The E symbol is used for entering numbers in scientific notation: it's short for *10^. This means that in many cases, its function is identical to that of the 10^( command (aside from the parenthesis). However, the exponent of E is limited to constant integer values ‾99 to 99.
Unlike the exponent of E, the mantissa (a special term for the A in A*10^B, in scientific notation) isn't limited in variable type: it can be a constant, a real or complex variable or expression, a list, a matrix, or even omitted entirely (and then it will be assumed to equal 1). The reason for this versatility is simple: internally, only the exponent is taken to be an actual argument for this command. The rest of the calculation is done through implied multiplication.
E99 and -E99 are often used for negative and positive infinity because the TI-83 series of calculators doesn't have an infinity symbol. Commands that often need to use infinity include solve(, fnInt(, normalcdf( (and the other distributions), and many others. The error introduced in this way is usually irrelevant, because it's less than the minimum calculator precision, anyway: E99 is mindbogglingly huge.
Don't add the mantissa when it's 1:
In addition, E2 or E3 can be used as shorthand ways of writing 100 and 1000 respectively. This could be continued, in theory, for higher powers of 10, but those aren't necessary as often.
E is much faster than using the 10^( command or typing out 10^. The drawback, of course, is that it's limited to constant values.