Direct Regulation of Extracellular Chitinase Production by the Transcription Factor LeClp in Lysobacter enzymogenes OH11

Lysobacter is a gram-negative bacterial genus belonging to the family Xanthomonadaceae (Christensen and Cook 1978). This genus consists of a wide range of species from diverse ecological niches, including plant surface, rhizosphere soil, and water (Christensen and Cook 1978; Hayward et al. 2010). Considerable research attention has been paid to this genus due to its capacity to produce a multitude of new antibiotics and abundant lytic enzymes (Christensen and Cook 1978; Hayward et al. 2010; Xie et al. 2012). These characteristics led some species, such as Lysobacter enzymogenes, L. antibioticus, and L. capsici, to be important biological control agents against a variety of crop diseases (Ji et al. 2008; Ko et al. 2009; Puopolo et al. 2014; Qian et al. 2009; Zhang and Yuen 1999). Among these species, the best-studied is L. enzymogenes (Christensen and Cook 1978; Kobayashi et al. 2005; Qian et al. 2012, 2013, 2014; Zhang and Yuen 1999). As denoted by its name, this species is shown to produce extracellular lytic enzymes such as chitinases, proteases, and β-1,3-glucanases (Kobayashi et al. 2005). These enzymes have important roles in the antifungal activity of L. enzymogenes because they can efficiently degrade the cell walls of fungi (Palumbo et al. 2005; Zhang and Yuen 2000).

The role of chitinases in control of fungal crop diseases has been well demonstrated in L. enzymogenes. In an early study, treatment of conidia of the fungus Bipolaris sorokiniana with chitinolytic fractions from L. enzymogenes strain C3 caused conidial deformation and abnormal germ tube formation; these effects were lost when the chitinolytic fractions were heated (Zhang and Yuen 2000). In a later report, L. enzymogenes C3 was found to produce at least two chitinases that are antifungal, including chitinase A and an unknown chitinase with a molecular weight of 48 kDa (Zhang et al. 2001). These early reports demonstrated the importance of chitinase production to L. enzymogenes biocontrol activity. In this regard, understanding of the molecular mechanisms involved in the regulation of chitinase biosynthesis in L. enzymogenes will be helpful in developing methods to enhance chitinase production and, thus, improve biocontrol effectiveness.

In the first report on the genetic regulation of chitinase biosynthesis in L. enzymogenes (Kobayashi et al. 2005), chitinase production by L. enzymogenes C3 was shown to be modulated by the cAMP receptor-like protein (Clp) transcription factor, with inactivation of clp resulting in near-complete deficiency in chitinase production. Furthermore, a recent screen of a transposon-based mutant library of strain C3 led to identification of four new regulators controlling the production of extracellular chitinase: glucose-galactose transporter, disulfide bond formation protein B, Clp protease, and polyamine synthase (Choi et al. 2012). However, it is not clearly understood as to how these regulators function in controlling extracellular chitinase production in L. enzymogenes C3.

In agreement with the findings using strain C3, we recently showed that extracellular chitinase activity in L. enzymogenes OH11, a strain isolate in China, is also regulated by Clp (Qian et al. 2009; Xu et al. 2015); mutation of clp also completely abolished extracellular chitinase production in strain OH11. A further study showed that, although strain OH11 possesses three predicted chitinase-encoding genes, only chiA, whose product is chitinase A, is required for extracellular chitinase activity (Xu et al. 2015). These results suggest that Clp most likely regulates extracellular chitinase activity in L. enzymogenes OH11 through control of chitinase A biosynthesis or secretion. The results from investigations on strains C3 and OH11 collectively suggest that Clp plays a conserved role in controlling extracellular chitinase activity; however, the precise mechanism by which Clp functions in each strain remains unclear.

The objective of the present work was to determine how Clp regulates extracellular chitinase production in L. enzymogenes, using strain OH11 as the subject. In the present study, the Clp of L. enzymogenes is designated “LeClp”. We provide abundant evidence to show that LeClp directly binds the promoter region of chiA under both in vitro and in vivo conditions, thus resulting in a positive regulation of chiA transcription. Our studies establish the first detailed report on the mechanism by which LeClp exerts its regulation on extracellular chitinase production in L. enzymogenes. Our results also show that the direct regulatory mode of LeClp on chiA appears to be a conserved attribute in diverse chitinase-producing Lysobacter spp.

MATERIALS AND METHODS

Bacterial strains, plasmids, and growth conditions.

The bacterial strains and plasmids used in this study are listed in Table 1. Escherichia coli strains used for plasmid construction were routinely grown in Luria broth (LB) or on LB agar plates at 37°C, supplemented with gentamicin (Gm) (25 μg/ml) or 5-bromo-4-chloro-3-indolyl-β-D-galacto-pyranoside (100 μg/ml) as needed. L. enzymogenes stains were grown in LB medium or 1/10 tryptic soy broth (TSB) at 28°C. When required, antibiotics were added into the medium according to the following final concentrations: kanamycin (Km) at 100 μg/ml, Gm at 150 μg/ml, and ampicillin (Amp) at 100 μg/ml.

TABLE 1 Bacterial strains and plasmids used in this study

Strains, plasmids	Characteristicsa	Source
Lysobacter enzymogenes
OH11	Wild-type strain, KmR	Qian et al. 2009
ΔLeclp	Leclp in-frame deletion mutant of strain OH11, KmR	Wang et al. 2014
ΔLeclp(Leclp)	Mutant ΔLeclp harboring plasmid pBBR-clp, KmR, GmR	Wang et al. 2014
ΔLeclp(pBBR)	Mutant ΔLeclp harboring plasmid pBBR1-MCS5, KmR, GmR	Wang et al. 2014
ΔchiA	chiA in-frame deletion mutant of strain OH11, KmR	Qian et al. 2012
ΔchiA(chiA)	Mutant ΔchiA harboring plasmid pBBR-chiA, KmR, GmR	Xu et al. 2015
ΔchiA(pBBR)	Mutant ΔchiA harboring plasmid pBBR1-MCS5, KmR, GmR	Xu et al. 2015
Δclp(chiA)	Mutant ΔLeclp harboring plasmid pBBR-chiA, KmR, GmR	This study
Escherichia coli
DH5a	F−, φ80dlacZ∆M15, ∆(lacZYA-argF)U169, deoR, recA1, endA1, hsdR17(rk−,mk+), phoA, supE44, λ−, thi-1, gyrA96	Lab collection
BL21 (DE3)	F−, ompT, gal, dcm, lon, hsdSB(rB−,mB−), λ(DE3)	Lab collection
XL1-Blue MRF´ Kan	Δ(mcrA)183, Δ(mcrCB-hsdSMR-mrr)173, endA1, supE44, thi-1, recA1 gyrA96, relA1, lac, [F´proAB lacIqZΔM15 Tn5 (KmR)]	Guo et al. 2009
Plasmids
pMD19-T	ColE1 origin, T Simple vector, AmpR	TaKaRa
pBBR1-MCS5	Broad host range cloning vector, lacZ, GmR	Kovach et al. 1995
pBBR-chiA	pBBR1-MCS5 with 2.5kb fragment including chiA, GmR	Xu et al. 2015
pGEX-6p-1	vector for protein expression, AmpR	Deng et al. 2012
pGEX-6p-1-clp	pGEX-6p-1 with the coding region of Leclp, AmpR	This study
pTRG	Plasmid used for protein expression in bacterial one-hybridization assay, TetR	Guo et al. 2009
pTRG-clp	pTRG with the coding region of Leclp, TetR	This study
pBXcmT	Plasmid used for DNA cloning in bacterial one-hybridization assay, ChloR	Guo et al. 2009
pBXcmT-chiA	pBXcmT with the predicted chiA promoter region, ChloR	This study

^a Km^R, Gm^R Amp^R, Tet^R, and Chlo^R = resistant to kanamycin, gentamicin, ampicillin, tetracycline, and chloramphenicol, respectively.

View as image HTML

Bacterial one-hybrid assay.

The bacterial one-hybrid reporter system was shown to be efficient for the testing of the physical interaction of the transcription factor with the promoter of target genes (Guo et al. 2009). This assay was applied to examine the potential interaction between the transcriptional regulator (LeClp) and the target DNA (the chiA promoter) in the present study. As described previously, the bacterial one-hybrid reporter system involved three components: plasmids pBXcmT and pTRG, which were used for cloning the bait DNA and expressing a target protein, respectively, and E. coli XL1-Blue MRF’ kan strain (Table 1), which is the host strain for propagating pBXcmT and pTRG recombinants (Guo et al. 2009). In the present study, the chiA promoter region (400 bp) of L. enzymogenes OH11 was cloned into pBXcmT, generating the recombinant vector pBXcmT-chiA (Table 1). Similarly, the coding region of Leclp (690 bp) was cloned into pTRG, creating the final construct pTRG-clp (Table 1). The vector pBXcmT-chiA and pTRG-clp were transformed into XL1-Blue MRF’ kan strain. If the direct physical binding occurs between LeClp and the chiA promoter, the positive-transformed E. coli strain containing both pBXcmT-chiA and pTRG-clp will be expected to grow well on the selective medium, which is the minimal medium containing 5 mM 3-amino-1,2,4-triazole, streptomycin at 8 μg/ml, tetracycline at 12.5 μg/ml, chloramphenicol at 34 μg/ml, and Km at 30 μg/ml, as described previously (Guo et al. 2009). Furthermore, the cotransformant containing the vector pBX-R2031/pTRG-R3133 served as a positive control (Guo et al. 2009), while the cotransformant containing the empty pTRG and pBXcmT-chiA was used as a negative control in the present study. All cotransformants were spotted onto the selective medium and grown at 28°C for 3 to 4 days, then photographed.

Protein expression and purification.

The assay of protein expression and purification was performed as previously described (Deng et al. 2012). In brief, the gene Leclp was amplified by polymerase chain reaction (PCR) with the designated primer pairs listed in Supplementary Table S1. The PCR product of Leclp was purified and digested by BamHI and XhoI, and further cloned into the BamHI/XhoI-digested expression vector pGEX-6p-1, creating the final construct pGEX-6p-1-clp (Table 1). This vector was transformed into E. coli strain BL21 (DE3) for protein expression. In brief, the resultant transformed strain was cultivated in LB medium containing Amp at 100 μg/ml overnight at 37°C. Then, a 5-ml overnight culture was transferred into a fresh 500-ml LB medium in the presence of Amp at 100 μg/ml and grown at 37°C with shaking at 200 rpm until the optical density at 600 nm (OD₆₀₀) value reached 0.4. Subsequently, isopropyl β-D-1-thiogalactopyranoside at a final concentration of 0.4 mM was added to the culture, followed by further growth at 28°C for 4 h. Then, the cells were collected by centrifugation (6,000 rpm) at 4°C and resuspended in 15 ml of phosphate-buffered saline (PBS) lysis buffer supplemented with phenylmethylsulfonyl fluoride (PMSF) at a final concentration of 1 mM. The cells were lysed by a brief sonication and the crude cell extracts were centrifuged at 13,000 rpm and 4°C. Soluble protein fractions were collected and mixed with preequilibrated Glutathione Sepharose 4B (GE) for 6 h at 4°C, which was placed into a column and extensively washed with PBS buffer. The proteins were subsequently eluted using elution buffer that contained the reduced glutathione. Protein purity was assessed by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and protein concentration was determined using a Bradford protein assay kit (Bio-Rad).

Electrophoretic mobility shift assay.

The electrophoretic mobility shift assay (EMSA) was performed as described previously (Wang et al. 2013). In brief, DNA probes (the 400-bp chiA promoter region) were amplified by PCR with the primers pchiA-F and pchiA-R. The DNA probes used in this study were labeled with [α-³²P]-dCTP by DNA polymerase Klenow fragments, and free [α-³²P]-dCTP was removed by ProbeQuant G-50 column (GE). The GST-tagged protein LeClp was incubated with 6 fmol of radiolabeled oligonucleotide duplex in a 20-μl reaction containing 10 mM Tris (pH 7.0), 50 mM KCl, 1 mM dithiothreitol (pH 7.5), 2.5% glycerol, 5.0 mM MgCl₂, 1.0 μg poly(dI:dC), 0.05% NP-40, and 10 mM EDTA for 20 min at room temperature. In competition experiments, unlabeled probes were added during the preincubation period. Free DNA and protein-bound DNA were separated on 5% nondenaturing polyacrylamide gel in 0.5× Tris-boric acid-EDTA. After electrophoresis, gels were blotted onto filter paper, dried, and autoradiographed.

Chromatin immunoprecipitation PCR.

The chromatin immunoprecipitation (ChIP) assay was conducted as previously described (Wang et al. 2013). In brief, L. enzymogenes stain OH11 was incubated in LB medium at 28°C with shaking at 200 rpm, until the value of OD₆₀₀ reached 1.0. Then, the cells were harvested by centrifugation at 6,000 rpm and suspended in an equivalent volume of 10% TSB broth at 28°C in a rotary shaker at 200 rpm for 1 h. All bacterial cells were incubated with 1% formaldehyde for 10 min at room temperature for cross-linking and subsequently quenched with 0.5 M glycine for 5 min. For DNA preparation, the cells were further washed with 10 ml of cold PBS buffer twice, then scraped into 1.2 ml of ChIP lysis buffer (10 mM Tris-HCl, pH 8.0; 20% sucrose; 50 mM NaCl; and 10 mM EDTA) and 4.8 ml of ChIP buffer (50 mM HEPESKOH, pH 7.5; 150 mM NaCl; 1 mM EDTA; 1% Triton X 100; 0.1% sodium deoxycholate; 0.1% SDS; and 1 mM PMSF). These lysates were sonicated using a Diagenode Bioruptor UCD-300 (Diagenode) to generate DNA fragments that ranged in the size from 100 to 500 bp. After centrifugation, a 100-μl aliquot was kept as the positive control DNA (Input). The 800-μl aliquot was incubated with 2 μl of the respective antibody (anti-Clp; gift from W. Qian, Chinese Academy of Science) and 30 μl of protein A/G agarose beads overnight at 4°C (Sample). As a negative control, only 30 μl of protein A/G agarose beads were added in another 800-μl aliquot (Mock). The sepharose beads were collected by centrifugation and washed with immunoprecipitation (IP) buffer, IP buffer with 500 mM NaCl, wash buffer, and Tris-EDTA buffer. Then, the beads were suspended in 100 μl of elution buffer (50 mM Tris-HCl, pH 7.5; 10 mM EDTA; and 1% SDS), and incubated for 1.5 h at 42°C with RNase A and an additional 6 h at 65°C with proteinase K to reverse protein/DNA cross-links. Finally, these samples were processed for DNA purification by phenol-chloroform extraction and ethanol precipitation. The quantity of DNA was detected by the NanoDrop ND-1000 (NanoDrop). Finally, the PCR assay was performed in 25 μl with the input DNA, sample DNA, and mock DNA, respectively, with the relevant primer pairs. The PCR products were separated on a 1.5% agarose gel and stained with ethidium bromide. For further validation, these PCR products were also sequenced in the present study.

Western blot analysis.

The Western blot analysis was performed according to the laboratory protocol, with a minor modification (Xu et al. 2015). In brief, GST-tagged protein LeClp was separated by SDS-PAGE and immobilized onto a polyvinylidenem difluoride membrane using the semidry blot machine (Bio-Rad). Membranes were probed with the specific primary anti-GST (M20007; Abmart), followed by an HRP-conjugated antimouse secondary antibody (M21001; Abmart) for signal detection.

Determination of extracellular chitinase activity.

The production of extracellular chitinase from L. enzymogenes strains was determined as described previously (Qian et al. 2009; Xu et al. 2015). In brief, the test L. enzymogenes strains were incubated in LB medium at 28°C with shaking at 200 rpm, until the value of OD₆₀₀ reached 1.0. Then, 2 μl of bacterial culture was spotted on LB plates contained colloidal chitin. After 96 h of incubation at 28°C, chitinase activity was assessed according to the cleaning zones around the bacterial colonies. Each treatment involved three replications, and the same experiment was performed three times.

RESULTS

LeClp bound directly to the promoter region of chiA in vitro.

To explore how LeClp regulates the transcription of chiA, we tested whether LeClp can directly bind the promoter of chiA via the C-terminal helix-turn-helix (HTH)-DNA binding domain of LeClp. To test this point, the characterized binding consensus sequence (5′-NTGTGA-N6-TCACANTTTN-3′) of E. coli CRP (http://regulondb.ccg.unam.mx/), the homolog of LeClp, was chosen for predicting whether the promoter of chiA possessed any potential DNA binding site of LeClp. This led to identification of a potential binding site (5′-TTGTCACGCGGCGCCAGTTTC-3′) of LeClp at the chiA promoter (Fig. 1A). This result supported our initial idea and further suggested that LeClp could directly bind to the chiA promoter. To validate this point, we employed the E. coli-based one-hybrid reporter system (described in detail in Materials and Methods). We clearly observed that the test E. coli strain containing both LeClp and the chiA promoter grew very well on the selective medium, similar to the positive control, whereas the negative control failed to grow (Fig. 1B). This result indicated that a direct binding of LeClp on the promoter of chiA occurred under the test conditions.

Download as PowerPoint

To further confirm the above finding, an EMSA/gel shift assay was performed. We found that the GST-tagged LeClp formed a clear protein/DNA complex with the test DNA probe, whereas no shift band was observed between the empty GST and DNA probe under the similar test conditions (Fig. 2). Furthermore, addition of an unlabeled probe (cold probe) at 10- and 1,000-fold excesses to the reaction mixtures competitively inhibited the binding of LeClp to the labeled DNA probe. This EMSA result, together with the data from Figure 1, strongly showed that LeClp directly bound to the promoter region of chiA under the in vitro condition.

Download as PowerPoint

A direct in vivo physical interaction occurred between LeClp and the chiA promoter.

Next, we attempted to test whether an in vivo binding of LeClp on the promoter of chiA occurs in L. enzymogenes cells. To achieve this point, the ChIP-PCR assay was carried out, whose scheme is presented in Figure 3A (described in detail in Material and Methods). Clearly, a 212-bp promoter region of chiA was amplified by PCR from the designated Sample (the chromatin DNA of strain OH11 immunoprecipitated with the respective antibody), which was similar to the case from the positive control Input (Fig. 3B). However, under the same condition, the negative control Mock (the chromatin DNA of strain OH11 immunoprecipitated without the respective antibody) failed to amplify the corresponding PCR products (Fig. 3B). These results suggested that L. enzymogenes LeClp bound to the promoter region of chiA in vivo. Collectively, all of the above results as well as our previous report (Wang et al. 2014) clearly reveal that LeClp exerted its positive regulation on chiA transcription through a direct interaction with the chiA promoter region in L. enzymogenes.

Download as PowerPoint

LeClp is an important but not the sole regulator of chiA transcription.

To address whether LeClp is the sole regulator of chiA transcription, the plasmid-borne chiA under the control of its native promoter was introduced into the Leclp mutant. This introduction caused a minor but visible effect in restoring the deficiency of the Leclp mutant in producing extracellular chitinase (Fig. 4). However, the Leclp mutant possessing an empty vector was still fully deficient in this ability. Furthermore, as controls, the Leclp or chiA mutant completely lost the ability to produce extracellular chitinase, where the trait was restored in the respective complemented strain to the wild-type level (Fig. 4). Taken together, these results suggested that, although LeClp was important, it was not the sole regulator of chiA transcription in L. enzymogenes.

Download as PowerPoint

The direct interaction between LeClp and the chiA promoter appears to be a conserved mechanism in diverse chitinase-producing Lysobacter spp.

Recently, the genomes of four representative chitinase-producing Lysobacter spp., including L. enzymogenes C3, L. antibioticus ATCC29479, L. capsici 55, and L. gummosus 3.2.11, were reported (de Bruijn et al. 2015). To explore whether the direct interaction between LeClp and the chiA promoter in strain OH11 is also conserved in those abovementioned Lysobacter spp., a comparative sequence alignment was performed. Clearly, all of the sequenced Lysobacter spp. possessed a highly conserved LeClp homolog (Fig. 5), and their corresponding HTH-binding domain at the amino-acid sequence level was almost same as that in strain OH11 (Fig. 5A). Furthermore, the promoter regions of chiA from all Lysobacter spp. were predicted to have a DNA binding consensus sequence of E. coli CRP (Fig. 5B), which was similar to the case from strain OH11. Collectively, the sequence analysis of the putative LeClp binding site indicated that the LeClp homolog could be involved in the regulation of extracellular chitinase production in those closely related Lysobacter spp. described above, by a mechanism similar to that in L. enzymogenes OH11.

Download as PowerPoint

DISCUSSION

As a global regulator, LeClp makes an important contribution to the production of diverse extracellular lytic enzymes in L. enzymogenes (Kobayashi et al. 2005; Wang et al. 2014); however, the corresponding molecular mechanism is unknown. In the present study, we provided the first report to show that the regulation of LeClp on the production of extracellular chitinase was, at least in part, due to the direct binding of LeClp to the promoter of chiA. Such a regulatory mechanism seems to be conserved in diverse chitinase-producing Lysobacter spp. This new knowledge will be helpful to generate high chitinase-producing L. enzymogenes strains for better application in the biocontrol of crop fungal disease.

Although this study provides abundant evidence of the direct interaction between LeClp and the chiA promoter, LeClp appears not to be the sole regulator of chiA transcription (Fig. 4). It is possible that, in addition to LeClp, there is another unidentified transcription factor that can bind with the native chiA promoter, causing some leaky transcription of chiA from the introduced chiA plasmids (maybe hundreds). However, in the chromosomal background of the Leclp mutant, there is only one copy of chiA; hence, the phenotype rescue phenomenon is less obvious.

The regulation of Clp on enzyme biosynthesis has been well characterized in the phytopathogenic bacterium Xanthomonas campestris pv. campestris (He et al. 2007). In this bacterium, the regulator Clp serves as a downstream regulator of the diffusible signal factor (DSF) cell-to-cell signaling to modulate diverse enzyme production (He et al. 2007). Although L. enzymogenes also contains a functional LeDSF system, this system did not regulate the biosynthesis of extracellular lytic enzymes such as chitinase (Qian et al. 2013). Therefore, in contrast to the findings in X. campestris, LeClp appears to utilize an LeDSF-independent manner to regulate the production of extracellular chitinase. These results suggested that the Clp family proteins have evolved to acquire species-adaptive mechanisms to control respective enzyme biosynthesis in X. campestris and L. enzymogenes, two related but different bacterial species at taxonomy.

In X. campestris, the regulation of Clp on extracellular enzyme biosynthesis is dependent on c-di-GMP, a newly identified bacterial second messenger (Chin et al. 2010; Tao et al. 2010). A representative example is the endoglucanase enzyme encoding by engXCA that contributes to X. campestris virulence through its cellulolytic activity (Gough et al. 1990). In X. campestris cells, Clp is a c-di-GMP receptor with high affinity (Chin et al.2010; Tao et al. 2010). Clp directly binds the promoter of engXCA, resulting in a positive control of its transcription. High-level c-di-GMP inhibits such binding, thus suggesting that c-di-GMP makes a negative contribution to formation of Clp-DNA complex (Chin et al. 2010; Tao et al. 2010). Although LeClp shares a high amino-acid similarity with the Clp of X. campestris (Kobayashi et al. 2005), it remains unclear whether the direct regulation of LeClp in transcription of chiA is c-di-GMP dependent, which needs further investigations.

In summary, our findings present new insights into how the global regulator LeClp controls the production of extracellular chitinase, which provides useful guidance to understand the regulation of LeClp in control of other extracellular lytic enzymes, such as β-1,3-glucanase in L. enzymogenes. Also, our studies may trigger the study of other chitinase-producing and underexplored Lysobacter spp. possessing the homologs of both LeClp and ChiA.

ACKNOWLEDGMENTS

G. Qian is a visiting scholar in M. Gomelsky’s laboratory at the University of Wyoming, supported by the China Scholarship Council. We thank W. Qian, from the Institute of Microbiology of Chinese Academy of Sciences, for providing the anti-Clp antibody as a gift, as well as his technical support in the EMSA; Z.-G. He, from Huazhong Agricultural University (China), for kindly providing the bacterial one-hybrid system; and G. Y. Yuen (University of Nebraska-Lincoln) for his critical editing. This study was supported by the National Basic Research (973) program of China (2015CB150600), National Natural Science Foundation of China (31371981, 31572046), Special Fund for Agro-Scientific Research in the Public Interest (No. 201303015) grants from Jiangsu Agriculture Science and Jiangsu Province (ZX(15)1006, BE2014386, BE2015354), Program for New Century Excellent Talents in University of Ministry of Education of China (NCET-13-0863), and Fundamental Research Funds for the Central Universities (KYTZ201403). Research in the Du lab is supported, in part, by NIH R01AI097260 and a UNL Redox Biology Center pilot grant. S.-H. Chou is supported by the Ministry of Education, Taiwan, under the ATU plan, and by the National Science Council, Taiwan (grants 102-2113-M-005-006-MY3).