Read Edit:
You are right! The definition of potential given by
$$V(\mathbf{r})\equiv -\int_{\mathcal{O}}^\mathbf{r}\mathbf{E}\cdot d\mathbf{l}$$
In your case the $\mathbf{E}$ is outward in x-direction while the displacement vector $d\mathbf{l}=\mathbf{x}-(\mathbf{x}+d\mathbf{x})=-d\mathbf{x}$. So yes the dot product between them will be negative.}
The following possibility may occur because they didn't give the value of $k$ it might be positive or negative, so may be they have taken it to be negative. Apart from this your reasoning is correct.
Edit : Actually there is flaw in this thinking. While we derive the expression for line intergral we don't care for the sign of $\Delta x$ because that already taken into account with limits of integration.For example : consider the example given below in the comment box.
So it's as given in the book.