Endopolygalacturonase PG1 in Different Formae Speciales of Fusarium oxysporum

Abstract

PG1, the major endopolygalacturonase of the vascular wilt pathogen Fusarium oxysporum, was secreted during growth on pectin by 10 of 12 isolates belonging to seven formae speciales, as determined with isoelectric focusing zymograms and sodium dodecyl sulfate-polyacrylamide gel electrophoresis gels. A Southern analysis of genomic DNA and PCR performed with gene-specific primers revealed that the pg1 locus was highly conserved structurally in most isolates. Two PG1-deficient isolates were identified; one lacked the encoding gene, and the other carried a pg1 allele disrupted by a 3.2-kb insertion with sequence homology to hAT transposases. The virulence for muskmelon of different F. oxysporum f. sp. melonis isolates was not correlated with PG1 production in vitro. We concluded that PG1 is widely distributed in F. oxysporum and that it is not essential for pathogenicity.

It has been proposed that endopolygalacturonases (endoPGs) (poly-α-1,4-galacturonide glycanohydrolases; EC 3.2.1.15) play a key role in fungal pathogenicity for plants by depolymerizing homogalacturonan, a major component of the plant cell wall (4). They may also function as avirulence determinants by releasing oligogalacturonide inducers of plant defense mechanisms (6) and interacting with plant proteins that modulate polygalacturonase (PG) activity (3). Fusarium oxysporum Schlecht. is an economically important soilborne plant pathogen that has a worldwide distribution and causes vascular wilt disease in a wide variety of crops. This species includes more than 120 described formae speciales that are defined on the basis of specificity for host species (1). The mechanisms of pathogenicity and wilt symptom induction by this fungus are poorly understood, although it has been suggested that endoPGs may be involved (2). Until now, no information concerning the occurrence and distribution of specific pectinolytic isozymes in this species has been available. Recently, PG1, the major endoPG produced by F. oxysporum f. sp. lycopersici during in vitro growth on pectin, was purified and characterized (7), and the corresponding gene was cloned (10). In the present study, we compared 12 F. oxysporum isolates belonging to seven different formae speciales to determine the occurrence and diversity of PG1 and the corresponding gene in the species and determined if PG1 production by F. oxysporum f. sp. melonis was correlated with virulence for muskmelon.

The F. oxysporum isolates used in the study were stored as microconidial suspensions in 30% glycerol at −80°C (Table 1). PG1 is the major pectinolytic enzyme secreted by F. oxysporum during early phases of growth on pectin (7, 10). Thus, to analyze PG1 production, 10⁸ microconidia obtained by filtering cultures grown for 4 days in potato dextrose broth (PDB) (Difco) were germinated by incubating them for 12 h in 50 ml of fresh PDB. The resulting germlings were washed twice in sterile water, transferred to 20 ml of synthetic medium (7) supplemented with 1% (wt/vol) citrus pectin (Sigma Chemical Co., St. Louis, Mo.), and incubated on a rotary shaker for 24 h at 150 rpm and 28°C. Total PG activity was determined in dialyzed culture filtrates by measuring the release of reducing groups from polygalacturonic acid by the method of Nelson-Somogyi. Appropriate control experiments without either enzyme or substrate were performed simultaneously. The quantity of reducing sugar released was calculated by using d-galacturonic acid standards (Sigma). Medium to high extracellular PG activity was detected in most filtrates; the only exceptions were the isolate 218 and 58110 filtrates, which contained low activity, and the 18M, A34, and 275 filtrates, which contained extremely low or no PG activity (Table 2). In order to check degradation of citrus pectin in the cultures, 1.2 ml of each culture filtrate was lyophilized in an Eppendorf tube. Total or partial degradation of the pectin polymers due to the action of endoPG was observed with all isolates except 18M, A34, and 275 (Fig. 1A). To visualize pectinolytic isozymes, filtrates from pectin cultures grown for 48 h were concentrated by acetone precipitation and were analyzed by using zymograms in isoelectric focusing gels (pH 3 to 9) (16). Two activity bands (pI 6.85 and 7.0) corresponding to PG1 isoforms (7, 10) were detected in all isolates except 18M and 275 (Fig. 1B). A faint activity band with a pI of 7.0 produced by these two isolates probably corresponded to an exopolygalacturonase (exoPG) (PG3) having the same pI as the main PG1 isoform that was masked in other isolates by the more active PG1 (11). After 48 h of growth, some isolates also produced significant amounts of other pectinolytic isozymes with pIs lower than the PG1 pIs.

TABLE 1.

F. oxysporum isolates used in the present study

Isolate	Taxon	Race	Host plant	Sourcea
42-87b	F. oxysporum f. sp. lycopersici	2	Tomato	I.N.I.A.
281b	F. oxysporum f. sp. lycopersici	2	Tomato	I.N.I.A.
77rb	F. oxysporum f. sp. radicis-lycopersici	NDc	Tomato	I.N.I.A.
58110	F. oxysporum f. sp. conglutinans	2	Radish	ATCC
15651	F. oxysporum f. sp. tuberosi	ND	Potato	ATCC
8503d	F. oxysporum f. sp. ciceris	5	Chickpea	I.A.S.
18Mb	F. oxysporum f. sp. melonis	1	Muskmelon	I.N.I.A.
1127b	F. oxysporum f. sp. melonis	2	Muskmelon	I.N.I.A.
Fo-8e	F. oxysporum f. sp. melonis	2	Muskmelon	UCB
A34e	F. oxysporum f. sp. melonis	0	Muskmelon	UCB
275f	F. oxysporum f. sp. melonis	1,2w	Muskmelon	C.I.D.A.
G-60301	F. oxysporum f. sp. niveum	1	Watermelon	I.P.O.

^aI.N.I.A., Instituto Nacional de Investigación Agraria, Madrid, Spain; ATCC, American Type Culture Collection, Rockville, Md.; I.A.S., Instituto de Agricultura Sostenible, Córdoba, Spain; UCB, University of California, Berkeley; C.I.D.A., Centro de Investigación y Desarollo Agraria, Almeria, Spain; I.P.O., Research Institute for Plant Protection, Wageningen, The Netherlands. 

^bProvided by J. Tello. 

^cND, not determined. 

^dProvided by R. Jimenez Díaz. 

^eProvided by T. R. Gordon. 

^fProvided by J. Gomez. 

TABLE 2.

Total PG activities in dialyzed culture filtrates of F. oxysporum isolates grown in synthetic medium containing 1% citrus pectin for 24 h, as determined by a reducing sugar assay

Isolate	PG activity (nkat ml−1)a
42-87	7.3 ± 1.1
281	2.1 ± 0.5
77r	9.6 ± 1.2
58110	3.5 ± 0.7
15651	5.9 ± 0.3
8503	5.7 ± 0.6
18M	0.2 ± 0.1
1127	6.3 ± 2.9
Fo-8	7.7 ± 1.8
A34	0.9 ± 0.6
275	0.4 ± 0.2
G-60301	6.2 ± 2.0

^aOne nanokatal was defined as the amount of enzyme that produced 1 nmol of galacturonic acid equivalent per s at 37°C. 

FIG. 1.

Analysis of pectin degradation by lyophilization (A), analytical isoelectric focusing followed by pectinolytic activity staining (B), and SDS-PAGE followed by silver staining (C) of culture filtrates of F. oxysporum isolates grown for 24 h (A and C) or 48 h (B) on synthetic medium supplemented with 1% (wt/vol) citrus pectin. Lane 1, strain 42-87; lane 2, strain 281; lane 3, strain 77r; lane 4, strain 58110; lane 5, strain 15651; lane 6, strain 8503; lane 7, strain 18M; lane 8, strain 1127; lane 9, strain Fo-8; lane 10, strain A34; lane 11, strain 275; lane 12, strain G-60301. In panel A the white residue that was visible after lyophilization was undegraded pectin. In panels B and C culture filtrates were concentrated 25-fold by acetone precipitation. The positions of isoelectric focusing and molecular weight markers are indicated on the left.

Cultures grown for 24 h on pectin were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) in 11% (wt/vol) discontinuous acrylamide gels, followed by staining with AgNO₃, which revealed two major protein bands (35 and 37.5 kDa) (Fig. 1C) that corresponded to differentially glycosylated isoforms of PG1 (7). After filtrates were concentrated up to 100-fold by acetone precipitation, the PG1 protein bands were detected in all isolates except 18M and 275 (results not shown). Three of the PG1 producers, 281 (F. oxysporum f. sp. lycopersici), 58110 (F. oxysporum f. sp. conglutinans), and A34 (F. oxysporum f. sp. melonis), secreted only small amounts of the enzyme during this growth phase (growth for 24 h) (Fig. 1C), although the zymogram results suggested that larger amounts of PG1 accumulated in filtrates of these isolates at later stages of growth (after growth for 48 h) (Fig. 1B). In summary, the combined results obtained from PG activity assays, pectin degradation studies, zymograms, and SDS-PAGE gels indicate that PG1 was the major pectinolytic enzyme secreted by F. oxysporum during the early phases of growth on pectin (growth for 24 h) and, due to its endo mode of action, was the main enzyme responsible for rapid degradation of the pectin polymer. Nevertheless, at later stages of growth (Fig. 1B) or on different pectic substrates, additional pectinolytic isozymes were produced. Thus, when the 12 isolates were grown on solid medium containing polygalacturonic acid (18), even PG1-deficient isolate 275 produced a clear halo, indicating that pectinolytic isozymes other than PG1 were present (data not shown). In fact, two additional exopolygalacturonases and one pectate lyase from F. oxysporum f. sp. lycopersici have been characterized previously (8, 9, 11).

The presence of the pg1 gene in the different F. oxysporum isolates was studied by performing a Southern hybridization analysis (17) of total genomic DNA extracted from mycelium (14) grown for 5 days in PDB on a rotary shaker at 150 rpm and 28°C. Two micrograms of DNA was digested with restriction enzymes EcoRV and HindIII and hybridized with a 735-bp internal PCR fragment (corresponding to nucleotides 114 to 849) of the pg1 gene of F. oxysporum f. sp. lycopersici 42-87 (10) labelled with a nonisotopic digoxigenin kit (Boehringer Mannheim). pg1 homologs were detected in most of the isolates studied (Fig. 2). A 2.3-kb HindIII fragment containing the pg1 promoter and coding region (10) was remarkably conserved in the majority of the isolates (Fig. 2B). One of the PG1-deficient isolates, 18M, produced band patterns consistent with the presence of at least one additional EcoRV and HindIII site, indicating that there was a possible insertion in the pg1 gene. Isolates A34 and 275 did not exhibit detectable hybridization with pg1. The same result was obtained when hybridization was performed under low-stringency conditions (56°C, one wash in 5× SSC [1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate]), indicating that no other genes with homology to pg1 were present (results not shown). Stripping and reprobing of the filters with the nit1 gene of F. oxysporum f. sp. lycopersici (12) resulted in the production of clear hybridizing bands by all isolates, demonstrating that the lack of hybridization with pg1 observed with isolates A34 and 275 did not result from degradation of DNA (Fig. 2C). Nevertheless, when a Southern analysis was performed with very large amounts of genomic DNA (>10 μg) and long exposure times, a high-molecular-weight hybridizing band was observed in A34 that comigrated with undigested genomic DNA (data not shown). In this isolate, none of the restriction enzymes used (EcoRV, HindIII, MvnI, PvuII, SmaI) produced hybridizing bands with molecular weights lower than 20 kb, suggesting that the DNA in the pg1 genomic region was not accessible to these enzymes. On the other hand, no hybridization signal was observed with isolate 275 even when large amounts of DNA and long exposure times were used, suggesting that this isolate lacks pg1.

FIG. 2.

Southern hybridization analysis of genomic DNAs of different F. oxysporum isolates. For lane contents see the legend to Fig. 1. DNA was digested with restriction enzymes EcoRV (A) and HindIII (B and C) and was probed with a digoxigenin-dUTP-labelled internal fragment of the pg1 gene (A and B) or with the nit1 gene (C) of F. oxysporum f. sp. lycopersici. Lane 5 contained no DNA. The sizes of DNA marker bands are indicated on the left.

To confirm these results, 50 ng of genomic DNA from each isolate was used for PCR amplification with the sense primer PG6 (5′-CAC TAC TGC CGA TAA CGA CT-3′) and the antisense primer PG7 (5′-CAA GAA TGA GCC CTG AGA TG-3′), which spanned a 325-bp internal region (nucleotides 326 to 650) of the pg1 gene of F. oxysporum f. sp. lycopersici (10). The PCR conditions used were as follows: denaturation at 94°C for 30 s, annealing at 55°C for 30 s, and extension at 72°C for 1.5 min, with an initial denaturation step consisting of 3 min at 94°C and a final elongation step consisting of 3 min at 72°C. Aliquots (12 μl) of the PCR products were analyzed on a 2% agarose gel. The expected 325-bp band was amplified with all of the isolates except 275, strongly suggesting that this isolate lacked the pg1 gene (Fig. 3A).

FIG. 3.

PCR performed with genomic DNAs of different F. oxysporum isolates and primers specific for pg1. (A) Primers PG6 and PG7 (see Materials and Methods). For the contents of lanes 1 through 12 see the legend to Fig. 1. Aliquots of the PCR products were electrophoresed on a 2% agarose gel along with a 100-bp ladder marker (lane M). (B) PCR performed with primers PG7 and PG8 (lanes 1 through 5) or primers PG6 and PG9 (lanes 6 through 11) and genomic DNAs of isolates 42-87 (lanes 1 and 6), 1127 (lanes 2 and 7), 18M (lanes 3, 8, and 11), A34 (lane 4 and 9), and 275 (lanes 5 and 10) by using an extension time of 2.5 min (lanes 1 through 10) or 4.5 min (lane 11). Aliquots were electrophoresed on a 1.4% agarose gel (or on a 0.7% agarose gel [lane 11]) along with lambda HindIII markers (lane M). The approximate sizes of amplified bands are indicated on the left.

An additional PCR analysis was performed with isolates 42-87, 1127, 18M, A34, and 275 by using the following primer combinations: PG7 and sense primer PG8 (5′-TCT TGT CTT TGT CTC ACT CG-3′), which spanned 669 bp of the first part of the pg1 coding region (nucleotides −18 to 650); and PG6 and antisense primer PG9 (5′-AGT GAA CAG GGA GTG ATG AT-3′), which encompassed 1,027 bp from nucleotide 326 to 34 nucleotides downstream of the pg1 stop codon. The PCR conditions were the same as those described above, except that the extension time was 2.5 min. Aliquots (12 μl) of the PCR products were analyzed on a 1.4% agarose gel. With primers PG7 and PG8 the expected fragment was amplified with all of the isolates except 275, whereas with primers PG6 and PG9, both the 18M preparation and 275 preparation lacked the expected amplification product (Fig. 3B). When a PCR was performed with primers PG6 and PG9 and a longer extension time (4.30 min), a 4.2-kb fragment instead of a 1-kb fragment was amplified from 18M, suggesting that this isolate carries a 3.2-kb insertion in the second part of the pg1 gene. A 0.8-kb EcoRV-HindIII fragment from this PCR amplification product was subcloned into pBluescript/SK⁺, and both strands were sequenced with a DyeDeoxy terminator cycle sequencing kit (Perkin-Elmer, Foster City, Calif.) and a model ABI Prism 310 genetic analyzer (Applied Biosystems, Foster City, Calif.). The deduced amino acid sequence exhibited significant homology to the sequences of the hAT family transposable elements, particularly the transposon restless from the fungus Tolypocladium inflatum (13).

To study the possible correlation between PG1 production and virulence, four F. oxysporum f. sp. melonis isolates, 18M, 1127, A34, and 275, which differed in PG1 production, were used in pathogenicity assays performed in a growth chamber. Ten-day-old seedlings of muskmelon (cultivar Amarillo Canario; susceptible to all F. oxysporum f. sp. melonis races) in the first true leaf stage were inoculated by dipping the roots for 30 min in a suspension containing 5 × 10⁶ microconidia ml of water⁻¹. Control plants were immersed in sterile water. Ten seedlings per treatment were planted in minipots containing vermiculite and were incubated in a growth chamber at 25°C with a photoperiod consisting of 14 h of light and 10 h of dark. At different times after inoculation, the severity of disease symptoms was recorded by using a scale ranging from 1 (healthy plant) to 5 (dead plant). The PG1 producer 1127 was highly virulent, killing the plants after 10 days (Fig. 4). PG1-deficient isolate 18M exhibited the same degree of virulence as 1127, whereas A34 and 275 exhibited lower levels of virulence (the disease symptoms were both delayed and not as severe as the disease symptoms observed with the other strains). The experiment was repeated twice with similar results.

FIG. 4.

Incidence of Fusarium wilt caused by different isolates of F. oxysporum f. sp. melonis on muskmelon plants (cultivar Amarillo Canario). The severity of disease symptoms was recorded at different times after inoculation by using a scale ranging from 1 (healthy plant) to 5 (dead plant). Symbols: ▾, isolate 1127; ○, isolate 18M; •, isolate A34; ▿, isolate 275; ◊, uninoculated control. The values are the means of the values from two independent experiments, each performed with 10 plants per treatment. The error bars indicate standard deviations.

The data obtained from enzyme, Southern hybridization, and PCR analyses suggest that PG1 is widely distributed and highly conserved in different formae speciales of F. oxysporum. This is consistent with the results of previous studies which showed that there is a close structural relationship between the endoPGs of F. oxysporum f. sp. lycopersici and Fusarium moniliforme (10). Three of the 12 isolates studied (18M, A34, and 275) secreted little or no PG1. The absence of PG1 production by isolate 275 presumably resulted from a lack of the pg1 gene. Conversely, in 18M the presence of additional restriction sites in pg1, amplification of a 4.2-kb PCR fragment instead of a 1.0-kb PCR fragment with pg1-specific primers, and homology between the sequence of this fragment and the sequence of a fungal transposase indicate that the lack of PG1 in this isolate may be due to disruption of the encoding gene by a transposable element. The presence of different types of mobile genetic elements in F. oxysporum has been described previously (5). In contrast to 275 and 18M, A34 was not completely PG1 deficient, but the production of this enzyme by A34 was strongly reduced. At the gene level, neither the highly conserved 2.3-kb HindIII fragment nor any other hybridizing restriction fragment smaller than 20 kb was detected. A possible explanation for this is that hypermethylation of the surrounding DNA region makes pg1 unaccessible to methylation-sensitive restriction enzymes. This status does not affect the total genomic DNA of the isolate, since no interference with restriction digests was detected in the nit1 region. It remains to be determined whether the phenomenon observed in this study is due to hypermethylation and whether this hypermethylation affects pg1 transcription, as reported for other eucaryotic genes (15), accounting for the low level of PG1 production in this isolate.

In a previous study, the PG1 activity band was detected in stems of tomato plants infected with F. oxysporum f. sp. lycopersici, suggesting that this enzyme may play a role in pathogenicity (7). The variability of PG1 production in isolates of F. oxysporum f. sp. melonis allowed us to study the correlation between in vitro secretion of the enzyme and virulence for muskmelon. The results indicate that at least in this forma specialis, PG1 was not essential for pathogenicity since the strong PG1 producer 1127 and the deficient isolate 18M exhibited the same degree of virulence for the cultivar tested. Thus, the reduced virulence of the remaining two isolates tested was probably due to factors other than PG1 production. Since F. oxysporum produces PGs other than PG1 in vitro, we cannot rule out the possibility that PG plays a role in pathogenicity, although PG1 does not appear to be involved.