We analyze the validity of a commonly used identification between structures of the virtual photon $\gamma^*\to Q\bar Q$ and vector meson $V\to Q\bar Q$ transitions. In the existing studies of $S$-wave vector-meson photoproduction in the literature, such an identification is typically performed in the light-front (LF) frame while the radial component of the meson wave function is rather postulated than computed from the first principles... The massive photon-like $V\to Q\bar Q$ vertex, besides the $S$-wave component, also contains an extra $D$-wave admixture in the $Q\bar Q$ rest frame. However, the relative weight of these contributions cannot be justified by any reasonable nonrelativistic $Q\bar Q$ potential model. In this work, we investigate the relative role of the $D$-wave contribution starting from the photon-like quarkonium $V\to Q\bar Q$ transition in both frames: in the $Q\bar Q$ rest frame (with subsequent Melosh spin transform to the LF frame) and in the LF frame (without Melosh transform). We show that the photon-like transition imposed in the $Q\bar Q$ rest frame leads to significant discrepancies with the experimental data. In the second case we find that the corresponding total $J/\psi(1S)$ photoproduction cross sections are very close to those obtained with the "$S$-wave only" $V\to Q\bar Q$ transition, both leading to a good description of the data. However, we find that the "$S$-wave only" transition leads to a better description of photoproduction data for excited heavy quarkonium states, which represent a more effective tool for study of $D$-wave effects. Consequently, the predictions for production of excited states based on the photon-like structure of $V\to Q\bar Q$ transition should be treated with a great care due to a much stronger sensitivity of the $D$-wave contribution to the nodal structure of quarkonium wave functions. (read more)