The answer seems somewhat complex for 229E. It seemingly evolved in us for better utilization of TMPRSS2, as Shirato et al. (2016) report:
Human coronavirus 229E (HCoV-229E), a causative agent of the common cold, enters host cells via two distinct pathways: one is mediated by cell surface proteases, particularly transmembrane protease serine 2 (TMPRSS2), and the other by endosomal cathepsin L. Thus, specific inhibitors of these proteases block virus infection. However, it is unclear which of these pathways is actually utilized by HCoV-229E in the human respiratory tract. Here, we examined the mechanism of cell entry used by a pseudotyped virus bearing the HCoV-229E spike (S) protein in the presence or absence of protease inhibitors. We found that, compared with a laboratory strain isolated in 1966 and passaged for a half century, clinical isolates of HCoV-229E were less likely to utilize cathepsin L; rather, they showed a preference for TMPRSS2. Two amino acid substitutions (R642M and N714K) in the S protein of HCoV-229E clinical isolates altered their sensitivity to a cathepsin L inhibitor, suggesting that these amino acids were responsible for cathepsin L use. After 20 passages in HeLa cells, the ability of the isolate to use cathepsin increased so that it was equal to that of the laboratory strain; this increase was caused by an amino acid substitution (I577S) in the S protein. The passaged virus showed a reduced ability to replicate in differentiated airway epithelial cells cultured at an air-liquid interface. These results suggest that the endosomal pathway is disadvantageous for HCoV-229E infection of human airway epithelial cells; therefore, clinical isolates are less able to use cathepsin.
IMPORTANCE: Many enveloped viruses enter cells through endocytosis. Viral spike proteins drive the fusion of viral and endosomal membranes to facilitate insertion of the viral genome into the cytoplasm. Human coronavirus 229E (HCoV-229E) utilizes endosomal cathepsin L to activate the spike protein after receptor binding. Here, we found that clinical isolates of HCoV-229E preferentially utilize the cell surface protease TMPRSS2 rather than endosomal cathepsin L. The endosome is a main site of Toll-like receptor recognition, which then triggers an innate immune response; therefore, HCoV-229E presumably evolved to bypass the endosome by entering the cell via TMPRSS2. Thus, the virus uses a simple mechanism to evade the host innate immune system. Therefore, therapeutic agents for coronavirus-mediated diseases, such as severe acute respiratory syndrome (SARS) and Middle East respiratory syndrome (MERS), should target cell surface TMPRSS2 rather than endosomal cathepsin.
The same authors have another (followup, or sorts) paper on two other coronaviruses:
Here, we show that clinical isolates of HCoV-OC43 and -HKU1 preferred the cell-surface TMRRSS2 to endosomal cathepsins for cell entry, similar to HCoV-229E.
Somewhat similar observations for NL63 (but these lack a historical/evolutionary perspective, since NL63 was only discovered in 2004.)
TMPRSS2 is important during early stages of the infection. It was previously suggested that coronaviruses may bypass the endocytic entry route employing transmembrane protease serine 2 (TMPRSS2), which primes the fusion protein and enables fusion of viral and cellular membranes on the cell surface (28, 29). We have tested whether inhibition of TMPRSS2 with camostat affects the HCoV-NL63 infection. We observed that inhibition of TMPRSS2 hampers virus infection in HAE cultures, while it has no effect on virus replication in LLC-MK2 cells. No inhibition of virus entry was observed in any of the models, as tracked with confocal microscopy visualizing the nucleoprotein.
Recent reports on other coronaviruses (28, 29, 35) suggested that these viruses may bypass the endocytic entry route using TMPRRS2 as the priming protease, enabling entry directly from the cell surface. Our experiments showed that inhibition of this protease indeed inhibited virus infection. Interestingly, it did not hamper virus internalization into the cell. Our data are consistent with the data presented by others (28, 29, 35), yet we believe that there is a different mechanistic explanation for the observed phenomenon. We believe that TMPRRS2 indeed is required for the virus-cell fusion, acting similarly to cathepsins, but it does not enable fusion on the cell surface, and the acidification of the microenvironment is required.
(HAE = "cultured human airway epithelium"; "LLC-Mk2 cells (ATCC CCL-7; Macaca mulatta kidney epithelial)" The original paper on the discovery of NL63 noted that "The in vitro host cell range of HCoV-NL63 is notable because it replicates on tertiary monkey kidney cells and the monkey kidney LLC-MK2 cell line.")
So for NL63 too, TMPRSS2 is basically a "catalyst" for cell entry, but its absence while slowing down NL63 didn't quite stop it from entering human airway epithelium cell... but (if I'm reading it correctly) the paper found that despite the virus still being pulled into the cell (in the absence of TMPRSS2) it wasn't as effective/infectious... (Given the latter para, which is from their discussion section, it seems this is a somewhat tentative/debatable conclusion.)