OUP user menu

Telomeres in the rice blast fungus Magnaporthe oryzae : the world of the end as we know it

Mark L. Farman
DOI: http://dx.doi.org/10.1111/j.1574-6968.2007.00812.x 125-132 First published online: 1 August 2007


The subtelomeres of many microbial eukaryotes are highly enriched in genes with roles in niche adaptation. Host and cultivar specificity genes in the rice blast fungus Magnaporthe oryzae also tend to be located near telomeres. In addition, the M. oryzae telomeres are highly variable chromosome regions. These observations suggested that plant pathogenic fungi might also use subtelomere regions for the amplification of genes with adaptive significance. Targeted sequencing of the M. oryzae telomeres provided an opportunity to test this hypothesis, and has yielded valuable insights into the organization and dynamics of these important chromosome regions.

  • helicases
  • transposons
  • adaptation


Telomeres are the sequences that form the ends of linear chromosomes and are essential for preventing the loss of DNA during DNA synthesis, and from exonucleolytic attack. In most organisms, the telomeres consist of short repeated sequence motifs that are added onto the chromosome ends by the enzyme telomerase (Zakian, 1996). The chromosomal sequences that occur adjacent to the telomere repeats are often duplicated at multiple chromosome ends and, as such, they define a specific subtelomere region. Usually, two major subtelomere domains are recognizable: the distal domains, which contain sequences that are duplicated at most and sometimes all chromosome ends, and the proximal domains, which usually consist of longer segments of duplicated DNA but are less widely dispersed among ends (Flint et al., 1997; Pryde et al., 1997; Nickles & McEachern, 2004).

In many microbial eukaryotes, the proximal subtelomere domains are highly enriched in genes involved in niche adaptation. The Saccharomyces cerevisiae subtelomeres contain genes for sugar catabolism, and the specific genes that are amplified in these regions vary depending on the strain's ecology (Ness & Aigle, 1995; Denayrolles et al., 1997). In human pathogens, such as Plasmodium falciparum, Trypanosoma brucei and Pneumocystis carinii, the subtelomeres contain large families of highly variable genes coding for antigenic surface proteins that would normally trigger the immune system. However, these pathogens stay one step ahead of their host's defenses by expressing only one gene at a time, and by periodically switching the copy that is expressed (Barry et al., 1998). It is believed that the subtelomeric localization contributes to this host evasion strategy by promoting the amplification and diversification of surface protein genes. It may also facilitate the switching process by promoting the recombination of silent genes into active expression sites (Wada & Nakamura, 1996; Donelson et al., 1998; Robinson et al., 1999; Barry et al., 2003) and/or by repressing gene expression through a telomere position effect (De Las Penas et al., 2003; Duraisingh et al., 2005; Freitas-Junior et al., 2005).

Magnaporthe oryzae genes controlling host and cultivar specificity are frequently found near telomeres

The fungus M. oryzae (formerly Magnaporthe grisea) causes a devastating disease of rice known as rice blast, and is a serious pathogen of other cereals and grasses. M. oryzae exhibits a high degree of host specificity, due to avirulence genes that code for proteins that trigger resistance in host plants with corresponding resistance genes. Two of the first avirulence genes to be studied, Avr2-YAMO (=AVR-Pita) and Avr1-TSUY, exhibited genetic instability, which led to the frequent appearance of virulent mutants. Intriguingly, both genes mapped to telomeric locations (Valent & Chumley, 1994; Orbach et al., 2000). Subsequently, the cloning of AVR-Pita revealed that the gene is located just 48 bp from the telomere repeats, and analysis of spontaneous virulent mutants showed that they frequently result from terminal deletions (Orbach et al., 2000). Additional evidence for telomere involvement in AVR gene instability in M. oryzae comes from the finding that gains of virulence to rice cultivar Tsuyake are associated with deletions of a specific telomere in isolate O-137 (Kang, 2001). However, the Avr1-TSUY gene remains to be cloned, and so the exact basis for its instability is not yet known (S. Kang, pers. commun.).

Despite these intriguing associations between AVR gene instability and telomere proximity, this relationship is not absolute because the Pwl2 gene is highly mutable, and yet maps in the middle of a chromosome (Sweigard et al., 1995). Conversely, several newly mapped AVR genes are located very close to telomeres but do not show high levels of spontaneous mutation (Table 1, and D. Tharreau, pers. commun.). One possibility is that the more recently mapped genes reside further from the telomere than AVR-Pita and AVR1-TSUY do, and are, therefore, less susceptible to terminal truncation. Hopefully, the cloning of additional AVR genes will shed new light on this matter.

View this table:
Table 1

Map locations of Magnaporthe oryzae avirulence genes

Avirulence geneStable?References
    Avr-K59Wang et al. (2005)
    Avr1-Ku86Dioh et al. (2000)
    Avr-MedNoïDioh et al. (2000)
    AvrPi15Ma et al. (2006)
    Avr-PiaChen et al. (2006a)
    Avr-PitaNoOrbach et al. (2000)
    Avr-PizYesLuo et al. (2005)
    Avr-PitWang et al. (2005)
    Avr–TSUY/Avr-C103TTPNoWang et al. (2005)
    Avr-Zh156Wang et al. (2005)
    Pwl1YesValent & Chumley (1994); Kang et al. (1995)
    Avr-C105TTPWang et al. (2005)
    AVR1-C101TTP5L23YesC. Kaye and D. Tharreau, pers. commun.
    AVR1-CO39Smith & Leong (1994); Farman & Leong (1998)
    Avr–IRAT7/ACE1Dioh et al. (2000), Bohnert et al. (2004)
    Avr1-MARAValent & Chumley (1994); Mandel et al. (1997)
    Avr-PikYesLuo et al. (2005)
    Avr-Piz-tYesLuo et al. (2005)
    Avr1–PTB25YesC. Kaye and D. Tharreau, pers. commun.
    Avr-Zuan1641Wang et al. (2005)
    PRE1Chen et al. (2006b)
    Pwl2NoSweigard et al. (1995)
    Pwl3Kang et al. (1995)
    Pwl4Kang et al. (1995)
  • A ‘–’ symbol indicates that this information was not reported.

Telomere-linked helicase genes

In an effort to isolate Avr1-TSUY, Kang and co-workers cloned and characterized two telomeric EcoRI fragments (Kang, 2001; Gao et al., 2002). One fragment contained a previously unidentified retrotransposon. The other contained a gene coding for a RecQ helicase protein with strong similarity to conceptual proteins encoded by the telomere-linked UTASa genes in Ustilago maydis (Sanchez-Alonso & Guzman, 1998). Southern hybridization studies showed that the M. oryzae RecQ helicase gene belongs to a gene family, and the cloning of two additional family members, as well as physical mapping, indicated that all of the genes in strains O-137 and Guy11 are within 10 kb of a telomere (Gao et al., 2002). For this reason, they were termed telomere-linked helicase (TLH) genes. Interestingly, the TLH genes were found to have a limited distribution among different host-specific forms of M. oryzae, being found almost exclusively in isolates from rice.

Telomere instability

Gao (2002) analyzed TLH gene stability through several single-spore generations and showed that deletions and amplifications can be detected even in the absence of selection. Rearrangements were also found when telomeric RFLPs were used for genetic mapping in a cross between two isolates that are pathogenic to rice (Farman & Leong, 1995). These findings suggested that rice-infecting M. oryzae isolates have highly dynamic telomeres. However, telomere fingerprinting of field isolates reveals very limited telomere variation among isolates collected from dispersed locations (an example is shown in Fig. 1a). This indicates that most telomeres are quite stable, or that purifying selection prevents the recovery of rearrangements at certain chromosome ends.

Figure 1

Telomere variation detected by Southern hybridization using a (TTAGGG)250–300 probe. (a) Telomere fingerprints in a collection of Magnaporthe oryzae isolates collected from rice fields in Arkansas. ‘a, b, c and d’ indicate race designations. (b) Changes in telomere fingerprints among single spores from a single culture of M. oryzae from perennial ryegrass. LpKY97-1 is the starting culture, SSA1_1 was derived from a single spore sampled from a plate culture of LpKY97-1. SSP1_2 and SSP1_5 were single spores collected from lesions in the first round of infection of perennial ryegrass plants. SSP2_4 and SSP2_6 were collected after a second round of infection. Black circles show ‘incipient’ telomeres that arose subsequent to single sporing.

In contrast, M. oryzae strains from perennial ryegrass (PRG) exhibit extreme telomere variability such that independent isolates usually have completely different telomere fingerprints, even when there are few detectable polymorphisms at internal loci (Farman & Kim, 2005). New telomere rearrangements also occur during routine culture of PRG-infecting isolates and during growth in planta (Farman & Kim, 2005). Analysis of single-spore cultures collected after two rounds of infection revealed isolates in which as many as eight of the 14 telomeres were rearranged (Fig. 1b). Therefore, it appears that completely new telomere profiles could arise after just a few disease cycles in the field. This level of telomere instability is unprecedented and, given the association between AVR genes and telomere regions, raises questions as to the possible adaptive significance of this phenomenon.

New insights into M. oryzae telomere organization and dynamics gained by sequencing chromosome ends

To explore the relationship between telomere variability and pathogenic adaptation, we sequenced the 14 telomeres of 70–15, the M. oryzae strain used for the genome sequence project (Dean et al., 2005), were sequenced. This yielded a total of 575 kb of sequence information, and contributed more than 270 kb of new sequence data that were not present in the original genome assembly.

Organization of chromosome ends in M. oryzae isolates from rice

In strain 70–15, one chromosome end consists of the ribosomal DNA array flanked directly by the telomere repeats, two ends lack any telomere-specific sequence features, and 11 ends contain a sequence that qualifies as a distinct subtelomere domain, due to its presence at multiple chromosome ends (sensu Pryde et al., 1997). The basic M. oryzae subtelomere unit contains the telomere-linked RecQ helicase (TLH) gene, another putative gene of unknown function, and several blocks of short tandem repeats (Fig. 2a). The subtelomere has two discrete boundaries, beyond which are sequences that are specific to the individual chromosome ends — the so-called ‘chromosome-unique’ sequences. At most ends, the transition to unique sequences occurs ∼1 kb upstream of the TLH gene, at the ‘near’ boundary, while others have an extended subtelomere that contains an additional gene and two transposon insertions, and diverges at a ‘far’ boundary located ∼10 kb upstream of the TLH start codon. The distal portion of the subtelomere is highly prone to truncation, and so the amount of subtelomere sequence downstream of the TLH genes varies among chromosome ends (Rehmeyer et al., 2006). Sequencing of telomeres in additional M. oryzae isolates (C. Rehmeyer, Y.-S. Kim and M. Farman, unpublished data), and comparison with the sequences reported by Gao (2002), indicate that this subtelomere structure is conserved among rice-infecting isolates.

Figure 2

Structural organization of subtelomeres in Magnaporthe oryzae isolates from rice and perennial ryegrass. The chromosome termini are drawn with the telomere repeats on the right-hand side (black circles), and the chromosome-unique sequences on the left (solid and dotted lines). The subtelomere regions are represented as rectangles. (a) Orange areas represent helicase-associated repeats (HARs), the red regions corresponds to an MgSINE element and the pink area represents a short repeat element. The TLH genes and a RETRO6-1 element are labeled in the figure. (b) MoTER elements are shown as colored rectangles. The orientations of the MoTER-coding regions are shown with arrows. The small circles represent short TTAGGG motifs that separate the MoTER elements from the chromosome-unique sequences and from each other. The open square at the 5′ end of MoTER2 is a variant telomere repeat sequence ([TTTGGG]8 [TTCGGG]2). The MoTER1 in the middle of the tandem array shown in the bottom diagram is a 5′-truncated element.

Strain 70–15 has four intact TLH genes and seven copies that apparently have been destroyed by mutation, truncation or transposon insertion. A multiple alignment revealed numerous point mutations in the TLH coding region, with a vast predominance of G–A and C–T transitions, implying that a RIP-like mechanism (Cambareri et al., 1989) was largely responsible for the TLH sequence divergence (C. Rehmeyer & M. Farman, unpublished data). A consensus gene prediction indicated that the full length of the TLH gene is 4638 bp — almost twice the length of the published TLH1 gene (Gao et al., 2002) (Genbank accession: AY077623). Subsequent inspection of the TLH1 sequence revealed a frameshift mutation that caused the gene to be split into two smaller ORFs. The TLH1 truncation apparently was not detected because the predicted ORF encoded all of the expected RecQ helicase domains.

The full-length TLH gene codes for a RecQ helicase with a long N-terminal extension. TLH proteins identified in Schizosaccharomyces pombe (Mandell et al., 2005) and Metarrhizium anisopliae (Inglis et al., 2005) have a similar extension that exhibits weak similarity with the M. oryzae TLH N-terminus, except for a highly conserved domain spanning ∼100 amino acids. Searches of the protein databases, as well as of fungal genome sequences, revealed that this domain is exclusive to fungi and almost always belongs to a helicase encoded by a subtelomeric gene (C. Rehmeyer, W. Li and M. Farman, in preparation). This finding suggests the subtelomerically encoded helicases have a function that is distinct from other RecQ helicases. Currently, the functions of the TLH genes are unknown. However, the S. pombe genes are expressed when the telomere is compromised (Mandell et al., 2004), and overexpression of a carboxy-terminal portion of the TLH protein can increase longevity in S. pombe cells lacking telomerase. This suggests a role for the TLH genes in telomere maintenance, but their precise function(s) in this process remains elusive. Initially we suspected that they were responsible for the increased telomere stability in the rice pathogens. However, this does not appear to be the case because M. oryzae isolates from foxtails, and M. oryzae isolates from crabgrass, have highly stable telomeres and yet they lack TLH genes (M. Farman, unpublished results).

Gene content in the terminal regions of the M. oryzae chromosomes

The terminal chromosome regions contained in the fosmid clones exhibited an overall gene density that is similar to the rest of the M. oryzae genome. Only two of the terminal genes encode predicted proteins that are likely to have roles in pathogenicity — these being cellulases. Three chromosome ends contain secondary metabolism gene clusters (Rehmeyer et al., 2006) that are sometimes involved in the production of toxins that contribute to plant pathogenesis (Panaccione et al., 1992; Proctor et al., 1995). However, the terminally located clusters in M. oryzae appear to be incomplete because they lack key genes and, therefore, are unlikely to be capable of toxin production. Consequently, the sequence data provide very little indication that the M. oryzae chromosome ends are enriched in pathogenicity genes. Nevertheless, these regions do contain a large number of genes coding for hypothetical and predicted proteins whose functions are unknown. Therefore, the relationship between telomere instability and pathogenic adaptation will need to be further addressed through comprehensive functional studies.

One thing that is clear, however, is that very few of the terminal genes in 70–15 are duplicated. This sets M. oryzae apart from other pathogenic eukaryotes in terms of its subtelomere organization, and makes it unlikely that this fungus uses switching mechanisms to evade the host defenses.

Despite the frequent association between AVR genes and telomere regions, there was also no evidence for the existence of telomere-linked AVR genes in 70–15. None of the terminal genes showed similarity to known AVR genes, nor do they encode cysteine-rich proteins — a common hallmark of AVR factors (Rohe et al., 1995; de Wit & Joosten, 1999; Rep et al., 2004). The terminal chromosome regions in 70–15 are also not enriched in genes coding for secreted proteins. Hence, although AVR genes are often found near telomeres in different isolates of M. oryzae, the current sequence data suggest that they do not exist in abundance in these regions.

Telomere rearrangements

Analysis of the genome sequence data revealed a number of novel telomeres that had arisen in the fungal culture used for sequencing. In one case, a TLH gene and the surrounding subtelomere sequence was lost due to a large (∼35 kb) deletion, while the other instances involved truncation of the rRNA gene repeat array joined to telomere 3. The exact sizes of the rRNA gene deletions were not determined but one was at least 20 kb in length (C. Rehmeyer & M. Farman, unpublished results).

The structures of the chromosome ends also suggested evolutionary histories rife with rearrangements. At least 10 of the TLH-containing chromosome ends showed evidence of past terminal truncations (Rehmeyer et al., 2006) indicating that the M. oryzae chromosome ends experience chronic telomere crisis, which can result in the loss of several kilobases of terminal sequence. In addition, the particular 70–15 culture that was sequenced had a novel version of telomere 14 that was not present in an archival culture. Analysis of the new telomere revealed that it contains a copy of an internal sequence that was captured from a location ∼500 kb from the chromosome end, and moved to a position distal to a TLH gene (C. Rehmeyer & M. Farman, unpublished results).

The chromosome-unique regions that are adjacent to the subtelomeres contain numerous transposon sequences whose specific organization indicates that, at one time, they experienced ectopic recombination with another copy (Rehmeyer et al., 2006). Such events frequently cause loss of intervening sequences and, therefore, one would expect any AVR genes in the terminal regions to be prone to deletion. Considering the stochastic nature of transposition and insertion processes, this could generate significant strain-to-strain variation in AVR gene content. Such a situation would clearly be advantageous for M. oryzae because AVR gene polymorphism will increase the virulence spectrum of the pathogen population. Therefore, it may be that the tendency for AVR genes to occur in telomere-linked locations is because these chromosomal regions experience frequent, yet random, mutations that generate significant AVR gene polymorphism within the pathogen population. In addition, the fact that M. oryzae telomeres are capable of capturing internal sequences indicates that internal genes can be moved close to telomeres, where they can benefit from the dynamic nature of the terminal regions.

Telomere-targeted retrotransposons

To investigate why the PRG pathogen telomeres are so much more unstable than those in the rice-infecting strains, telomeres from three different PRG-infecting isolates were sequenced. This revealed a subtelomere structure that was completely different from what was found in the rice-infecting isolates. Instead of a TLH-containing region, the PRG pathogens’ subtelomeres consisted of two types of repetitive elements (Fig. 2b). Both elements are found exclusively at telomeres and for this reason, they are referred to as MoTERs (M. oryzae Telomere Exclusive Repeats). MoTER1 is 4.6 kb in length and codes for a reverse transcriptase. It lacks long terminal repeats and, therefore, resembles non-LTR retrotransposons. MoTER2 is smaller (only 1.7 kb) and encodes a hypothetical protein of 204 amino acids. Some chromosome ends contain a single MoTER element separating the telomeres from the chromosome-unique sequences, while others possess tandem arrays, comprised of single elements or mixtures of the two (Fig. 2b). Several truncated copies of MoTER1 were identified in the arrays but MoTER2 always seems to be intact. All MoTER copies are found in the same orientation — with their 5′-end directed toward the telomere, and each element is separated from the next by less than one to three copies of the TTAGGG telomere motif.

Telomere-specific retrotransposons have been identified in three other organisms. The best-known example is Drosophila, which lacks the short tandem repeats that cap the chromosome ends of most organisms. Instead, telomere functions are assumed by the non-LTR retrotransposons TART and Het-A, which periodically transpose onto the chromosome ends, thereby preventing the loss of terminal DNA that accompanies replication (Pardue et al., 2005). The protozoon Giardia lamblia has two retroelements, GilM and GilT, that also exist in tandem arrays at its chromosome ends. However, unlike Drosophila, G. lamblia has canonical telomere repeats at the ends of the retroelement arrays (Arkhipova & Morrison, 2001). This organization suggests that GilM and GilT transpose onto degraded ends during periods of telomere crisis, with canonical telomeres being reconstituted later on (Arkhipova & Morrison, 2001). In the silkworm, Bombyx mori, the telomere arrays contain retrotransposon insertions. These elements belong to two major families, TRAS and SART (Okazaki et al., 1995; Takahashi et al., 1997), and encode endonucleases that cleave within the TTAGG array, thereby providing target sites for insertion (Anzai et al., 2001).

The idea that MoTERs are active transposons was particularly attractive because it could potentially explain why the PRG pathogen's telomeres are so unstable. Therefore, to determine whether MoTER transposition is responsible for the observed telomere rearrangements, a number of novel telomere fragments that arose during the culture of PRG-infecting strain LpKY97-1 were cloned and characterized. This led to the identification of two new MoTER1 insertions at chromosome ends that originally lacked MoTER elements, thereby demonstrating that MoTER1 is an active transposon (J. Starnes & M. Farman, unpublished data). In both cases, the new insertions were within the existing telomere repeat array. At present, it is not possible to tell whether the MoTER elements insert into intact telomere arrays, as in B. mori, or whether the PRG pathogens experience chronic telomere crisis with MoTERs adding onto the ends of critically short telomeres, providing temporary protection until the canonical telomere is reconstituted.

In some isolates, the MoTER:chromosome-unique sequence boundaries are associated with large (>20 kb) terminal deletions encompassing several genes. This suggests that the MoTER elements can destabilize the chromosome termini, making them prone to degradation and loss of terminal genes. It is also possible that MoTER transposition can alter the expression of neighboring genes, as has been shown for Het-A insertions in Drosophila (Golubovsky et al., 2001; Mason et al., 2003). Whether these activities have any impact on the pathogenic capabilities of the fungus will need to be addressed through functional studies of the genes that reside at the chromosome ends of the ryegrass pathogens. Interestingly, it is worth noting that several of these genes are not present in 70–15.

Concluding remarks

The fact that so many AVR genes map near telomeres originally led to the speculation that the M. oryzae chromosome ends might be organized in a manner similar to the subtelomeres of other eukaryotic pathogens, being highly enriched in genes coding for surface proteins that tend to trigger host defenses. However, the sequencing of chromosome ends of one strain of M. oryzae failed to provide evidence in support of such an arrangement. Specifically, the paucity of duplicated genes in the terminal regions indicates that M. oryzae does not use subtelomeres to amplify genes involved in interactions with the host, nor are the terminal regions particularly enriched for genes that encode secreted proteins. Indeed, it is possible that the sequenced regions of the 70–15 chromosome termini are devoid of AVR genes altogether.

If the organization of terminal regions in 70–15 is a faithful representation of what occurs in most rice-pathogenic isolates, it seems unlikely that the tendency for AVR genes to occur near chromosome ends is due to the existence there of active systems for host defense evasion. Instead, it appears that this situation may arise simply through selection for AVR gene diversity within the pathogen population, and that this occurs most often when the AVR genes are in terminal locations.

In retrospect, it is perhaps not all that surprising that the M. oryzae subtelomeres lack active systems for host evasion. Unlike humans, who have a dynamic immune response that requires the pathogen to undergo antigenic switching if it is to remain undetected, the rice plant's resistance system is relatively static — there is no evidence that the plant's surveillance system ‘evolves’ during the course of an infection. Therefore, if the fungus can enter the plant without triggering defenses, it should not need to alter its secreted protein profile to remain undetected. Consequently, rice-infecting isolates of M. oryzae may lack large, subtelomeric families of surface protein genes because there is little pressure to change their ‘coats.’

The ecology of M. oryzae may be another reason why its subtelomeres are not dedicated largely to the amplification of surface protein genes. Plasmodium falciparum, T. brucei and P. carinii are obligate parasites, and so all stages of their life cycles require association with living hosts. As such, the predominant challenge to pathogen survival is to avoid recognition, and subsequent eradication by the host's immune system. M. oryzae, on the other hand, is a facultative pathogen that is also capable of growing on nonliving substrates. Indeed, its life cycle is completed through spore production, which usually occurs on dead or dying tissues. Without the shelter of a living host, the fungus probably has to contend with a variety of environmental insults from which human pathogens would be insulated. In addition, sporulating lesions are invariably co-colonized by a variety of other fungi and bacteria. Thus, in its natural environment, the fungus likely has to counter not only host defenses but must also compete with microbial co-colonizers for resources, and may also need to defend itself against antibiotics produced by these organisms. Therefore, if the M. oryzae subtelomeres do have roles in adaptation, their gene content probably represents a solution to a number of ecological challenges.


The author acknowledges David Thornbury for technical assistance and Lisa Vaillancourt for commenting on this review. The author also thanks Jim Correll and Ed Boza for providing a Southern blot (Fig. 1a), and Claudia Kaye and Didier Tharreau for sharing unpublished data. The research was supported by the National Science Foundation (MCB 0135462), National Center for Research Resources (5P20RR016481-03) and United States Department of Agriculture-Cooperative State Research, Education and Extension Service special grant 2001-34457-10343.


  • Editor: Richard Staples


View Abstract