molecular biology - Why would we overexpress Sir2 by overexpressing its hypomorph (dSir2-EP2300) in C. elegans?

I have had a little time to look over this paper.

They do overexpress a native sir2 clone in high doses, called sir2.1 OE. Which seems to be native and also appears in high copy numbers. This strain was found in previous publication to have a high lifespan... that is old news.

This paper sees sir2.1 expression levels as an oversimplification of the causes of longevity. When they create crosses of the sir2.1 OE strain with wildtype, you can see the Outcross, which is verified to have a high level of sir2 expressed no longer has an extraordinary lifespan.

This can be seen in Figure 1. http://www.nature.com/nature/journal/v477/n7365/fig_tab/nature10296_F1.html

So this paper is now asking, if sir2 levels do not convey the information that creates an extended lifespan, then what does? It must be some modulation of some protein that sir2 affects. They implies that the actual cause of the lifespan increase may have been a mutation somewhere else in the organism.

"However, longevity was not suppressed by sir-2.1 RNA interference (RNAi) ... indicating causation by factors other than sir-2.1, either on mDp4 or elsewhere in the genome."

"This implies that lifespan extension is due to transgene-linked genetic effects other than the overexpression of dSir2."

In the second half of the paper (figure2) the investigators have moved on to drosophila work, where they look at how expressing constructs or inhibiting sir2 protein levels might affect lifespan. Although the dSir(EP2300) / + construct (drosophila sir2) with a wild type gene promoter did not live quite as long as dSir2(EP2300) with a fancy promoter (dSir2(EP2300) / tub-GAL4), the promoter construct (tub-GAL4/+) alone also had just as long a lifespan. How can this be? Not sure, but the expression of sir2 is clearly not the panacea we had hoped. Note however that the reduced power gene still gave the same boost to lifetime. This shows dramatically that sir2 activity alone does not drive the longer lifetime.

Lastly, deletion constructs (dSir4.5/1.7) , which should have lower than wild type protein levels for sir2 had completely normal lifetimes.

So the answer to your question; you need to test a hypothesis going in both directions - does increase of the protein really create a strong effect? Does decreasing it have a negative effect then? There are lots of other reasons to use knockouts and non functional genes in such an experiment, but those are the broad strokes.