Appending just one column of a dataframe to another dataframe

df_1 =

metadata_id     col_1     col_2      col_3
        100         1         2          3
        101         1         2          1
        102         2         5          3
        103         1         2          6
        104         1         4          2
        105         6         2          3


df_2 =

      my_metadata_id       col_4      col_5     version
                 100           1          2          10
                 101           1          7          11
                 201           2          2          11
                 102           1          2          12
                 202           1          2          12
                 302           6          5          12
                 103           1          2          13
                 104           2          1          14
                 105           1          2          15

Desired output:

metadata_id     col_1     col_2      col_3     version
        100         1         2          3          10
        101         1         2          1          11
        102         2         5          3          12
        103         1         2          6          13
        104         1         4          2          14
        105         6         2          3          15

It looks trivial, but when I do:

df_1 = df_1.join(df_2.select("version"), df_1("metadata_id") == df_2("my_metadata_id"), how="left")  

I get: TypeError: 'DataFrame' object is not callable.

I would just join normally and then drop the extra columns, but in reality there is a lot of them and I wouldn't do it manually when there must be an elegant solution...