Studies published in Nature typically have flashy science. Coupled with the fact that their relatively short format makes it difficult to determine what exactly was done, much less replicate it, I'm actually less inclined to believe something in Nature as compared to, for example, Cell or PNAS.

It's also worth considering that there's a positive correlation between impact factor and retraction rates. http://retractionwatch.wordpress.com/2011/08/11/is-it-time-f...

All in all, while Nature doesn't publish obviously bad science, it's less clear that it publishes better science than average.