A new tool for functional genomics in maize


Maize provides one of the fundamental sources of nutrition for the world’s population. Continued improvement in the performance of maize is essential to meet the needs as the global population increases. In many ways maize research is poised for a wave of progress as its rich genetic diversity can now be more effectively exploited through resources such as the recently completed maize B73 genome sequence (Schnable et al., 2009), other genome sequences that are nearing completion (e.g. Mo17 at http://www.phytozome.net/maize and Palomero Toluqueno http://www.palomerotoluqueno.org/) and the 5000-line nested association mapping population (McMullen et al., 2009). However, utilization of genome sequence and the linking of DNA sequence to function often require research tools that have been lacking in maize, in which it is particularly laborious to generate transgenic plants. In this issue of New Phytologist, van der Linde et al. (pp. 471–483) report significant progress that should facilitate the process of establishing the function of maize genes.

‘The development of efficient VIGS tools for maize is a very significant advance for maize research.’

In van der Linde et al.’s report, a virus-induced gene silencing (VIGS) system is utilized to assess whether candidate genes have essential functions in determining the outcome of interactions between maize and the biotrophic fungal pathogen Ustilago maydis, the causal agent of corn smut. In addition to being a major pathogen of maize, the U. maydis–maize interaction is one of the best-developed model fungal pathosystems of maize. During the initial 12 h of infection, defense genes are highly expressed in both compatible and incompatible U. maydis–maize interactions (Doehlemann et al., 2008, 2009). At the 24 h time-point during the compatible interaction, when the biotrophic interface has been established, expression of defense genes is suppressed and expression of the well-known suppressor of cell death, Bax-inhibitor 1 (BI1), is strongly induced (Doehlemann et al., 2008). By contrast, in the incompatible interaction with a mutant strain of U. maydis lacking the effector protein, Pep1, suppression of the defense genes does not occur, the pathogen is unable to complete penetration of the epidermis and a massive defense response is triggered with the collapse of the infected cells (Doehlemann et al., 2009).

A strong hypothesis for how changes in gene expression may determine the outcome of U. maydis–maize interactions can be derived from these data. However, ultimately the functions of these genes must be tested in vivo. To accomplish this, van der Linde et al. adapted the Brome mosaic virus (BMV) VIGS system, first described by Ding et al. (2006), for use in the U. maydis pathosystem. Key questions needing answers were: Can conditions be established that will support sufficient development of BMV-VIGS as well as U. maydis infection? Where in the plant will maximal and the most consistent VIGS occur? Will the U. maydis–maize interaction occur normally in BMV-infected plants, or will the viral infection per se perturb the outcome of the fungal interaction?

This research team was able to find workable solutions to these issues. While it is the pathogenesis of the maize ear that causes the major disease problem of corn smut, protocols for silencing in the ear are not yet developed for BMV-VIGS. Fortunately, the fungus is able to induce tumors and grow on all parts of the maize plant and so the researchers performed the VIGS assays using leaves of maize seedlings. When utilizing a phytoene desaturase silencing construct, the researchers found that gene silencing occurred in the fourth, fifth and sixth leaves; however, it was only consistent in the sixth leaves. They assert that an important step for obtaining uniform silencing was to standardize the viral-inoculum titer across all constructs and experiments. In this case they did this by first infecting Nicotiana benthamina with the in vitro transcripts necessary to initiate BMV-VIGS and then used quantitative reverse transcription–polymerase chain reaction (qRT-PCR) to measure the BMV titer in the sap prepared from these plants. These relative quantifications were then used to dilute the virions to a uniform titer for infection of maize. They found that initiating VIGS in seedlings grown at 22°C and then shifting the temperature to 28°C for U. maydis inoculation was conducive to the development of efficient VIGS as well as to fungal growth. Additionally, to assess the effect of BMV-VIGS silencing on U. maydis pathogenesis, they performed experiments with a yellow fluorescent protein (YFP) silencing construct on transgenic maize that constitutively expressed YFP. These crucial controls demonstrated that no statistically significant difference in the rate of tumor formation was observed between BMV-infected and control plants.

When the researchers assessed the possible functions of their candidate genes in the U. maydis–maize interaction they developed a useful quantitative polymerase chain reaction (qPCR) strategy to quantify the degree of maize gene silencing and fungal growth from DNA and RNA prepared from the same maize tissue. With this qRT-PCR-based assay to determine the outcome of the pathology assay and plant gene silencing in place, they assessed whether silencing Bax inhibitor 1 (bi-1), an endochitinase B (ecb), terpene synthase 6/11 (tps6/11), or a Bowmann–Birk type trypsin inhibitor (bti) had detectable effects on the outcomes of the U. maydis–maize interaction. Two of the genes –ecb and bti– were effectively knocked-down by BMV-VIGS, with no significant effect on fungal growth. However, in the cases of bi-1 and tps6/11, significant alterations in the pathology assay were observed in plants with high levels of gene silencing. Interestingly, opposing effects in the pathology assay were observed: bi-1 silencing caused plants to become less susceptible, whereas tps6/11 silencing resulted in increased fungal growth. At the level of the pathogenesis mechanism, these results indicate that blocking the inhibition of cell death through silencing of bi-1 thwarts the growth of the biotrophic fungus, while it would appear that tps6/11 may defend the plant against U. maydis.

To conclude this impressive study, the researchers demonstrated that two transgenic maize lines expressing a tps6/11 RNA interference (RNAi) construct are also more susceptible to U. maydis, consistent with the VIGS result. Taken together, a great deal of groundwork has been accomplished in this report so that other maize researchers should be able to begin to adapt the BMV-VIGS system for their own applications.

Even though this report represents a significant advance in functional gene studies in maize, the BMV-VIGS system still has significant limitations. Maize is an extremely diverse crop (Buckler et al., 2006) and much of the power of maize genetics derives from harnessing this diversity. Importantly, the applicability of the BMV-VIGS system has only been demonstrated in one specific maize genotype, Va35. Different maize lines can display quite contrasting responses to BMV infection. For instance, in B73, BMV infection causes systemic necrosis and ultimately leads to plant death, whereas in Va35, the viral symptoms consist of chlorotic streaks (Ding et al., 2001). If an investigator’s allele of interest is not present in a BMV-VIGS compatible line, as it stands it would preclude the investigation of its putative role in the maize–pathogen interaction. In the study of van der Linde et al., U. maydis could successfully establish an infection in Va35, which allowed the roles of potential candidate genes to be assessed.

The availability of a greater range of VIGS-compatible maize lines should greatly improve the utility of VIGS in maize. The development of any effective VIGS system must strike a fine balance between the aggressiveness of viral infection and the plant’s ability to defend itself by targeting the viral genome for degradation by its silencing machinery. An efficient silencing response is critical for successful VIGS, but this may cause a reduced level of viral infection and limit the spreading of VIGS. On the other hand, a strong infection may lead to confounding symptoms as a result of viral disease, or even plant death. Presently, taking into consideration the limited understanding of plant–virus interactions and particularly its players at the molecular level, the identification of additional VIGS-efficient maize lines may be the quickest and shortest route to speed up efforts to broaden the applicability of the BMV-VIGS system in maize.

A few other aspects of the BMV-VIGS system in maize require further consideration. Silencing in Va35 is transient and has only been demonstrated in leaves, where it displayed a patchy pattern. A similar result was described for the YFP transgenic line used by van der Linde et al. It is unclear if phenotypes in other organs of the maize plant are accessible to VIGS using BMV. The transient nature of the silencing phenotype is probably correlated to insert instability, which has been documented by others in both the BMV system and the barley stripe mosaic virus (BSMV)-VIGS (Scofield et al., 2005; Ding et al., 2006; Brunn-Rasmussen et al., 2007). Curiously, in rice, the efficiency and stability of the silencing phenotype induced by BMV-VIGS has also been shown to vary among the two tested target genes – phytoene desaturase and actin1 (Ding et al., 2006). Although the mechanisms causing this difference are not clear, one might expect that, in maize, silencing effectiveness using BMV-VIGS may also differ between target genes, and perhaps even between different fragments of the same target gene used for generating the silencing constructs.

The current BMV-VIGS system relies on the production of capped in vitro transcripts, which are expensive to generate. Adaptation of the current system to Agrobacterium binary-based vectors for viral inoculation would be a step towards a system that would allow a cost-effective, high-throughput analysis of multiple candidate genes. Alternative methods of viral inoculation in maize could also be explored. For instance, direct inoculation into the coleoptilar node of Agrobacteria carrying T-DNA constructs that initiate infection of maize streak virus (MSV) is routinely used to screen different maize lines for MSV resistance (Grimsley et al., 1987; Martin et al., 1999). Alternatively, a vascular puncture inoculation method has also been successfully used to inoculate many different maize viruses recalcitrant to traditional inoculation procedures, such as leaf rubbing (Redinbaugh et al., 2001).

The development of efficient VIGS tools for maize is a very significant advance in maize research. The availability of functional genomics tools for studying maize geneticists’ favorite candidate genes, together with all the powerful resources available for maize genetics, brings the promise of accelerated discoveries.