There are different ways to model a straight line $L$ through the origin.
One way is using a direction vector $d=(d_1, d_2)$ (which can be any vector different from the null vector) and have the points of the line created by all endpoints of the scalar multiples of $d$:
$$
L = \{ t d \mid t \in \mathbb{R} \}
$$
In two dimensions we can describe the lines by a normal vector $n$ of the line, the endpoints of all vectors perpendicular to that normal vector form the line:
\begin{align}
L
&= \{ u \mid u \perp n \} \\
&= \{ u \mid n \cdot u = 0 \} \\
&= \{ u = (x, y) \mid n_x x + n_y y = 0 \}
\end{align}
The description as functions
$$
y(x) = m x
$$
is slightly less general, as a vertical line would need infinite slope.
A rotation of a vector in the plane by an angle $\alpha$ around the origin can be described by a rotation matrix
$$
R_\alpha=
\begin{pmatrix}
\cos \alpha & -\sin \alpha \\
\sin \alpha & \cos \alpha
\end{pmatrix}
$$
Its first column is the rotated vector $(1,0)^T$, its second column the rotated vector $(0,1)^T$.
For the above descriptions it could be applied to a direction vector and its scalar multiples or to a normal vector of a line.
For $\alpha=45^\circ$ it would be
$$
R =
\begin{pmatrix}
1/\sqrt{2} & -1/\sqrt{2} \\
1/\sqrt{2} & 1/\sqrt{2}
\end{pmatrix}
$$
and applied as
$$
f(u) = R u \iff \\
f(x,y) = R \, (x,y)^T = ((x-y)/\sqrt{2}, (x+y)/\sqrt{2})^T
$$
for any vector $u$.
The line $x + y = 1$ can be written as
$$
(x, y)^T = (x, 1-x)^T = x(1,-1)^T + (0,1)^T \quad (x \in \mathbb{R})
$$
so it does not intersect the origin.
We can apply the rotation to those vectors:
$$
f(x, 1-x)
= ((x-(1-x))/\sqrt{2}, (x+(1-x))/\sqrt{2})^T \\
= ((2x-1)/\sqrt{2}, 1/\sqrt{2})^T \quad (x \in \mathbb{R}) \\
= (t, 1/\sqrt{2})^T \quad (t \in \mathbb{R})
$$
which is the horizontal line $y = 1/\sqrt{2}$.

The image above shows the line $x+y=1$ in red colour, and the rotated line in green colour.