Genomic RNA of hepatitis C virus (HCV) has an internal ribosome entry site (IRES), a specific highly structured fragment responsible for its non-canonical translation initiation. The HCV IRES contains a major part of the 5′-untranslated region of the viral RNA and a small portion of the open reading frame (ORF). At the first step of initiation, IRES directly binds to 40S ribosomal subunits so that the AUG start codon appears at the P site region without scanning and without involving initiation factors. However, it is still not entirely clear whether the IRES ORF is correctly loaded into the 40S ribosomal mRNA binding channel in the resulting binary complex. To address this issue, we applied site-directed cross-linking using HCV IRES derivatives bearing a perfluorophenyl azide cross-linker at nucleotides in definite positions relative to the adenine of the AUG start codon. We found that the modifier at the IRES position −3 cross-links to ribosomal proteins uS11 and eS26. These proteins have been identified together with uS7 as those interacting with the mRNA nucleotide in position −3 relative to the first nucleotide of the codon directed to the P site by a cognate tRNA. Thus, our results indicate a certain difference in the locations of the above parts of HCV IRES and canonical mRNAs on 40S subunits. The modifier at the IRES positions +4/5 was attached to uS19, which is specific for ribosomal complexes with the P site tRNA and similar derivatives of model canonical mRNAs when the modifier is in the same positions. However, the cross-linking efficiency of the IRES derivative was drastically lower than that previously observed with derivatives of model mRNAs. This implies that the IRES ORF portion is correctly loaded into the mRNA binding channel only in a tiny fraction of the binary complexes.