sktime: [BUG] ColumnTransformer Unnecessary ValueError in ForecastingPipeline

Describe the bug When using ColumnTransformer in ForecastingPipeline,

• model.predict gives ValueError: If passed as a pd.DataFrame, X must be a nested pd.DataFrame, with pd.Series or np.arrays inside cells.
• model.update gives TypeError: y must be in an sktime compatible format, of scitype Series, Panel or Hierarchical, for instance a pandas.DataFrame with sktime compatible time indices, or with MultiIndex and last(-1) level an sktime compatible time index. Allowed compatible mtype format specifications are: ['pd.Series', 'pd.DataFrame', 'np.ndarray', 'nested_univ', 'numpy3D', 'pd-multiindex', 'df-list', 'pd_multiindex_hier'] . See the data format tutorial examples/AA_datatypes_and_datasets.ipynb. If you think the data is already in an sktime supported input format, run sktime.datatypes.check_raise(data, mtype) to diagnose the error, where mtype is the string of the type specification you want. Error message for checked mtypes, in format [mtype: message], as follows

To Reproduce

from sktime.transformations.series.boxcox import LogTransformer
from sktime.transformations.series.exponent import ExponentTransformer
from sktime.datasets import load_longley
from sktime.transformations.panel.compose import ColumnTransformer
from sktime.forecasting.sarimax import SARIMAX
from sktime.forecasting.model_selection import temporal_train_test_split
from sktime.transformations.compose import OptionalPassthrough

y, X = load_longley()
y_train, y_test, X_train, X_test = temporal_train_test_split(y, X)

selected_columns_1 = ["GNP", "UNEMP"]
selected_columns_2 = ["GNPDEFL"]

pipe = ColumnTransformer(
    [
        ("log", LogTransformer(), selected_columns_1),
        ("exp", ExponentTransformer(power=2), selected_columns_2),
    ],
    remainder="passthrough",  # "drop"
)

# Workaround for ValueError just for `predict` method
# pipe = OptionalPassthrough(pipe)

model = pipe ** SARIMAX()
model.fit(y_train, X_train)

# Gives ValueError if OptionalPassthrough is not used
model.predict(fh=[1, 2, 3, 4], X=X_test)

# Gives TypeError either way
model.update(y=y_test.iloc[:2], X=X_test.iloc[:2, :])

Expected behavior No ValueError should be thrown.

Additional context

Versions

python: 3.8.15 executable: /usr/bin/python machine: Linux-5.19.0-26-generic-x86_64-with-glibc2.17

Python dependencies: pip: 22.3.1 setuptools: 65.5.0 sklearn: 1.2.0 sktime: 0.15.0 statsmodels: 0.13.5 numpy: 1.22.4 scipy: 1.9.3 pandas: 1.5.2 matplotlib: 3.6.2 joblib: 1.2.0 numba: 0.56.4 pmdarima: None tsfresh: 0.20.0