No. Why would it be universal?

'Universal' would mean present in all languages.

Anyway, in languages with i-mutation the change a→e is quite common. However, it's not so much the sound change but the actual underlying process that is important. Many languages can share the change a→e, but it's the reason why that is of interest.

Well, i-mutation is caused by i/j in nearby (or same) syllables, causing a vowel to move further forward, i.e. to be fronted. The equivalent sound change for consonants would be palatalisation, as consonants themselves cannot be 'fronted' as such. This because i/j to varying degrees possess the qualities 'front' and 'palatal', which can colour nearby sounds.

E.g.

back [ɑ] → centre [ɜ] → front [ɛ]
back [ɔ] → centre [ɞ] → front [œ]
back [u] → centre [ʉ] → front [y]
front low [æ] → front mid-low [ɛ] → front mid-high [e] → front high [i]

As can be seen from the examples, i-mutation commonly causes non-front vowels to be fronted, and front vowels to be raised — they move closer towards the sound i.

This is not the only reason [a] might change to [e], but it is a fairly common change, and probably quite well studied.

Can this change happen the otherway,that is from  '|e|'to '|a|'?

Yes,Why it happens is an interesting question.Can you suggest some literature that shed some light on it?

Actually answering "why" is all but impossible. But we can talk in more general terms than just [a] becoming [e]. We can look at how this relates to more general patterns. For example, is it the only change at that time? Was the [e] space previously unfilled? If we look at Arabic dialects that previously had just three vowels and now have five, we note that diphthongs are more articulatory work and then they (ai, au) settled into the mid vowel space (e, o). That's not entirely a "why" answer, but it's getting closer.

Another way of looking at what freknu said is that it's entirely irrelevant whether one particular vowel becomes another particular vowel. It might be a little interesting for people who like languages, but beyond that it's of absolutely no value to know the details. It is more interesting to know the FACT that a vowel changed to a different vowel. And it is also interesting to know which vowels most often become which other vowels. Generally, as I said, they're not too far off in the articulatory space and they may be open spaces (or might result in a push chain where what was there also moves somewhere else).

Another way to look at this would be to ask WHEN this happens (not WHY). Are there any typological patterns to that? This is a kind of fronting. It probably often occurs in cases of vowel harmony (see Turkish). Maybe other cases too. That tells us something about the interaction between sounds and how they're represented as mental categories (involving, for example, phonological features).

Most stochastic models of diachronic phonology offer fairly compelling explanations of why changes happen. There's no one size fits all answer for *e->a, but (like freknu suggests) stress patterns are going to be a good first place to look for clues.