Mating-type genes of the anamorphic fungus Ulocladium botrytis affect both asexual sporulation and sexual reproduction

Abstract

Ulocladium was thought to be a strictly asexual genus of filamentous fungi. However, Ulocladium strains were shown to possess both MAT1-1-1 and MAT1-2-1 genes as observed in homothallic filamentous Ascomycetes. Here, we demonstrate that the U. botrytis MAT genes play essential roles for controlling asexual traits (conidial size and number). Using reciprocal genetic transformation, we demonstrate that MAT genes from the related heterothallic species Cochliobolus heterostrophus can also influence U. botrytis colony growth, conidial number and size, and have a strong effect on the range of the number of septa/conidium. Moreover, U. botrytis MAT genes can also affect similar aspects of asexual reproduction when expressed in C. heterostrophus. Heterologous complementation using C. heterostrophus MAT genes shows that they have lost the ability to regulate sexual reproduction in U. botrytis, under the conditions we employed, while the reciprocal heterologous complementation demonstrates that U. botrytis MAT genes have the ability to partially induce sexual reproduction in C. heterostrophus. Thus, the genetic backgrounds of C. heterostrophus and U. botrytis play significant roles in determining the function of MAT genes on sexual reproduction in these two fungi species. These data further support the role of MAT genes in controlling asexual growth in filamentous Ascomycetes but also confirm that heterothallic and homothallic Dothideomycete fungi can be interconverted by the exchange of MAT genes.

Introduction

Sexual reproduction in filamentous ascomycetes is controlled by a single regulatory mating-type locus referred to as the mating-type locus or MAT ^{1, 2}. The mating-type locus consists of two dissimilar DNA sequences in the mating partners, termed MAT1-1 and MAT1-2 idiomorphs³. MAT1-1 encodes a protein with an alpha-box (α-box) DNA-binding domain, whereas MAT1-2 encodes a protein with an HMG-box (high mobility group) DNA-binding motif. The α-box or HMG-box domain proteins specify two alternative transcription factors that permit each mating type to induce specific expression of many other genes required during and after mating, in particular, the genes that regulate pheromone precursors and pheromone receptors that are essential for cells of opposite mating types to attract each other and cause fertilization^{2, 4–6}.

Mating behavior in filamentous ascomycetes can be either homothallic (self-fertile) or heterothallic (self-sterile) in the same genus¹. Initiation of the sexual cycle is the step that mainly distinguishes heterothallic and homothallic species. The heterothallic species of filamentous ascomycetes are known to possess either one or the other idiomorph at the MAT1 locus. In contrast to the heterothallic species, homothallic species carry both MAT idiomorphs in a single genome, usually closely linked or fused^{2, 7}. Over the past decade, mating-type genes have been identified and characterized in an increasing number of filamentous ascomycetes, where their function as master regulators of sexual reproduction has been conserved^7–16. However, for approximately half of all filamentous ascomycetes species there is no known sexual state¹⁷. Presently, the question is whether these fungi, which only reproduce in a vegetative state, have abandoned sexual reproduction altogether. Alternatively, their sexual states could be small, inconspicuous, or only initiated under unusual conditions. Evidence suggests, following molecular investigation, that even the many putatively asexual filamentous ascomycetes species have genomes with MAT genes¹⁸, some of which are constitutively transcribed, providing appropriate evidence for sexual potential that is morphologically absent^19–25. Therefore, it is of great interest to determine whether the occurrence of MAT genes in an asexual species is a sign of realized mating, a relictual unused gene set, or a pathway that evolved to regulate another function.

Traditionally, the only way to determine whether any filamentous ascomycete species can reproduce sexually is by observation of their reproductive characteristics. Currently, the recent breakthroughs in the understanding of mating in ascomycetes following the cloning of mating-type genes in combination with genomics has made it possible to answer questions about the role of MAT genes in presumably asexual fungi. First, since the primary function of MAT genes is regulatory control of sex, their presence in asexual fungi can be presumed a necessary condition for sex to occur. Second, it can be assessed whether the MAT genes are properly expressed under controlled conditions and developmentally regulated in a manner consistent with sexual reproduction. Last, it can be tested whether mutations in MAT genes in asexual species occur more frequently and unpredictably than mutations in sexual species as a process of accumulation of mutations in the unused or neo-functionalized MAT genes.

The MAT genes have been well-studied in putatively asexual ascomycete species^{24, 26–28}. Ulocladium is genus of ascomycetes closely allied with the anamorphic (asexual) genera Alternaria, Embellisia, Nimbya and Stemphylium in the order Pleosporales (Dothideomycetes)²⁸. Ulocladium contains more than 29 species^29–33 and is closely allied with Alternaria and Stemphylium. Teleomorphs are known from several species in these two allied genera; their sexual states patterns are Alternaria/Lewia and Stemphylium/Pleospora, respectively^{34, 35}. However, no sexual state has yet been identified for Ulocladium. Ulocladium is therefore thought to be strictly asexual. In addition, most species within these five genera are only allied to asexual states. Alternaria is considered to be a largely asexual genus because most of the members have no known teleomorph yet are still known to carry expressed MAT genes in a heterothallic arrangement^{19, 24}. The genus Stemphylium is the anamorphic stage of the teleomorph Pleospora ^36–38. Some MAT loci of the Stemphylium species contain a single idiomorph (self-sterile), either MAT1-1 or MAT1-2, whereas others contain a unique fusion of MAT1-1 and MAT1-2 regions (self-fertile)^{2, 39}. However, the sexual state has not been identified in most species of Stemphylium. The MAT locus organization is unknown for most members of the genus Alternaria ³⁹.

Previously, we identified the full-length sequences of MAT1-1-1 and MAT1-2-1 genes for 26 Ulocladium species. Notably, both MAT1-1-1 and MAT1-2-1 genes were detected in the same haploid genome of all 26 Ulocladium species which appear to similar to that of Ophiocordyceps sinensis ⁴⁰, and thus all of the Ulocladium species have the potential to be homothallic³⁰. Transcriptional analysis on the basis of qRT-PCR showed that both MAT1-1-1 and MAT1-2-1 genes were expressed and may be functional in all 26 Ulocladium species, suggesting that all these Ulocladium species might have the potential to reproduce sexually³⁰.

In this study, we focused on the type species U. botrytis of Ulocladium ²⁹ and addressed the question of whether U. botrytis MAT1-1-1 or MAT1-2-1 genes lost the ability for sexual reproduction using genetic disruption and heterologous expression. In addition, we tested whether MAT genes influence asexual reproduction of Ulocladium species under natural conditions. Here, we first demonstrated that U. botrytis MAT1-1-1 and MAT1-2-1 play essential roles in colony growth and conidial size and number in U. botrytis using both separate MAT1-1-1 or MAT1-2-1 deletions and double deletions. Then, using heterologous expression, we showed that mating-type genes, regardless of whether they come from a heterothallic fungus (C. heterostrophus) or the anamorphic fungus (U. botrytis), regulate the expression of only asexual reproduction in the anamorphic fungus, whereas MAT genes from both the asexual and sexual species are capable of inducing sexual development when tested in the sexual species. This study provides insights into the functional role of MAT genes in asexual filamentous fungi where sexual reproduction is rare or absent and provides additional evidence that MAT genes may regulate important processes not directly related to sexual reproduction, i.e., asexual sporulation.

Results

Influence of U. botrytis MAT1-1-1 and MAT1-2-1 on vegetative growth and asexual sporulation

We have previously cloned and described the structural organization of MAT1-1-1 and MAT1-2-1 from the asexual species U. botrytis ³⁰. To test the functions of MAT1-1-1 and MAT1-2-1 genes in U. botrytis, we created MAT1-1-1 or MAT1-2-1, and MAT1-1-1/1-2-1 deletion mutants using the split-marker method. The single gene deletion mutants ΔmatUbMAT-1 and ΔmatUbMAT-2 and double mutants DmUbMAT-1 shown in Table S1 were confirmed by PCR (Fig. 1F), Northern blot (Fig. 1G) and Southern blot assays (Fig. 1H). The colony diameters of ΔmatUbMAT-1 and ΔmatUbMAT-2 (Fig. 1A,C: c,d) were very similar to those of WT and CK (Empty vector transformant) (Fig. 1A,C: a,b), whereas the colony diameters of DmUbMAT-1 (Fig. 1A,C: e) were slightly smaller than those of the two controls and either of the two single mutants (P < 0.05). The colony borders of these three mutants (Fig. 1A: c,d,e) were loose in contrast to WT and CK (Fig. 1A: a,b), and DmUbMAT-1 showed significant incompactness around the colony borders (Fig. 1A: e). MAT expression levels influenced the size of the conidia (Fig. 1B,D), and the conidial sizes of ΔmatUbMAT-1 (17 × 14 μm²) and ΔmatUbMAT-2 (16 × 13 μm²) were slightly smaller than those of WT (20 × 17 μm²) and CK (19 × 16 μm²) (P < 0.05). The conidial sizes of DmUbMAT-1 (13 × 11 μm²) were significantly smaller than those of conidia formed by WT, CK and either of the two single mutants (P < 0.01). MAT expression levels also were correlated with the total number of conidia produced by three different mutants compared with two controls (Fig. 1B,E). The two single mutants ΔmatUbMAT-1 and ΔmatUbMAT-2 produced fewer conidia than did WT and CK (P < 0.05; Wilcoxon rank-sum test), but DmUbMAT-1 produced the least conidia. These data demonstrate that MAT1-1-1 and MAT1-2-1 play roles in both colony growth and conidial size and number in U. botrytis.

Figure 1.

Effect of deletion of the U. botrytis MAT genes on colony morphology, conidial size and number. (A,C) Growth and diameter of the colonies of different mutants at 12 days after incubation. Colony growth rates were determined from at least 25 plates. (B,D) Variation in conidial size. L = Length, W = Width. The average size of conidia was determined from at least 50 conidia. Photographs were taken 12 days after incubation. (B,E) Variation in conidial number. The number of conidia produced per plate from cultures grown on PCA plates for 12 days under standard conditions. Error bars represent standard errors calculated using three replicates for each sample. ‘*’ indicates a significant difference from WT (P < 0.05) using Student’s t-test. ‘**’ indicates a significant difference from WT (P < 0.01) using Student’s t-test. (F) PCR analysis of the transcription of the MAT genes in different deletion lines. D-DNA template of WT. (G) Northern blot analysis. Twenty micrograms of total RNA, isolated from WT, CK, and all mutant strains, were loaded per lane. The Northern blot was probed using MAT1-1-1 and MAT1-2-1 gene-specific probes. A 5.8S rRNA-specific probe was used as positive control. (H) For Southern blot analysis, both hygB and G418 specific probes were used to detect transgene insertion. WT and CK have no hygB and G418 specific insertion. a WT (Wild-type U. botrytis). b CK is an empty vector transformant. c ΔmatUbMAT-1, G418 was used to detect transgene insertion. d. ΔmatUbMAT-2, hygB was used to detect transgene insertion. e DmUbMAT-1, hygB and G418 were individually used to detect transgene insertion. Each experiment was repeated at least three independently times.

Heterothallic C. heterostrophus MAT1-1-1 and MAT1-2-1 also influence vegetative growth and asexual sporulation in U. botrytis

To test the functions of the C. heterostrophus MAT1-1-1 and MAT1-2-1 in the stable U. botrytis deletion mutants ΔmatUbMAT-1, ΔmatUbMAT-2, and DmUbMAT-1, we created transformants ΔmatUbMAT-1{ChMAT}, ΔmatUbMAT-2{ChMAT}, DmUbMAT-1{ChMAT}-1, DmUbMAT-1{ChMAT}-2, and DmUbMAT-1{ChMAT}-3 using previously described methods. Each of the U. botrytis MAT deletion mutants was transformed with the corresponding gene from C. heterostrophus, i.e., ΔmatUbMAT-1{ChMAT} contains the MAT1-1-1 gene of C. heterostrophus. Each of these transgenes conferred to the MAT deletion mutants of U. botrytis the same phenotypes of colony growth, conidial number and size, and compartmentalization (data not shown). Thus, we only analyzed the five typical transformants ΔmatUbMAT-1{ChMAT}-1, ΔmatUbMAT-2{ChMAT}-1, DmUbMAT-1{ChMAT}-1-1, DmUbMAT-1{ChMAT}-2-1, and DmUbMAT-1{ChMAT}-3-1 in subsequent experiments (Table S1). The PCR, Southern blot, and qRT-PCR analyses of these five typical transformants are shown in Fig. 2F,G,H.

Figure 2.

Effect of heterologous expression of C. heterostrophus MAT genes on asexual morphology in U. botrytis strains of MAT deletion lines. (A,C) Growth and diameter of colonies across the different transgenic lines at 12 days after incubation. Colony growth rates were determined from at least 25 plates. (B,D) Variation in conidial size. L = Length, W = Width. The average size of conidia were determined from at least 50 conidia. Photographs were taken at 12 days after incubation. (B,E) Variation in conidial number. Number of conidia produced per plate from cultures grown on PCA plates for 12 days under standard conditions. (F) PCR analysis of MAT gene transcription in different transgenic lines. D-DNA template of WT. (G) qRT-PCR analysis of mRNA expression levels of MAT1-1-1 and MAT1-2-1 in individual heterologous transgenic lines, relative to the constitutive control genes. WT and CK were used as negative controls. Actin gene was used as the reference gene. Error bars represent standard errors calculated using three biological replicates for each sample. ‘*’ indicates a significant difference from WT (P < 0.05) using a Student’s t-test. ‘**’ indicates a significant difference from WT (P < 0.01) using a Student’s t-test. (H) For Southern blot analysis, both hygB and G418 specific probes were used to detect transgene insertion as shown in Table S1. WT and CK have no hygB and G418 specific insertion. a WT (Wild-type U. botrytis), b CK is an empty vector transformant. c ΔmatUbMAT-1{ChMAT}-1, d. ΔmatUbMAT-2{ChMAT}-1, e DmUbMAT-1{ChMAT}-1-1, f. DmUbMAT-1{ChMAT}-2-1, g DmUbMAT-1{ChMAT}-3-1. Each experiment was repeated at least three times.

The colony diameters of these five transformants (Fig. 2A,C: c,d,e,f,g) were very similar to WT and CK (Fig. 2A,C: a,b). The colony borders of ΔmatUbMAT-1{ChMAT}-1 and ΔmatUbMAT-2{ChMAT}-1 (Fig. 2A: c,d) are incompact in contrast to WT and CK (Fig. 2A: a,b) and three other typical transformants in the DmUbMAT-1 background (Fig. 2A: e,f,g). Notably, no significant differences in the colony borders between ΔmatUbMAT-1{ChMAT}-1 and ΔmatUbMAT-2{ChMAT}-1 (Fig. 2A: c,d) and the two single mutants ΔmatUbMAT-1 and ΔmatUbMAT-2 (Fig. 1A: c,d) were found under standard conditions. As shown in Fig. 2A: e,f,g, the cultures of these three DmUbMAT-1 transformants became loose and ringed in a slight gray color while the single gene deletion background transformants (Fig. 2A: c,d) and either of the two controls (Fig. 2A: a,b) were very compact and pigmented in a constant dark color.

C. heterostrophus MAT gene heterologous expression in asexual U. botrytis can affect the variation of conidial sizes and number in these different transformants. As shown in Fig. 2B,D,E (c,d), the conidial sizes and numbers of the ΔmatUbMAT-1{ChMAT}-1 and ΔmatUbMAT-2 {ChMAT}-1 were very similar to those of WT and CK (Fig. 2B,D,E: a,b). The conidial sizes of DmUbMAT-1{ChMAT}-1-1 (30 × 22 μm²) and DmUbMAT-1{ChMAT}-2-1 (29 × 23 μm²) (Fig. 2B,D: e,f) were significantly larger than those of WT (20 × 17 μm²) and CK (22 × 16 μm²) (Fig. 2B,D: a,b) and of the two other transformants ΔmatUbMAT-1{ChMAT}-1 (19 × 17 μm²) and ΔmatUbMAT-2 {ChMAT}-1 (20 × 16 m²) (P < 0.01) (Fig. 2B,D: c,d). The conidial sizes of the DmUbMAT-1{ChMAT}-3-1 (23 × 19 μm²) (Fig. 2B,D: g) were slightly larger than those of the two controls and ΔmatUbMAT-1{ChMAT}-1 and ΔmatUbMAT-2 {ChMAT}-1 (P < 0.05) (Fig. 2B,D: a,b,c,d) but were also significantly smaller than those of DmUbMAT-1{ChMAT}-1-1 and DmUbMAT-1{ChMAT}-2-1 (P < 0.01) (Fig. 2B,D: e,f). On the other hand, the number of conidia produced by DmUbMAT-1{ChMAT}-1-1, DmUbMAT-1{ChMAT}-2-1, and DmUbMAT-1{ChMAT}-3-1 (Fig. 2B,E: e,f,g) were significantly fewer than those of WT, CK, and of the two single gene deletion backgrounds with transgenes (Fig. 2B,E: a,b,c,d) (P < 0.01). The range in the number of septa/conidium was 0-1 within CK, WT, ΔmatUbMAT-1{ChMAT}-1 and ΔmatUbMAT-2{ChMAT}-1 (Fig. 2B: a,b,c,d), whereas DmUbMAT-1{ChMAT}-1-1 and DmUbMAT-1{ChMAT}-2-1 had 1-4 septa/conidium and most had 2-3 (Fig. 2B: e,f). However, the mature conidia of DmUbMAT-1{ChMAT}-3-1 (Fig. 2B: g) was restored to 0-1 septa/conidium as in the WT and CK and became more darkly pigmented and distinctly different from the two controls and each of the four other transformants (Fig. 2B: a,b,c,d,e,f). These results indicated that the C. heterostrophus MAT1-1-1 and MAT1-2-1 transgenes could regulate similar asexual reproduction traits as observed for U. botrytis MAT genes.

Expression of U. botrytis MAT1-1-1 and MAT1-2-1 in C. heterostrophus influences vegetative growth and asexual sporulation

To determine whether U. botrytis MAT1-1-1 and MAT1-2-1 are involved in controlling colony growth and size and number of conidia in C. heterostrophus, three transformants were created and were used for subsequent analyses, including ChΔmat0{UbMAT}-2, ChΔmat0{UbMAT}-3 and ChΔmat0 {UbMAT}-4 (Table S1). The genetic composition of these three transformants were confirmed by PCR, Southern blot and qRT-PCR (Fig. 3E,F,G: e,f,g) and compared to C. heterostrophus (2847), C. heterostrophus C4-41.7 (MAT-0), C. heterostrophus C5 (2829) and C. heterostrophus C4 (2849), which served as controls (Fig. 3E,F,G: a,b,c,d). The cultures of the three heterologous transformants (Fig. 3A,D: e,f,g) were often very compact and darkly pigmented in a constant manner with slightly small diameters compared with each of the four controls (Fig. 3A,D: a,b,c,d). The conidial sizes of ChΔmat0 {UbMAT}-2 (110 × 15 μm²) and ChΔmat0 {UbMAT}-3 (117 × 14 μm²) (Fig. 3C: e,f) were nearly the same as those of C. heterostrophus C5 (108 × 16 μm²) and C. heterostrophus C4 (112 × 14 μm²) (Fig. 3C: c,d). The conidia produced by C. heterostrophus C4-41.7 were the smallest in size (89 × 12 μm²) of all the untransformed strains (Fig. 3C: b). However, the conidial sizes of ChΔmat0 {UbMAT}-4 (124 × 17 μm²) (Fig. 3C: g) were the largest and most similar to that of C. heterostrophus (2847) (120 × 19 μm²) (Fig. 3C: a). No clear differences in the number of conidia produced by these three transgenic strains and four controls were found under standard conditions (data not shown). For C. heterostrophus C4-41.7, the range in number of septa/conidium was 1–6, with a mean of 3–4 (Fig. 3B: b). Notably, ChΔmat0{UbMAT}-2, ChΔmat0{UbMAT}-3, and C. heterostrophus C5/C4 had 3–9 septa/conidium and most had 5-7 (Fig. 3B: c,d,e,f). In addition, ChΔmat0{UbMAT}-4 was nearly restored to the wild type strain C. heterostrophus (2847) that had 7–12 septa/conidium and most had 7–9 (Fig. 3B: a,g). Therefore, we concluded that the U. botrytis MAT1-1-1 and MAT1-2-1 genes could also affect asexual reproduction in C. heterostrophus.

Figure 3.

Effect of transformation of U. botrytis MAT genes on asexual morphology of C. heterostrophus C4-41.7 (MAT-0). (A,D) Growth and diameter of the colonies of the different strains at 12 days after incubation. Colony growth rates were determined from at least 25 plates. (B,C) Variation in conidial size. L = Length, W = Width. The average size of conidia was determined from at least 50 conidia. Photographs were taken 12 days after incubation. (E) RT-PCR analysis of the transcription of MAT genes in different transgenic lines. D-DNA template of WT. (F) qRT-PCR analysis of mRNA expression levels of MAT1-1-1 and MAT1-2-1 in individual heterologous transgenic lines as described above, relative to the constitutive control lines. C. heterostrophus strains 2847, C4-41.7 (MAT₀), 2829 and 2849 were used as negative controls. Actin was used as a positive control. Error bars represent standard errors calculated using three biological replicates for each sample. ‘*’ indicates a significant difference from WT (P < 0.05) using a Student’s t-test. ‘**’ indicates a significant difference from WT (P < 0.01) using a Student’s t-test. (G) For Southern blot analysis, both hygB and G418 specific probes were used to detect transgene insertion as shown in Table S1. WT and WT1 have no hygB and G418 specific insertion. a WT is C. heterostrophus (2847). b WT1 is C. heterostrophus C4-41.7 (MAT-0). c WT2 is C. heterostrophus C5 (2829). d WT3 is C. heterostrophus C4 (2849). e ChΔmat0 {UbMAT}-2. f ChΔmat0 {UbMAT}-3. g ChΔmat0 {UbMAT}-4. Each experiment was repeated at least three times.

Effect of C. heterostrophus MAT1-1-1 and MAT1-2-1 genes on sexual reproduction in the anamorphic U. botrytis

To confirm whether a mating phenotype of the asexual U. botrytis was conferred by C. heterostrophus MAT transgenes, we conducted cross mating using DmUbMAT-1 {ChMAT}-1-1× DmUbMAT-1{ChMAT}-2-1 strains that carried compatible C. heterostrophus MAT genes and three tests of self-fertilization of strains with gene combinations expected to confer self-compatibility, including DmUbMAT-1{ChMAT}-3-1, ΔmatUbMAT-1{ChMAT}-1, and ΔmatUbMAT-2{ChMAT}-1 (Table S2). The DmUbMAT-1{ChMAT}-3-1 strain was transformed with both the C. heterostrophus MAT1-1-1 and MAT1-2-1 genes. The ΔmatUbMAT-1{ChMAT}-1 strain contained U. botrytis MAT1-2-1 transformed with C. heterostrophus MAT1-1-1. The ΔmatUbMAT-2{ChMAT}-1 strain contained U. botrytis MAT1-1-1 transformed with C. heterostrophus MAT1-2-1. DmUbMAT-1{ChMAT}-1-1× DmUbMAT-1{ChMAT}-2-1 did not produce pseudothecia or asci after incubating on the surface of corn leaf substrates (Fig. 4A: a, Table S2), and these results were consistent with the three self matings which were also sterile (Fig. 4A: b,c,d, Table S2). Moreover, cross mating of ΔmatUbMAT-1× ΔmatUbMAT-2 (Fig. 4A: e, Table S2) and self-mating of U. botrytis (Fig. 4A: f, Table S2) did not produce pigmented pseudothecia and asci on the surface of corn leaf substrates. In contrast, the number of pseudothecia per square centimeter and the number of asci per pseudothecium were much greater in cross matings of C. heterostrophus C5 × C. heterostrophus C4 (W1) (Fig. 4A: g and g₁, Fig. 4B,C: g, Table S2) and were indistinguishable from those produced in self mating of C. heterostrophus (W2) (Fig. 4A: h and h₁, Fig. 4B,C: h, Table S2). These data demonstrate that although the C. heterostrophus MAT genes are expressed in the U. botrytis transgenic strains (Fig. 2F,G), the C. heterostrophus MAT genes can not regulate sexual reproduction in the genetic background of the anamorphic U. botrytis strains.

Figure 4.

Effect of transformation of C. heterostrophus MAT genes on pseudothecia and asci formation in U. botrytis. (A) Pseudothecia formation was tested in different crosses or self matings on the surface of corn leaf substrates. (B) Average number of pseudothecia per square centimeter on the surface of the corn leaf. Error bars indicate 95% confidence intervals. No significant differences were observed in the number of pseudothecia between W1 and W2 (P > 0.05). (C) Average number of asci per pseudothecium. At least 10 pigmented pseudothecia were opened and the number of asci in each pseudothecium were recorded. Error bars indicate 95% confidence intervals. No significant differences were observed in the number of asci per pseudothecium between WT1 and WT2 (P > 0.05). a Cross-mating pattern DmUbMAT-1{ChMAT}-1-1× DmUbMAT-1{ChMAT}-2-1. b Self-mating pattern DmUbMAT-1{ChMAT}-3-1. c Self-mating pattern ΔmatUbMAT-1{ChMAT}-1. d Self-mating pattern ΔmatUbMAT-2{ChMAT}-1. e Cross-mating pattern ΔmatUbMAT-1× ΔmatUbMAT-2. f Self-mating pattern U. botrytis strain. All these crosses or self matings were completely sterile–no pseudothecia and asci were produced on the surface of corn leaf substrates. g and g₁ Cross mating of C. heterostrophus C5× C. heterostrophus C4 (W1). h and h₁ Self mating of C. heterostrophus (2847) (W2). Following W1 and W2 crosses, pseudothecia and asci were produced on the surface of corn leaf substrates.

Effect of U. botrytis MAT1-1-1 and MAT1-2-1 genes on sexual reproduction in C. heterostrophus

To test whether a mating phenotype was conferred by U. botrytis MAT transgenes expressed in the heterothallic C. heterostrophus, we conducted three cross matings ChΔmat0 {UbMAT}-2 × ChΔmat0{UbMAT}-3, C. heterostrophus C5× ChΔmat0{UbMAT}-3 and C. heterostrophus C4× ChΔmat0{UbMAT}-2 (Table S2) and one test of self-fertilization ChΔmat0{UbMAT}-4 (Table S2). As a result, numerous and tiny pigmented pseudothecia were produced by a cross mating ChΔmat0{UbMAT}-2× ChΔmat0{UbMAT}-3 that were very similar to those of a self mating of ChΔmat0{UbMAT}-4 (P > 0.05) on the surface of corn leaf substrates (Fig. 5A,B: a,b, Table S2). Note that ChΔmat0{UbMAT}-2 and ChΔmat0{UbMAT}-3 contain U. botrytis MAT1-1-1 or MAT1-2-1, while ChΔmat0{UbMAT}-4 contain U. botrytis MAT1-1-1 and MAT1-2-1 (Table S2). In addition, two other cross matings, C. heterostrophus C5× ChΔmat0{UbMAT}-3 and C. heterostrophus C4× ChΔmat0{UbMAT}-2, produced almost the same numerous and slightly larger pigmented pseudothecia on the surface of corn leaf substrates (P > 0.05) (Fig. 5A,B: c,d, Table S2). For the two other cross matings, half were crossed to a transgenic strain carrying C. heterostrophus MAT1-1-1 or MAT1-2-1, and half were crossed to a transgenic strain carrying U. botrytis MAT1-1-1 or MAT1-2-1. As shown in Fig. 5A,B (a,b,c,d,e,f), the number and the sizes of the pigmented pseudothecia were gradually increased or enlarged on the surface of corn leaf substrates, respectively. Interestingly, no asci were noted when all the pseudothecia from self or cross mating strains were examined (Table S2, Fig. 5A: a₁,b₁,c₁,d₁) compared with those of WT1 (Fig. 4A: g,g₁, Fig. 4B,C: g. Table S2) and WT2 (Fig. 4A: h and h₁, Fig. 4B,C: h, Table S2). These data demonstrate that the heterologous U. botrytis MAT genes are not only strongly expressed in the C. heterostrophus transgenic strains but also have the ability to induce a sexual mode of reproduction in the genetic background of the heterothallic C. heterostrophus strains.

Figure 5.

Effect of transformation of U. botrytis MAT genes on pseudothecia and asci formation in C. heterostrophus. (A) Pseudothecia formation in different cross or self matings on the surface of corn leaf substrates. (B) Average number of pseudothecia per square centimeter on the area of the corn leaf. Error bars indicate 95% confidence intervals. No significant differences were observed in the number of pseudothecia between W1 and W2 (P > 0.05). a and a₁ Cross mating of ChΔmat0 {UbMAT}-2× ChΔmat0 {UbMAT}-3. b and b₁ Self mating of ChΔmat0 {UbMAT}-4. A few, tiny pigmented pseudothecia were discovered in a and a₁ or b and b₁. c and c₁ Cross mating of C. heterostrophus C5× ChΔmat0{UbMAT}-3. d and d₁ cross mating of C. heterostrophus C4× ChΔmat0 {UbMAT}-2. A medium number of slightly large pigmented pseudothecia were discovered in c and c₁ or d and d₁. e and e₁ Cross mating of C. heterostrophus C5× C. heterostrophus C4 (W1). f and f₁ Self mating of C. heterostrophus (2847) (W2). The maximum number of the largest pigmented pseudothecia were discovered in e and e1 or f and f₁.

Discussion

Fungi are a group historically considered to present a high proportion of asexual species; a fifth of the species were once thought to exclusively reproduce asexually⁴¹. For example, most species in the filamentous ascomycetes genera Alternaria, Stemphylium, and Ulocladium are only known to reproduce asexually^{29, 30}. Possible reasons for the absence of sex are that the suitable factors and conditions needed to induce sex have not been determined and further research needs to be conducted to identify suitable environmental conditions for sex, or that these fungal genomes lack the equipment to engage in sex^{42, 43}. However, MAT genes have also been cloned and characterized from putatively asexual fungi and have been shown to be functional when expressed in closely related sexual species^{19, 24}, even when they have not been demonstrated to function in the asexual progenitor. In asexual fungi, the functions of mating-type genes have proven particularly useful in molecular phylogenetic studies^{24, 44, 45}. Our previous study demonstrated that the MAT genes are suitable for phylogenetic analysis for the four closely allied genera Ulocladium, Alternaria, Cochliobolus, and Stemphylium ³⁰ and support a similar functional role in all four asexual genera. In this study, our experiments have demonstrated that U. botrytis MAT1-1-1 and MAT1-2-1 could influence colony growth and conidia size and number when deleted in U. botrytis or when expressed in C. heterostrophus (Figs 1 and 3) and that C. heterostrophus MAT1-1-1 and MAT1-2-1 could also exert similar effects when expressed in U. botrytis (Fig. 2). Thus, the mating-type genes in these two closely related fungi are functional and influence both sexual and asexual characteristics. The presence of mating-type genes in both taxa with and without a known sexual stage allow these genetic characters to be integrated across both anamorphs and teleomorphs and are particularly useful for consolidating the taxonomy of these two groups⁴⁶.

Several putatively asexual species have been previously reported to contain functional, constitutively transcribed MAT genes^{19, 36, 39}. Among these species are plant pathogens such as A. alternata, S. herbarum, S. triglochinicola and S. eturmiunm, as well as biotechnologically relevant anamorphic fungi, including Aspergillus fumigatus, and Penicillium marneffei ^{24, 47, 48}. Analyses of the MAT gene sequences of these asexual fungi revealed the presence of transcriptionally active MAT genes which are normally associated with sexual reproduction^{47, 48}. These reports indicate that the absence of detectable sexual reproduction in the asexual filamentous ascomycetes is not due to the lack of mating-type genes nor is it due to the occurrence of disruptive mutations within MAT genes or other sex-related genes. Thus, sexual reproduction in the filamentous ascomycetes is universally genetically controlled by a sex-specific region referred to as the mating-type locus^{1, 2}. Our previous study demonstrated that all Ulocladium species usually carry both MAT1-1-1 and MAT1-2-1 in a single genome, which provides further evidence supporting that all Ulocladium species may have the potential to reproduce sexually during the life cycle³⁰. However, no sexual state has yet been identified for Ulocladium, which is therefore thought to be a strictly asexual filamentous ascomycete genus. It is possible that the MAT genes within Ulocladium species can not effectively regulate sexual reproduction. The U. botrytis MAT1-1-1 and MAT1-2-1 sequences are homologous to MAT-1-1-1 and MAT1-2-1 of the related heterothallic species C. heterostrophus. The coding sequences of the α-box domain of both MAT1-1-1 genes (Fig. S1) and HMG-box domain of both MAT1-2-1 genes (Fig. S2), apart from their 47 or 43 nonhomologous sequences, are 71.43% or 72.44% identical, respectively. The U. botrytis MAT1-1-1 and MAT1-2-1 sequences are thus lowly similar to those of C. heterostrophus MAT-1-1 and MAT1-2-1, respectively. When either of the C. heterostrophus MAT1-1-1 or MAT1-2-1 genes were transformed into U. botrytis, the recipient could neither self nor cross with other U. botrytis strains, in contrast to wild type C. heterostrophus strains and transgenic C. heterostrophus strains which can do both. Notably, all the mating patterns of the transgenic U. botrytis strains containing U. botrytis genes did not induce sexual reproduction (Table S2, Fig. 4A). On the other hand, introduction of the U. botrytis MAT1-1-1 and MAT1-2-1 into the C. heterostrophus C4, C5, and C4-41.7 strains induced either by self-mating or cross-mating a varying degree sexual reproduction (Table S2, Fig. 5A), suggesting the U. botrytis MAT genes have not lost the ability for initiating sexual reproduction. Thus, the lack of sexual reproduction in U. botrytis is not due to either absence or mutation of MAT genes, as was observed for A. alternata and B. sacchari ²⁴, nor is it due to the low similarity of the MAT1-1-1 and MAT1-2-1 sequences between U. botrytis and C. heterostrophus (Figs S1 and S2). We hypothesize that there are multiple possible reasons that U. botrytis MAT genes are not triggering sexual reproduction in the laboratory conditions tested. First, MAT genes encode transcriptional regulators that normally control the expression of many genes required for sexual reproduction, including the mating pheromones and their G-protein–coupled receptors⁴⁹, and these MAT-regulated genes may have evolved to not control sexual reproduction in U. botrytis. Alternatively, the genetic background of U. botrytis may restrict the roles of MAT genes in sexual reproduction to environmental conditions not tested here. However, another explanation is that U. botrytis may have a cryptic sexual cycle similar to the human pathogen Coccidioides immitis ⁵⁰, but sexual reproduction may be a rare event that is hard to detect as it was for the presumed asexual barley pathogen Septoria passerinii ⁵¹ and thus remains to be described.

Mating-type genes have been characterized in a number of heterothallic and homothallic filamentous ascomycetes, where they function as master regulators of sexual reproduction⁵². MAT genes govern both the ability of a strain to undergo sexual reproduction but are also critical in the evolution of heterothallic and homothallic modes of mating by exchange or rearrangement of MAT genes^52–55. In this study, we addressed the function of MAT genes of U. botrytis by expressing heterothallic C. heterostrophus MAT1–1–1 or MAT1-2-1 genes in single or a double MAT-deleted U. botrytis strains and evaluating if the C. heterostrophus MAT genes could promote sexual reproduction in U. botrytis. Unexpectedly, our results demonstrate that both ChMAT1-1-1 and ChMAT1-2-1 could not trigger sexual reproduction in all transgenic U. botrytis strains despite the multiple tests of different mating specificity (Fig. 4A,B: a,b,c,d,e, Table S2), as observed in the wildtype U. botrytis strain (Fig. 4A,B: f, Table S2). However, the MAT genes of both U. botrytis and C. heterostrophus were shown to be able to influence asexual characteristics in both species. These observations are consistent with studies showing that expression of genes during asexual growth is also dependent on MAT, such as in isogenic Neurospora crassa and Aspergillus oryzae strains^{4, 14}. MAT gene regulation of diverse functions has been observed in asexual fungi such as Fusarium graminearum ⁵⁶, Penicillium chrysogenum ⁵⁷ and in sexual fungi Podospora anserina ⁵⁸, Sordaria macrospora ⁵⁹ and Neurospora crassa ¹⁸, including metabolism, cell wall organization, cellular response to stimuli, cell adhesion, fertilization, information pathways, transport, and developmental processes. A broader understanding that MAT genes pleiotropically control both asexual and sexual reproduction is provided by these studies and our study on U. botrytis. For these reasons, the function of MAT genes in fungi with no known sexual cycle needs to be carefully scrutinized before concluding that they promote outcrossing and meiotic reproduction.

In all C. heterostrophus transgenic strains, the heterothallic transgenic ChΔmat0{UbMAT}-4 strain was changed to homothallic when U. botrytis MAT1-1-1 and MAT1-2-1 were co-introduced into the C4-41.7 (MAT₀) strain, but all other C. heterostrophus transgenic strains still mated in a heterothallic manner, including crosses between ChΔmat0 strains carrying complementary U. botrytis MAT genes (Table S2). Thus, all C. heterostrophus transgenic strains were able to cross in a heterothallic manner or self in a homothallic manner using the U. botrytis genes, although the phenotypes were different from those of the genetic background of C. heterostrophus. Specifically, all these self and cross phenotypes were able to produce fewer and smaller pseudothecia (Fig. 5A: a,b,c,d, Table S2) but were not able to produce asci compared to those of wild type C. heterostrophus crosses (Fig. 5A: e,f, Table S2). These observations suggest that partial characteristics of sexual reproduction in these C. heterostrophus transgenic strains are attributable to the introduction of U. botrytis MAT genes into the genetic background of the heterothallic C. heterostrophus. Thus, these results suggest that the genetic backgrounds of the C. heterostrophus and U. botrytis strains may play significant roles in determining the potential effect of MAT genes on sexual reproduction in heterothallic and homothallic strains. In summary, this study reveals that U. botrytis MAT1-1-1 and MAT1-2-1 may have not lost the ability for sexual reproduction in this species which has only been observed reproduce asexually and that the MAT genes play a major role in controlling asexual characteristics.

Methods

Strains, culture conditions, and crosses

The U. botrytis strain²⁹ (CBS 198.67) (MAT1-1-1: KF533878, MAT1-2-1: KF533888)³⁰ was grown on potato carrot agar (PCA) under standard conditions³³. Some test strains, including C. heterostrophus strains C5 (ATCC48332) only containing MAT1-1-1 (X68399), C4 (ATCC48331) only containing MAT1-2-1 (X68398), C. heterostrophus strain 2847 carrying MAT1-2-1/1-1-1, and a double mat-deleted C4-41.7 (MAT₀) strain, were obtained from O. C. Yoder and B. G. Turgeon of Cornell University (Ithaca, NY, U.S.A). Note that the C4-41.7 strain is derived from C4 that lacks the whole mating-type locus⁶⁰. These test strains were cultured on complete medium with xylose (CMX)¹¹ and incubated under 16 h light/8 h dark at approximately 22 °C for 12 days. In this study, selfing or crossing of U. botrytis, C. heterostrophus and all transgenic strains were performed using procedures previously described for C. heterostrophus ^{32, 61}.

Amino acid alignment and phylogenetic analysis

Assembled U. botrytis MAT1-1-1 and MAT1-2-1 sequences were aligned with MAT1-1-1 and MAT1-2-1 sequences from C. heterostrophus (X68399, X68398, respectively), A. alternata (AB009451, AB009452, respectively) and S. eturmiunum (EGS29-099, EGS29-099, respectively). Assembled sequences were analyzed for putative open reading frames and introns using Genetyx Mac v.11.2 software (Genetyx, Shibuya, Tokyo, Japan). Putative introns were spliced from the open reading frames, conceptually translated using Jellyfish software (Lab Velocity, San Francisco, CA), and aligned in ClustalX BLAST⁶² searches for similar nucleotide and protein sequences were carried out against the National Center for Biotechnology Information (NCBI) databases.

Deletion of MAT1-1-1 and MAT1-2-1 of homothallic U. botrytis

Fungal transformation and molecular characterization of gene knockout mutants were conducted according to Leng et al.⁶³. The split-marker system⁶⁴ was used for gene deletion, and hygB ^R or G418 transformants were purified by successive transfer of young hyphal tips of U. botrytis to selective medium and screened for self-sterility. The MAT1-1-1 and MAT1-2-1 genes in the asexual U. botrytis were identified in a previous study³⁰. U. botrytis MAT1-1-1 or MAT1-2-1 was deleted using the split-marker method, with the exception that the entire selectable marker cassette was amplified from plasmid pUCATPH⁶⁵, then fused to the 5′ and 3′ flanking fragments of the MAT1-1-1 or MAT1-2-1. Transformation was conducted following a described protocol⁶⁶. Single mutant ΔmatUbMAT-1 or ΔmatUbMAT-2 was individually constructed as shown in Table S1. The double mutant DmUbMAT-1 (ΔmatUbMAT1-1-1/1-2-1) was constructed by deletion of UbMAT1-2-1 from the single mutant ΔmatUbMAT-1 (Table S1). For the deletion, the 5′ and 3′ flanking fragments of MAT1-2-1 were fused to the NPTII selectable marker cassette from pII99⁶⁷ by overlapping PCR, and the fused fragment was used for transformation of the ΔmatUbMAT-1 strain (Table S1). Transformants were subjected to RT-PCR, Southern blot and Northern blot analysis to confirm deletion of MAT1-1-1, MAT1-2-1, and MAT1-1-1/1-2-1 which were performed as described below. ΔmatUbMAT-1 strain was chosen as the recipient for heterologous expression of C. heterostrophus MAT1-1-1 ΔmatUbMAT-2 strain was chosen as the recipient for heterologous expression of C. heterostrophus MAT1-2-1. DmUbMAT-1 was chosen as the recipient for heterologous expression of C. heterostrophus MAT1-1-1/1-2-1.

Transformation of C. heterostrophus and U. botrytis

Plasmid pBG, carrying bar-encoding resistance to hygB ^{R 68}, was obtained from Tsutomu Arie¹⁹. For transformation procedures, C. heterostrophus C4-41.7 (MAT₀), DmUbMAT-1, ΔmatUbMAT-1, and ΔmatUbMAT-2 strains were cultivated as described above. The preparation of C. heterostrophus C4-41.7, ΔmatUbMAT-1, ΔmatUbMAT-2, and DmUbMAT-1 protoplasts was performed as described previously^{14, 66}. Bar ^R transformants were selected on a selective regeneration medium. The segregation of antibiotic-resistant phenotypes in the sexual crosses was then scored on PCA or CMX medium.

Crossing: determination of mating phenotypes of U. botrytis transgenic strains carrying C. heterostrophus MAT genes

U. botrytis transgenic strains carrying opposite C. heterostrophus MAT genes were crossed and selfed as indicated in Table S2. The unsuccessful crosses were as follows: DmUbMAT-1{ChMAT}-1-1× DmUbMAT-1{ChMAT}-2-1, and ΔmatUbMAT-1× ΔmatUbMAT-2. The successful self matings were as follows: U. botrytis strain (Wild type), DmUbMAT-1{ChMAT}-3-1, ΔmatUbMAT-1{ChMAT}-1, and ΔmatUbMAT-2{ChMAT}-1. The negative controls were as follows: a self-mating U. botrytis strain and a cross mating ΔmatUbMAT-1× ΔmatUbMAT-2. The positive controls were as follows: a self-mating C. heterostrophus (2847) and a cross mating C. heterostrophus C5 × C. heterostrophus C4. All cross and self-mating strains were cultured on the corn leaf substrate as described above. Fertility from self or cross mating was determined by checking the number of pseudothecia per square centimeter of area on the corn leaf substrates, the number of asci in individual pigmented pseudothecia, and the number of ascospores in individual asci. For the initial screening, at least 10 pseudothecia were opened and the number of asci per pseudothecium were recorded. Each experiment was repeated at least three times.

Crossing: determination of mating phenotypes of C. heterostrophus transgenic strains carrying U. botrytis MAT genes

The transgenic strains ChΔmat0{UbMAT}-2, ChΔmat0{UbMAT}-3 and ChΔmat0 {UbMAT}-4 carrying U. botrytis MAT1-1-1 or MAT1-2-1 were mated in pairs as indicated in Table S1. One cross was performed with a heterothallic MAT gene pattern: ChΔmat0 {UbMAT}-2 was crossed to ChΔmat0{UbMAT}-3 on the surface of corn leaf substrates. Control cross patterns: C. heterostrophus C5 was crossed to C. heterostrophus C4; ChΔmat0{UbMAT}-4 or C. heterostrophus (2847) was individual selfed. Fertility from self or cross mating was determined by checking the number of pigmented pseudothecia per square centimeter on the surface of corn leaf substrates, the number of asci in individual pseudothecia, and the number of ascospores in individual asci. For the initial screening, at least 10 pseudothecia were opened and the number of asci in each pseudothecium were recorded. Each experiment was repeated at least three times.

Nucleic acid manipulation

U. botrytis strain cultivation and DNA extraction were conducted as previously described³⁰. C. heterostrophus strain growth and genomic DNA purification followed the procedures described by Turgeon et al.¹¹. Total RNA was extracted using the TRrizol reagent (Invitrogen, USA) according to the manufacturer’s protocol. PCR amplifications were performed in a total volume of 20 μl containing 0.4 μM of each dNTP, 5 μM of each primer, 1 unit of easy Taq or 2 units easy Pfu DNA polymerase (Trans, China), 2.0 μl of 10 reaction buffer, and 10 to 20 ng of genomic DNA. Southern blotting and Northern blotting were adjusted slightly according to previous descriptions⁶⁹. For Southern blot analysis of MAT genes in the transgenic strains of U. botrytis deletion lines and C. heterostrophus C5, C4 and C4-41.7 (MAT ₀), MAT-specific probes were prepared by PCR amplification (Table S3) of MAT1-1-1 and MAT1-2-1 from U. botrytis strain (CBS 198.67), C. heterostrophus strains C5 and C4, respectively, using primers UMAT1-1F and UMAT1-1R to amplify MAT1-1-1 from U. botrytis, and UMAT1-2F and UMAT1-2R to amplify MAT1-2-1 from U. botrytis; using primers CMAT1-1F and CMAT1-1R to amplify MAT1-1-1 from C5, and CMAT1-2F and CMAT1-2R to amplify MAT1-2-1 from C4. For Southern blot analysis of MAT deletion lines in U. botrytis, both hygB and G418 probes were used detect transgene insertion. PCR amplicons were column purified and approximately 1 μg of DNA was random prime labeled with digoxigenin-11-dUTP using the DIG DNA Labeling and Detection Kit (Roche Diagnostics, Indianapolis, IN) according to the manufacturer’s instructions. Hybridization, washing, and chemiluminescent detection with CSPD were carried out with the same kit. Hybridization was detected by exposing the membranes to Kodak X-OMAT film (Kodak, Rochester, NY) for 15–30 min and developed under standard conditions. The Northern blot was also adjusted slightly according to previous descriptions⁶⁹.

The expression of U. botrytis MAT1-1-1 and MAT1-2-1 loci in C. heterostrophus MAT deletion lines and C. heterostrophus C5, C4 and C4-41.7 strains was analyzed for RNA expression using qRT-PCR. RT-PCR was performed with the PrimeScript strand cDNA Synthesis Kit (Takara, Japan) following the supplier’s instructions. Transcript levels were quantitated using either the threshold cycle (ΔΔCT) method or a relative standard curve. SYBR green sequence detection was performed using the StepOne real-time PCR system (Applied Biosystems)⁷⁰. To monitor the expression of U. botrytis MAT1-1-1 or MAT1-2-1 in reference transgenic C. heterostrophus strains C5, C4 and C4-41.7, we used the primers listed in Table S3. The C. heterostrophus actin gene (AY748990) was used as the endogenous control to normalize the expression of MAT1-1-1 or MAT1-2-1 in all transgenic lines of C. heterostrophus. To monitor the expression of the C. heterostrophus MAT1-1-1 or MAT1-2-1 in reference transgenic U. botrytis strains, we used the primers listed in Table S3. The actin gene was used as the endogenous control to normalize the expression of MAT1-1-1 or MAT1-2-1 genes in all transgenic lines of U. botrytis. Actin-F and Actin-R primers were used to amplify the actin gene in all tested strains (Table S3). Validation experiments of target genes and control genes for the comparative ΔΔCT method were performed according to the instructions of Applied Biosystems⁷⁰. For a valid ΔΔCT method calculation, the efficiency of the target amplification and the efficiency of reference amplification must be approximately equal. Relative quantitation is expressed as a difference in target gene expression with respect to an endogenous control in different samples. Each cDNA sample was assayed in triplicate, and RNAs were obtained from three separate biological samples.

Light Microscopy

For microscopic studies, all transformants, C. heterostrophus or U. botrytis wild-type and MAT-deleted strains were cultivated using standard conditions^{11, 32}. Microscopy was performed using an Olympus BX-53 microscope (Tokyo, Japan). The preparations of fruiting bodies and asexual spores of C. heterostrophus or U. botrytis were conducted following the procedures described by Wang et al.³² and Turgeon et al.¹¹. The pseudothecia and asci produced by the different transformants from the cross or self matings were stained with cotton blue. Photographs were subsequently processed using the Autolevel and Autocontrast features of Adobe Photoshop 9.0. Each experiment was repeated at least three times.

Author Contributions

All authors contributed extensively to the work presented in this paper. Xiu Guo Zhang conceived and designed the experiments. Qun Wang, ShiWang and Chen Lin Xiong conducted the experiments. Qun Wang and Shi Wang contributed equally to the research: Xiu Guo Zhang and Timothy Y. James wrote the manuscript. All authors reviewed the final version of the manuscript.

Competing Interests

The authors declare that they have no competing interests.