP1, 2 and 4, and it was therefore impossible to construct the corresponding phylogenetic trees. For LsCysRP3, the BLAST^® result revealed a sequence identity of 66.98–64.15% with various Gossypium species and 63.81% with Vitis pseudoreticulata. Finally, LsCysRP5 showed 73.83–65.42% identity with different Quercus species, 74.77% identity with Castanea mollissima, and 67.29% identity with Durio zibethinus.

3.5. Expression of CysRP mRNAs of L. styraciflua and N. parvum during Early Stages of the Infection Process

To explore whether the transcription of CysRP mRNAs of L. styraciflua and N. parvum was modulated at early stages of the interaction, quantitative polymerase chain reaction (qPCR) tests were performed (Figure 8). LsPR1, an ortholog of the Nicotiana tabacum gene encoding pathogenesis-related protein 1 (PR1), a protein involved in the defense response in plants and usually used as a defense marker. The qPCR results showed an increase in LsPR1 mRNA content at 1 and 3 dpi. The opposite profile was observed for all LsCysRP transcripts, since the expression decreased after 1 dpi, and LsCysRP2 presented the lowest level at this time post-infection. At 3 dpi, LsCysRP2 presented the most significant change, and LsCysRP3 remained unchanged, while LsCysRP1, 4, and 5 presented a mild increase (Figure 8a).

Finally, the experimental design allowed comparing the expression of NpCysRP transcripts between 1 and 3 dpi. There was an increase in the expression of NpCysRP1, 2, and 5 mRNAs, with NpCysRP5 presenting the most significant increase (Figure 8b). In contrast, NpCysRP4 showed a significant decrease.

4. Discussion

Liquidambar styraciflua L. (Altingiaceae) is used for reforestation and landscaping and has been proposed as an attractive hardwood species for potential bioenergy production [23]. Given its importance, it is important to determine the factors that affect the integrity of this tree, i.e., phytosanitary problems. It is generally assumed that L. styraciflua is associated with certain pathogens [57]. In this context, information collected by Hepting in 1971 [58] listed Cercospora liquidambaris, Septoria liquidambaris, Exosporium liquidambaris, Leptothyriella liquidambaris, and Gloesporium nervisequm as common foliar pathogens of the sweetgum. On the other hand, L. theobromae and Botryosphaeria dothidea have been identified as responsible for provoking stem cankers and dieback in seedlings located in nurseries and outplantings in the USA [57,58]. Currently, there is no new information about the nature of pathogenic fungi infecting L. styraciflua, but there is a prevalence of the incidence of symptoms associated with diseases. Here, we report that N. parvum is associated with foliar damage in L. styraciflua. The pathosystem established in this study identified Liqui 1−3 as a very aggressive strain, since it provokes several symptoms at early stages post infection (Figure 1 and Figure 2). The SEM analysis showed that the fungus was able to grow and develop pycnidia to produce a large number of mature spores, which acts to increase its infective potential (Figure 4). Our results are consistent with those reported for other Botryosphaeria species. Amponsah et al. [59] reported that, during the interaction of grapevine and N. luteum, conidial germination occurred at a faster rate (3 h after inoculation) on detached and wounded leaf and shoot surfaces than on attached and non-wounded leaf surfaces, indicating that conidium adhesion, germination and development were affected by the condition of the host plants. Moreover, in mamey sapote stem cuttings infected with L. theobromae at 30 dpi, it was possible to identify fruiting bodies embedded in the host tissue [60]. It will be interesting in future studies to characterize the Liqui 1−3 strain in greater detail, with respect to its capacity to produce lytic enzymes and secondary metabolites with phytotoxic effects in comparison to other strains isolated from other ecological niches. Finally, attention should also be given to other fungi isolated from L. styraciflua with pathogenic potential, such as Liqui 1–02 and Liqui 2–2 (Figure 1 and Figure S1).

In addition to the phenotypic effects associated with the pathogenicity caused by N. parvum in L. styraciflua, the identification of CysRPs was considered part of the molecular events triggered by both organisms during infection and defense response. Five CysRPs were identified for each organism (Table 1) and bioinformatic analyses revealed that LsCysRP1, 2, and 4 are proteins with unknown functions, suggesting that they are species-specific. LsCysRP3 was recognized as an LTP. This protein is considered an antifungal protein classified as PR-14 [61] with yet unknown mode of action, but in vitro conditions have the ability to enhance intermembrane exchange, causing a posteriori fungal cell death [62]. In barley leaves, LsCysRP3 inhibits the growth of Fusarium solani [63], while transgenic Populus tomentosa overexpressing an LTP of Leonurus japonicus is resistant to Alternaria alternata and Colletotrichum gloeosporioides [64]. LsCysRP5 was identified as a gibberellin-regulated protein 1-like protein. These proteins are named the GASA family in A. thaliana and Snakins in Solanum tuberosum. This class of proteins forms 5 or 6 disulfide bonds necessary for structure as a consequence of protein function [65]. Moreover, the overexpression of Snakin protein 1 in potato enhances resistance to relevant pathogens [66]. The Snakin-2 of French bean may form a 42-kDa protein complex with a proline-rich protein and could participate in the plant defense process [67]. Thus, in comparison with the reported orthologs of LsCysRP3 and 5, these L. styraciflua proteins may present antifungal activity. The expression patterns of the respective mRNAs of LsCysRP3 and 5 during the interaction with N. parvum are particularly interesting because of the repressive behavior profile of both mRNAs (Figure 8).

It is well known that many pathogens have evolved tools to evade the immune system of the host, so it will be interesting to investigate this effect further in future studies. Through genomic and transcriptomic analyses, it has already been recognized that the pathogenic and virulent protein arsenal of N. parvum potentially comprises enzymes that facilitate wood degradation and host colonization [68,69]. To the best of our knowledge, our research is the first to identify new CysRPs with potential roles in pathogenesis. Following examination of the genomic information of N. parvum, five sequences with high cysteine content were selected, and the analysis identified NpCysRP5 as a CFEM domain protein, distinguishable from other cysteine-rich domains and inherent to fungi [56]. Although this class of motifs is well represented in the Ascomycota phylum, pioneering research on the CFEM domain [56,70] did not include members of the Botryosphaeriaceae family. In this sense, NpCysRP5 is identified for the first time as an orthodox CFEM protein of N. parvum that is expressed early during the infection process in Liquidambar. Interestingly, NpCysRP4 (a protein that apparently lacks a transmembrane domain) has a CFEM-like domain with two additional well-conserved cysteines (Figure S6). The presence of these cysteines in NpCysRP4 could cause a disulfide rearrangement and increase conformational stability [71,72], and could increase the possibility of creating additional proteoforms with complementary functions, contributing to virulence. In accordance with the report by Zhang et al. [56], a positive correlation between CFEM domain occurrence and fungal pathogenicity was shown, and it seems that the occurrence of CFEM domain proteins is independent of the lifestyle of pathogen (i.e., biotrophic, hemibiotrophic, or necrotrophic). For example, in Botrytis cinerea (a necrotrophic fungus), BcCFEM1 is highly expressed at early stages of infection in Phaseolus vulgaris, and gene disruption acts to decrease virulence [73]. For Magnaporthe grisea, a hemibiotrophic filamentous ascomycete, the mutant pth11 (mutated in a CFEM transmembrane protein) is impaired during appressorium maturation, which influences infection capacity [74]. However, it is important to note that not all CFEM proteins play a role in pathogenicity. For example, three CFEM motif GPI-anchored proteins from Aspergillus fumigatus participate in cell wall stability but not in fungal virulence [75]. Here, we identified a CFEM motif-containing protein of N. parvum (NpCysRP5) that also contains a putative transmembrane helix, and is expressed during early infection stages in L. styraciflua. It will be of great interest to elucidate further the exact contribution of this protein in pathogenesis. The phylogenetic analyses identified orthologs of NpCysRP4 and NpCysRP5 in additional Botryosphaeriaceae species; thus, contributing to the knowledge of this important pathogen of woody plants (Figure 7). Moreover, the expression analysis indicated contrasting mRNA expression profiles for NpCysRP4 and NpCysRP5 (Figure 8). Since only the NpCysRP5 mRNA expression showed a positive correlation during early stages of infection, we believe that NpCysRP5 is a better candidate with a significant role during pathogenesis, highlighting the fact that NpCysRP4, with a CFEM-like domain, has another important function that could be identified in further study. NpCysRP2 also introduces a new cysteine-rich domain with the following characteristics: (1) the deduced consensus motif is C1[Y/F]xPx9-10C2x6-8C3C4x4C5x2Nx2C6x10-23C7Tx8C9x3C10; (2) all cysteines are conserved; and (3) a transmembrane helix toward the carboxyl terminus is also conserved. All of these data indicate that NpCysRP2 is a new protein, which we refer to as a fungal cysteine-rich transmembrane protein (FCRTP). Finally, the phylogenetic analyses highlighted two important points: (1) NpCysRPs are present in various fungal families, and (2) each NpCysRP is associated with a particular species. For example, for NpCysRP4, the order Eurotiales is well represented by several Penicillium species, while for NpCysRP5, the order Hypocreales is represented by members of the Fusarium genus, indicating that N. parvum is a very versatile fungus.

5. Conclusions

In this study, we showed that N. parvum is an aggressive pathogen of L. styraciflua. We also identified novel CysRPs of both interacting organisms, such as NpCysRP2, which seems to be conserved in various fungi and contains a novel cysteine-rich motif and a putative transmembrane domain. In addition, NpCysRP4 was identified as a novel CFEM-like domain. In the case of LsCysRPs, three of the identified proteins did not play a predicted putative function. Analysis of expression during the plant–pathogen interaction revealed that all of the LsCysRPs presented a repressive expression profile in comparison with that of NpCysRPs, suggesting the involvement of an interesting molecular mechanism during this interaction. Finally, all of the CysRPs identified in this study are suitable candidates for further investigation in order to increase the knowledge regarding the molecular processes that occur during the defense response of woody species to by N. parvum infection.